Proefschrift Definitief Lage Resolutie

On-ward observations
in neonatal intensive care:

Towards safer supplemental oxygen & IV therapy
BW_Proefschrift 2 aug.indd 1
02-08-12 11:36
02-08-12 11:36
in neonatal intensive care:
Proefschrift
ter verkrijging van de graad van doctor
aan de Technische Universiteit Delft,
op gezag van de Rector Magnificus prof. ir. K.C.A.M. Luyben,
voorzitter van het College voor Promoties,
in het openbaar te verdedigen op maandag 24 september 2012 om 15.00 uur
door
Anne Catherine VAN DER EIJK
Ingenieur in Biomedical Engineering
geboren te Zwolle
02-08-12 11:36
Dit proefschrift is goedgekeurd door de promotoren:

Prof. dr. J. Dankelman
Prof. dr. H.J. Simonsz
Copromotor:
Dr. B.J. Smit
Samenstelling promotiecommissie:
Rector Magnificus,
Technische Universiteit Delft, voorzitter
Prof. dr. J. Dankelman,
Technische Universiteit Delft, promotor
Prof. dr. H.J. Simonsz,
Erasmus Medisch Centrum Rotterdam, promotor
Dr. B.J. Smit,
Erasmus Medisch Centrum Rotterdam, copromotor
Prof. dr. ir. C.A. Grimbergen,
Technische Universiteit Delft
Academisch Medisch Centrum Amsterdam
Prof. dr. S. Bambang Oetomo,
Technische Universiteit Eindhoven
Maxima Medisch Centrum Veldhoven
Prof. dr. F. van Bel,
Universitair Medisch Centrum Utrecht
Prof. dr. ir. R.M. Verdaasdonk,
VU Medisch Centrum Amsterdam
Prof. dr. ir. R.H.M. Goossens,
Technische Universiteit Delft, reservelid
The research presented in this thesis was partially supported by

Philips Medical Systems, Boeblingen, Germany
ODAS foundation, Delft, The Netherlands
Lay-out Louise de Kruijf

ISBN 978-94-6191-355-5
Copyright 2012, A.C. van der Eijk
All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means,
without the prior written permission of the author.
02-08-12 11:36
Wires
You got wires, going in
You got wires, coming out of your skin
You got tears, making tracks
I got tears, that are scared of the facts
Running down corridors, through automatic doors
Got to get to you, got to see this through
I see hope is here, in a plastic box
Ive seen christmas lights, reflect in your eyes
You got wires, going in
You got wires, coming out of your skin
Theres dry blood, on your wrist
Your dry blood on my fingertip
Running down corridors, through automatic doors
Got to get to you, got to see this through
First night of your life, curled up on your own
Looking at you now, you would never know
I see it in your eyes, I see it in your eyes
Youll be alright
Artist: Athlete
02-08-12 11:36
02-08-12 11:36
Table of Contents
Chapter 1
General introduction
11
1.1 Introduction
1.2 Preterm infants, definitions & prevalence
1.3 Prematurity, causes & outcome
1.4 Prematurity & oxidative stress related
diseases
1.4.1 Bronchopulmonary dysplasia
1.4.2 Infant respiratory distress syndrome
1.4.3 Patent ductus arteriosus
1.4.4 Retinopathy of prematurity
1.5 Supplemental oxygen therapy
1.5.1 A brief overview of history of supplemental
oxygen therapy
1.6 IV therapy
1.6.1 A brief overview of history of IV therapy
1.7 Problem statements
1.7.1 Supplemental oxygen therapy
1.7.2 IV therapy
1.8 Objectives
1.8.1 Part I Supplemental oxygen therapy
1.8.2 Part II IV therapy
1.9 Thesis outline
1.10 References
13
14
15
Part I Supplemental oxygen therapy
33
Chapter 2
Oxygenation in preterm infants; background,
target ranges & monitoring techniques
35
2.1 Introduction
2.2 Oxygen in the human body
2.2.1 The oxygen dissociation curve
2.3 Monitoring of oxygenation
2.3.1 Physical assessment of the skin
2.3.2 Blood gas analysis
2.3.3 Continuous intra-arterial blood gas
monitoring
2.3.4 Transcutaneous oxygen measurement
2.3.5 Pulse oximetry
2.3.6 Near infrared spectroscopy
2.3.7 Capnography
2.4 Reference values for blood oxygen levels
2.4.1 Reference values for oxygen saturation
2.4.2 Reference values for partial pressure of
oxygen
2.4.3 Target ranges & outcome
2.5 Monitoring oxygenation in preterm infants:
future perspectives
2.6 References
15
18
Chapter 3
New-generation pulse oximeters in extremely
low birth weight infants: how do they perform
in clinical practice?
3.1 Introduction
3.2 Methods
3.2.1 Patients
3.2.2 Study set-up
3.2.3 Experimental set-up
3.2.4 Data analysis
3.3 Results
3.3.1 SpO2 values
3.4 Discussion
3.5 References
57
59
60
62
69
72
19
21
24
25
26
37
38
41
47
49
51
Chapter 4
Manual adjustments of the inspired oxygen
fraction in extremely low birth weight infants
4.1 Introduction
4.2 Methods
4.2.1 Patients
4.2.3 Data collection
4.2.4 Target levels for SpO2
4.2.5 FiO2 adjustments
4.3 Results
4.3.1 Patients
4.3.2 Manual FiO2 adjustments
4.3.3 SpO2 levels
4.4 Discussion
4.5 References
Chapter 5
Pulse oximetry alarm limits in extremely low birth
weight infants: when do deviations from the
protocol occur?
5.1 Introduction
5.2 Methods
5.2.1 Working situation
5.2.2 Policy for pulse oximetry alarm limits
5.2.3 Data analysis
5.3 Results
5.3.1 Occurrence of alarm limits
5.3.2 FiO2 levels
5.3.3 SpO2 levels
5.3.4 Characteristics of the alarm limit
adjustments
5.4 Discussion
5.5 References
75
77
78
80
86
89
91
93
94
95
102
105
02-08-12 11:36
Chapter 6
Manual control of oxygenation in extremely low
birth weight infants: what is the nurses point
of view?
6.1 Introduction
6.2 Methods
6.2.1 The questionnaire
6.3 Results
6.3.1 Manual control of oxygenation
6.3.3 Pulse oximetry alarm limits
6.3.4 Suggestions for improvement
6.4 Discussion
6.5 References
Chapter 7
Defining hazards of supplemental oxygen therapy
in neonatology using the Failure Mode and Effects
Analysis (FMEA) tool
7.1 Introduction
7.1.1 FMEA
7.2 Methods
7.3 Results
7.3.1 Step 1. Defining the topic
7.3.2 Step 2. Team assembly
7.3.3 Step 3. Process analysis
7.3.4 Step 4. Hazard analysis
7.3.5 Step 5. Develop risk reduction methods
7.4 Discussion
7.4.1 Lessons learnt by the FMEA-team
7.5 References
Part II IV therapy
Chapter 8
Flow-rate variability in neonatal IV therapy: what
do we know about the flow?
8.1 Introduction
8.2 Methods
8.3 Results
8.3.1 Factor 1: Vertical syringe or patient
displacement
8.3.2 Factor 2: Syringes
8.3.3 Factor 3: Infusion tubing
8.3.4 Factor 4: Check valves & anti-siphon valves
8.3.5 Factor 5: Inline filters
8.3.6 Factor 6: Add-on devices
8.3.7 Factor 7: Vascular access devices
8.4 Discussion
8.5 References
Chapter 9
Flow-rate variability in neonatal IV therapy
caused by the use of check valves
161
109
111
112
112
116
118
9.1 Introduction
9.2 Methods
9.2.1 Check valves
9.2.3 Study set-up
9.3 Results
9.3.1 Experiment I: Adding syringes
9.3.2 Experiment II: Changing height
9.4 Discussion
9.5 References
Chapter 10
General discussion
121
123
124
126
163
164
168
173
176
179
10.1 Main conclusions

10.1.2 IV therapy
10.2 On the research approach
10.2.2 IV therapy
10.3 The need for standardization
10.4 Future work
10.4.1 Changes in culture
10.4.2 Technical improvements
10.5 Conclusion
10.6 References
181
Appendices
191
Summary
201
Samenvatting
207
Dankwoord
213
About the Author
221
182
183
185
187
188
131
133
135
137
139
140
140
155
157
02-08-12 11:36
02-08-12 11:36
10
02-08-12 11:36
CHAPTER 1
02-08-12 11:36
12
CHAPTER 1
02-08-12 11:36
13
1.1 INTRODUCTION
Preterm infants are infants that are born before the expected date of birth. Due to their immaturity, they often require intensive care to survive with a good health outcome. Intensive
care is the division of medicine that is concerned with the continuous monitoring and support of vital functions of critically ill patients. To be able to meet the specific needs of preterm
infants, dedicated neonatal intensive care units (NICUs) have been established worldwide. In
these NICUs virtually all preterm infants receive supplemental oxygen therapy and intravenous (IV) therapy. Both therapies, essential but potentially dangerous, will be studied in this
thesis.
Supplemental oxygen therapy refers to the therapy where a gas mixture with >21% of oxygen is supplied to the patient via (mechanical) ventilation. Due to the immaturity of the preterm infants lungs, supplemental oxygen therapy is often needed immediately after birth to
reach and maintain adequate oxygenation of the preterm infant. Unfortunately supplemental oxygen therapy is not without risk. Both too high and too low blood oxygen levels may
have severe consequences for the development of the preterm infant.
The immaturity of organs and/or severe illness are also reasons why IV therapy is essential
for preterm infants hospitalised on a NICU. In IV therapy various types of nutrition, drugs,
and/or fluids are administered directly into the veins of the patient via a vascular access
device. Because of the limited vascular access possibilities, multi-infusion is used. In multiinfusion therapy, several infusions are supplied to the infant via a single catheter. To administer the IV fluids with a pre-programmed flow-rate into the patient, syringe pumps are
frequently used. Although it is expected that the IV substances are supplied to the patient
with the pre-programmed flow-rate, it has been shown that the actual volume delivered
to the patient can vary over time. Especially in preterm infants, these changes in delivered
volume can have severe consequences.
In the next paragraphs the background of prematurity, supplemental oxygen therapy, and
IV therapy are discussed in more detail. In the final paragraphs of this chapter (1.7 to 1.9)
the problem statements, objectives, and thesis outline are presented.
To be able to meet the specific needs of preterm

infants, dedicated neonatal intensive care units
have been established worldwide.
02-08-12 11:36
14
CHAPTER 1
1.2 Preterm infants, definitions & prevalence

A term, healthy newborn infant is born after a pregnancy duration of 37 to 42 weeks with
a birth weight (BW) of approximately 3.5 kg. The pregnancy duration, or gestational age
(GA), is calculated from the first day of the last menstruation of the mother. When an infant is born with a GA 37 weeks it is classified as preterm, very preterm (<30 weeks),
or extremely preterm (<28 weeks). When the birth weight of the infant is less than 2500
grams, it is defined as low birth weight (LBW), very low birth weight (1500 g.; VLBW) or,
extremely low birth weight (1000 g.; ELBW).
In the past, infants born preterm and/or with a low birth weight had little chance of survival
due to a lack of knowledge about the needs of this specific group of patients. Since the term
neonatology was first mentioned in a textbook in 1960, major developments in neonatal
care have been made.1, 2 Thanks to on-going research and improved treatment, both survival
rates and long term outcome of preterm infants have improved enormously.3 Nowadays,
infants born after only 24 weeks of pregnancy are more likely to survive than ever before.
While the medical care for (preterm) newborn infants has advanced, the prevalence of preterm births has also been increasing.4-6 Currently, in the United States the preterm delivery
rate is about 12 to 13%; in Europe it is lower, about 5 to 9%. Worldwide, about 80% of the preterm infants is born after a pregnancy duration of 32 to 36 weeks. About 15% is born between
28 and 31 weeks, and 5% is born before 28 weeks.7 In the Netherlands, in the first decade of
the 21st century, the percentage of births <32 weeks (including still births) was between 1.4
and 1.6%. Figure 1.1 presents the corresponding number of births <32 weeks and the number
of survivors at postnatal day 28.
22.0-23.6 wks
24.0-24.6 wks
25.0-25.6 wks
3000
26.0-27.6 wks
28.0-31.6 wks
Number of births
2500
2000
1500
1000
500
2008, Survivors
2008, Births
2007, Survivors
2007, Births
2006, Survivors
2006, Births
2005, Survivors
2005, Births
2004, Births
2004, Survivors
2003, Survivors
2003, Births
2002, Survivors
2002, Births
2001, Births
2001, Survivors
Figure 1.1 Number of births (including still births) <32 weeks, and the number of survivors at postnatal
day 28 in the Netherlands from 2001 to 2008. (Graph based on data obtained from Stichting Perinatale
Registratie Nederland).
02-08-12 11:36
15
1.3 Prematurity, causes & outcome

Preterm infants are either born spontaneously (40 to 45%), by induced labour or by
caesarean section. The latter two are performed because of preterm premature rupture of
membranes (25 to 30%) or because of maternal or foetal indications (30 to 35%).7 The cause
for prematurity is multifactorial and includes maternal characteristics like race, age, weight,
health in general, and pregnancy history. The educational status, socio-economic status,
and marital status of the mother are also known to influence the pregnancy duration.8-14
Women who are exposed to drugs, heavy alcohol use, or tobacco use during pregnancy are
known to be more likely to have babies with a low birth weight.15-17 Characteristics of the
pregnancy, like assisted reproductive technologies and multiple gestation, are also risk factors for preterm labour: about 50 to 60% of all multiple gestation pregnancies end in preterm birth. While only 2 to 3% of the infants is part of a multiple gestation, they account for
15 to 20% of all preterm births.7, 18
Although the chances of survival for preterm infants have increased enormously in the last
decades, there is still a large part of the surviving infants that suffer from disorders or disabilities. The disabilities cover, amongst others, cerebral palsy, developmental delay, visual
or hearing impairment, speech and language difficulties, and chronic lung disease.19, 20
The risk of adverse outcome is strongly related to the pregnancy duration and birth weight.
Hille et al.21 showed that in the Netherlands 36% of the ELBW infants had moderate to severe problems in overall outcome at the age of 19 years. However, it is difficult to interpret
these numbers because the medical care provided to newborn infants develops continuously, and long term outcome data are always behind on the current status of neonatal care.
1.4 Prematurity & oxidative stress related diseases

Being born is, from a physiological point of view, a very dramatic event. The intrauterine
environment (i.e. in the womb) is warm, sterile, and dark. Oxygen and nutrition are supplied
from the mother to the foetus via the umbilical cord. One of the major changes between
the intrauterine environment and the extrauterine environment (i.e. outside world) is the
difference in oxygen tension, the partial pressure of oxygen (PO2). The intrauterine PO2 is
about 3 kPa (20 to 25 mmHg), this is comparable with the atmosphere at the top of the Mount
Everest. Thus, foetal development takes place in a relative hypoxic environment compared to
the atmosphere at sea level where the PO2 is about 21 kPa (155 to 160 mmHg).22-24
The sudden change in oxygen tension after birth results in a sharp increase in reactive oxygen species (ROS). ROS are chemically reactive molecules that contain oxygen. At low levels,
ROS contribute to homeostasis and cell signalling processes. At higher levels, ROS can lead
to oxidative stress in the cell. Oxidative stress is defined as the imbalance between oxidants
02-08-12 11:36
16
CHAPTER 1
and antioxidants. Although oxidants and antioxidants both are necessary for maintaining
life, an imbalance in favour of the oxidants may result in cell damage or even cell death. The
imbalance can occur due to an increase in oxidant production and/or an inadequate antioxidant production. To prevent oxidative stress after birth, several processes take place in
the foetus in the final weeks before term birth. These processes comprise, amongst others,
an increase in both antioxidants and lung surfactant. Surfactant is a fluid that lowers the
surface tension of the lungs, making it easier to inflate them.23, 25
Preterm infants lack the preparation for the sudden increase in oxygen tension because
they are born too early. Consequently, their defence system for antioxidants is immature,
and therefore, they are at increased risk for oxidative stress after birth. The risk for oxidative stress increases even more when preterm infants receive supplemental oxygen therapy
and/or develop infections.26-28
In 1988 Saugstad was the first to mention the term oxygen radical disease of the newborn.29
He stated that several diseases in preterm infants have a common pathogenesis via oxidative stress. Since then, it became clear that diseases typical for neonatal intensive care like
bronchopulmonary dysplasia, infant respiratory distress syndrome, necrotizing enterocolitis, retinopathy of prematurity, patent ductus arteriosus and periventricular leukomalacia
are associated with oxidative stress. However, it is not always clear whether the presence of
oxidative stress is a cause or a result of the disease process.26, 29-31 To show examples of the
possible consequences of suboptimal or incorrect use of supplemental oxygen therapy, four
of the disorders mentioned above are discussed in more detail in the next paragraphs.
1.4.1 Bronchopulmonary dysplasia
Bronchopulmonary dysplasia (BPD) is a disorder characterised by respiratory distress and
airway inflammation.32 The disorder was first described in 1967 by Northway et al.,33 and
diagnosed when there was a need for supplemental oxygen therapy at a postnatal age of
28 days. Because since then the GA of surviving preterm infants decreased, the definition
is not valid anymore. Therefore, the term new-BPD was introduced. New-BPD is diagnosed when there is need for supplemental oxygen therapy or ventilatory support at a postmenstrual age of 36 weeks, regardless of the GA or postnatal age.
Studies on the prevalence of BPD found the disorder in 22% of all infants with a birth weight
between 501 and 1500 grams.34, 35 The exact pathogenesis of BPD is unclear, but it seems to be
multifactorial. Risk factors are, amongst others, low birth weight, mechanical ventilation,
and supplemental oxygen therapy. To treat the symptoms of BPD several types of medication, like corticosteroids, are available. In some cases special mechanical ventilation is
needed to prevent further lung damage.36-39
02-08-12 11:36
17
1.4.2 Infant respiratory distress syndrome

In order to breathe normally, the alveoli in the lungs need to be inflated. To inflate the
alveoli, surfactant is required. Due to the short pregnancy duration, in very preterm infants, there is often a surfactant deficiency. As a result of this lack of surfactant, the alveoli
collapse and the total lung capacity decreases. This phenomenon is referred to as (Infant)
respiratory distress syndrome ((I)RDS). Symptoms include laboured and fast breathing,
cyanosis, grunting, and nasal flaring.
To diagnose IRDS, radiography is used: a low lung volume is one of the signs for IRDS.
Due to the decreased lung capacity, mechanical ventilation and supplemental oxygen
therapy are required. The incidence of IRDS is inversely related to the pregnancy duration.
Thanks to the use of both antenatal steroids to promote lung maturation and the use of
surfactant therapy after birth, the incidence of IRDS in preterm infants has been reduced
enormously.36, 40
1.4.3 Patent ductus arteriosus
The blood circulation of a foetus includes a connection between the main pulmonary
artery and the aorta. This connection, the ductus arteriosus, allows the blood to bypass
the not yet ventilated lungs. After birth, several physiological changes cause the closure of
the ductus arteriosus to make sure that oxygen-poor blood starts to enter the lungs. When
the ductus arteriosus does not close (completely) within 24 to 72 hours after birth, a patent
ductus arteriosus (PDA) is diagnosed.41
About 65% of the infants born with a GA <30 weeks suffers from PDA.42 A PDA may have a
negative effect on (amongst others) the cerebral oxygenation and the pulmonary function
of the infant.43 The diagnosis of PDA is typically confirmed by echocardiography and can be
treated with medication or surgical ligation.44
1.4.4 Retinopathy of Prematurity
Retinopathy of prematurity (ROP) is a vasoproliferative disorder of the retina in preterm
infants and an important, potentially preventable, cause of blindness in childhood. In infants with ROP, the vascularisation of the retina is disturbed during the first weeks of extrauterine life.45, 46 The most important risk factors for the development of ROP are prematurity and supplemental oxygen therapy. Furthermore, factors like low birth weight, blood
transfusions, neonatal sepsis, and BPD are also associated with higher rates of ROP.47-67
ROP was first described by Terry in the 1940s.68 Nowadays the disorder is diagnosed
most frequently in middle income countries where there is a lack of qualified personnel,
resources, and screening and treatment programs.69-71 However, even in the more equipped
centres with a long established neonatal intensive care, the incidence of ROP in infants
with a birth weight 1250 gram is still 10% to 15%.72, 73 Treatment of ROP includes the retinal
ablation of the avascular retina by cryotherapy or laser photocoagulation. The best method
to prevent ROP is adequate screening and restricted use of oxygen.60, 74
02-08-12 11:36
18
CHAPTER 1
1.5 Supplemental oxygen therapy

After the outline about the diseases that are caused or followed by suboptimal oxygenation
it is obvious why maintaining adequate oxygenation in preterm infants is a major issue. In
healthy adults and children adequate oxygenation is maintained by a network of complex
processes with multiple parameters influencing each other. To reach and maintain adequate
oxygenation in preterm infants, it would be desirable for neonatologists to be able to monitor and control these parameters (continuously). Although current mechanical ventilators
and monitoring techniques are sophisticated, it is still difficult to monitor and control the
relevant parameters in preterm infants.
One of the parameters that can be monitored is the arterial oxygen saturation (SaO2). In
preterm infants the SaO2 is determined intermittently by blood gas analysis, and continuously with a non-invasive sensor, the pulse oximeter. The SaO2 determined by pulse
oximetry is referred to as SpO2. When the SpO2 level is outside the desired range an alarm
sounds. To recover the SpO2 level NICU staff can, amongst others, adjust the fraction of
inspired oxygen (FiO2) in the gas mixture supplied to the infant. The FiO2 can be adjusted
from 21% (room air) to 100% (pure oxygen).
Currently, in neonatal intensive care the FiO2 level is adjusted manually, mainly by the
nursing staff. A simplified block scheme of this process is shown in Figure 1.2. Although
supplemental oxygen therapy increased chances of survival after preterm birth, it has been
known for about 60 years that supplemental oxygen therapy in preterm infants is not without risks.75-77
Target
level
SaO2
Human
controller
Patient
Alarm
SaO2
level in
blood
Sensor
Figure 1.2 The process of controlling oxygenation. The input of the system is the target level of oxygen
saturation (SaO2) at the left side of the figure. The desired SaO2 level is compared with the SaO2 level
measured by a sensor (e.g., by a pulse oximeter). When the difference between the desired SaO2 level
and the measured SaO2 level is too large, an alarm sounds. The human controller can decide to take
action, for instance adjusting the fraction of inspired oxygen (FiO2). The FiO2 is supplied to the patient
by a ventilator used for respiration. In the blood of the patient, the SaO2 level changes due to the change
in FiO2. The sensor measures a new SaO2 level which is again compared with the desired SaO2 level.
02-08-12 11:36
19
1.5.1 A brief overview of history of supplemental oxygen therapy

Soon after the discovery of oxygen in 1774, this life-giving gas was used for medical purposes. In 1780 the Frenchman Chaussier was the first who used oxygen for newborn infants
with respiratory problems. From then it took one and a half century before oxygen was
widely and liberally used for respiratory support in newborn infants in the 1940s.76-78 The
negative effects of the use of oxygen became clear several years later. In 1951, Dr. Campbell was the first who assumed there was a relation between supplemental oxygen therapy
and retrolental fibroplasia, the blindness disorder now better known as ROP (see 1.4.4).79
Several other studies confirmed her hypothesis.80, 81
In 1954 a large trial was performed to investigate the risks of supplemental oxygen therapy.
The conclusion of this trial was that it was safe to give oxygen to newborn infants as long
as the FiO2 was below 40%.82, 83 Although some serious methodological errors were made
in this trial, the results were widely accepted. It was so strongly believed that FiO2 >40%
was harmful that, when an infant developed ROP, the hospital was accused for malpractice.
After all, the ROP was the proof that FiO2 had exceeded the 40%.75, 84
In the years after the large trial the incidence of ROP reduced dramatically. However, the
incidence of cerebral palsy and mortality increased.85, 86 As a result of this change in prevalence of both mortality and morbidity and the fact that it became possible to monitor blood
oxygen levels, clinicians recognised more and more that they should not restrict the supply
of FiO2, but that they should restrict the actual blood oxygen levels of the preterm infant
itself. This understanding led to a more accurate control of blood oxygen levels and a more
tailored approach to the use of supplemental oxygen therapy in the NICU.87-89
1.6 IV therapy
In IV therapy, various types of parenteral nutrition, drugs, and/or fluids are administered
directly into the veins of the patient. To deliver these IV substances with a pre-programmed
flow-rate to the patient, a mechanical pump pushes the plunger of a syringe with a preprogrammed velocity into the syringe. The IV substance in the syringe flows into the
patient via flexible tubing and a vascular access device.
Often, because intravenous access in preterm infants is limited and several IV substances
need to be administered simultaneously, multi-infusion is used. In multi-infusion, several
syringe pumps are connected via tubing to a single vascular access device. These connections are conducted with add-on devices. A schematic overview of a multi-infusion set up
is shown in Figure 1.3. Although multi-infusion creates the possibility to provide various
substances simultaneously, the accuracy and the predictability of the volumes delivered to
the patient is limited.
02-08-12 11:36
20
CHAPTER 1
Syringe pumps with

syringes
Infusion tubing
Stopcock with
3-way valves
Vascular
acces device
To patient
Figure 1.3 Schematic overview of multi infusion intravenous (IV) therapy. Two or more syringe pumps
are connected to one vascular access device via infusion tubing and an add-on device (e.g., a stopcock
with 3-way valves).
1.6.1 A brief overview of history of IV therapy

The first attempts of IV therapy were already made in the Middle Ages. In 1492 the ill Pope
Innocent VIII was transfused with blood from three boys via vein-to-vein anastomosis. Regrettably the pope and donors died. From then until the second half of the 17th century the
knowledge of blood, the blood circulation, and IV therapy increased enormously. This increase in knowledge was realised by performing numerous experiments with blood transfusions in and between animals and humans. However, because these experiments frequently
resulted in the death of the subjects, several governments and churches decreed the performance of blood transfusions as a criminal act.90, 91
In the 1800s, the work of Dr. William Brooke OShaughnessy and his student Thomas Latta
formed the basis for modern IV therapy. During the cholera epidemic in England in the
1830s, OShaughnessy realised that the typical thick black blood of the cholera victims was
a result of a shortage of water, saline, and alkali. Therefore, he indicated that the patients
needed injections of water and salts in the bloodstream. In 1832, Thomas Latta applied the
recommendations, and saved 8 of the 25 victims he treated with intravenous saline using a
small silver tube attached to a syringe.90, 92
02-08-12 11:36
21
After the results of OShaughnessy and Latta, it lasted until the two World Wars before
further innovations in IV therapy were made. In 1933 IV solutions came on the market in a
vacuum bottle, which eliminated microbial growth and pyrogens. In 1940, the Massachusetts General Hospital started the first IV team. This idea of organizing special IV teams
became very popular during the 1970s. Since then great advances in IV therapy were made.
These advances were, amongst others, the result of the development of plastic IV materials.
Today, virtually all hospitalised patients receive IV therapy.90, 91
In newborn infants IV therapy is more complicated than in adults due to, amongst others,
the small size of the vessels. Fortunately, together with the progress in the development of
materials for IV therapy in adults, the manufacturing of dedicated materials for (preterm)
newborn infants advanced as well. For example, today it is possible to buy vascular access
devices with an outer diameter of only 0.35 mm to serve the needs for ELBW infants. However, despite the presence of dedicated materials for the smallest patients, the accuracy of
the actual delivered IV substances still needs to be improved.93-101
1.7 Problem statements

Both supplemental oxygen therapy and IV therapy are necessary, but potentially dangerous
therapies. In the next paragraphs, the problem statements for both therapies are discussed
separately.
In supplemental oxygen therapy the FiO2 can be adjusted from 21% (room air) to 100%
(pure oxygen) manually. Manual control of the oxygenation as described in Figure 1.2 is
time consuming and very difficult to do accurately, mainly because of the frequent and unpredictable fluctuations of the SpO2.102-107 When SpO2 is outside the desired range an alarm
sounds. The high rate of alarms may lead to anxiety of patients and their family, and to a
reduced or delayed reaction of the nursing staff.108 Although supplemental oxygen therapy
has been widely used in newborn infants for more than 60 years, there is still no consensus
about the target ranges for blood oxygen levels, and the best methods to give (preterm)
newborn infants this treatment.75-77
To increase the quality of supplemental oxygen therapy and to reduce workload of the
nursing staff, several groups worldwide have developed devices for (semi-)automatic control
of the oxygenation. Most of the devices for (semi-)automatic control of oxygenation adjust
the FiO2 supplied to the patient (semi-)automatically when SpO2 deviates from the target.
The first devices that were developed were dedicated servo systems.109-112 With the development of computer aided control the devices became more advanced: Proportional-IntegralDerivative control with or without adaptive models,52, 56, 103, 113-117 dual control methods,118-120
state machine106, 121 and fuzzy logic control122 have all been developed and tested.
To test the devices some groups used patient simulators,52, 114, 119 or animals,111, 117 but most of
the groups tested their controller on patients.106, 109, 110, 112, 115, 116, 118, 122-125 The first studies did not
02-08-12 11:36
22
CHAPTER 1
use an objective way to test the effectiveness of their automatic controller. This resulted in
no or weak conclusions.56, 111, 113, 116, 119, 121, 126-128 Some more recent papers included tests that actually show significant improvement. Amongst the tests was the time spent within the target
range for SpO2 for manual, dedicated, and (semi-)automatic control (Table 1.1). Comparing
the devices with each other is hampered by the fact that the nurse:patient ratio, the subject
characteristics, the study period, and the study methods varied between studies. However,
although the results of the developed devices are promising, differences between manual and
(semi-)automatic control are major, and effects on long term outcome are still unknown.129
table 1.1 Time spent within the target range for SpO2 for periods on manual, dedicated manual, and
(semi-)automatic control.
Time SpO2 is within target range
Manual control [%]
Dedicated control [%]
(semi-)
Automatic control [%]
Beddis109, Collins110
72
88
Dugdale112
45
75
Taube113
54
69
81
Bhutani115
54
69
81
121
Morozoff
17
46
Morozoff116
39
50
Sun122, 128
58
72
66
75
Urschitz106 Open loop
80
86
85
Urschitz106 Closed loop
82
91
91
Morozoff State machine
57
71
Morozoff123 Adaptive model
57
73
Morozoff123 Closed loop, PID
57
70
Claure124
42
58
Claure
32
40
First author
Claure105, 118
123
125
The term manual control is used to refer to the situation in normal daily NICU care. Dedicated control is a situation where a physician or nurse stays at the bedside of the patient and is focussing on
the control of the SpO2 only. In (semi-)automatic control a device controls SpO2 automatically, and/or
advises the NICU staff to make an adjustment in FiO2. Amongst others, the study methods and patient
characteristics differed, thus results between studies are difficult to compare.
02-08-12 11:36
23
To improve outcome for those situations where oxygenation is controlled manually, two
studies focused on the development of protocols to standardise when, why, and how FiO2
should be adjusted.130, 131 One of these studies actually showed a reduction in the incidence
of ROP.131 However, both studies mention difficulties with implementation of the protocol
and compliance to it. Because none of the studies actually quantified the performed manual
FiO2 adjustments, knowledge about the actual behaviour of NICU personnel with respect
to control of the oxygenation is still lacking.
1.7.2 IV therapy
IV therapy is hampered by a number of complications and limitations. The most wellknown are related to infections, and extravasation. While these complications are very
relevant, they are outside the scope of this thesis. The focus in this thesis is on a, probably
underestimated, limitation in IV therapy: flow-rate variability. This flow-rate variability
is caused by multiple factors and complicates the accuracy and predictability of the actual
volumes delivered to the patient. Moreover, the flow-rate variability can lead to, for instance,
changes in the haemodynamics and oxygenation of newborn infants.93, 97, 99, 132, 133 Therefore, it
is important to minimise flow-rate variability in IV therapy in clinical practice.
Two of the factors that affect the flow-rate variability are backflow and siphonage. In siphonage, there is uncontrolled emptying or free flow of substances from a syringe into
the patient. Siphonage can occur when the syringe is not clamped or is poorly clamped in
the syringe pump or when there are air leaks in the IV-administration set.101, 134-136 Backflow
can occur when multiple infusions are interconnected to each other (e.g., via a stopcock).
Because of differences in resistance in the IV-administration set, it is possible that fluids do
not flow from the syringe into the patient but into another line instead.137
To prevent backflow and/or siphonage there are various types of check valves available that
can be inserted in the IV-administration set. Paradoxically, while check valves are implemented in IV-administration sets to minimise backflow and/or siphonage, it has been shown
that the presence of these valves can enhance flow-rate variability as well.101, 138 Thus, to be able
to increase the accuracy and predictability of the volume delivered to the patient, the factors
influencing the flow-rate variability should be known and, where possible, controlled.
02-08-12 11:36
24
CHAPTER 1
1.8 Objectives
The primary objective of this thesis is to determine the limitations of supplemental oxygen
therapy and IV therapy in current neonatal intensive care and to identify areas for improvement. To distinguish between supplemental oxygen therapy and IV therapy, the thesis is
divided in two parts. The secondary objectives are listed in the next two paragraphs.
1.8.1 Part I Supplemental oxygen therapy
In Part I of the thesis, the work related to supplemental oxygen therapy in preterm infants
is discussed. The secondary objectives are:
I.I To obtain background information regarding oxygenation of the human body, to get
an overview of literature on target ranges for blood oxygen levels in newborn infants,
and to evaluate methods for monitoring oxygenation in neonatology.
I.II To evaluate the performance of new-generation pulse oximeters of three different
brands in ELBW infants.
I.III To quantify manual adjustments in the FiO2 performed by NICU personnel in ELBW
infants, in relation to SpO2 and bedside care.
I.IV To study the compliance to the protocol for pulse oximetry alarm limits in ELBW
infants in relation to FiO2, SpO2, and bedside care.
I.V To explore the decision making processes and obtain insight in the knowledge,
opinions, and attitude of the nursing staff towards supplemental oxygen therapy
in ELBW infants.
I.VI To prospectively evaluate hazards in the process of supplemental oxygen therapy
in very preterm infants hospitalised in a NICU.
1.8.2 Part II Intravenous therapy
In Part II of the thesis the work related to IV therapy in newborn infants is discussed. The
secondary objectives are:
II.I To study which factors are responsible for flow-rate variability in IV therapy with
syringe pumps.
II.II To evaluate the effect of three different types of check valves on flow characteristics
in a low-flow multi-infusion set.
02-08-12 11:36
25
1.9 Thesis outline

To meet the objectives, five studies were performed. These studies form, together with the
literature reviews on the discussed subjects, the basis of this thesis.
The first part of this thesis is about supplemental oxygen therapy in preterm infants. Chapter 2 is the first chapter of Part I and provides a literature review to meet objective I.I. First
the background information regarding oxygenation of the human body and especially that
of preterm infants is described. Next, the working principles and (dis)advantages of current
developed methods for monitoring oxygenation are elaborated on. Thereafter, an overview of literature on target ranges for blood oxygen levels in (preterm) newborn infants
is provided. Finally, a future perspective of the needs for oxygen monitoring in (preterm)
newborn infants is discussed. Chapter 3 describes a study to the performance of newgeneration pulse oximeters in ELBW infants to meet objective I.II. In this study three different brands of pulse oximeters were compared by dual SpO2 monitoring in nine ELBW
infants. In Chapter 4 and Chapter 5 an observational study is discussed. During this observational study on-ward video and data recording was performed to obtain insights in the
manual control of oxygenation in ELBW infants by healthcare professionals. This observational study was set up to meet both objectives I.III and I.IV. Chapter 6 provides the results of a survey amongst 24 NICU nurses. The questionnaire in this study was developed
to meet objective I.V. The questions assessed the knowledge, opinions, and attitude of the
nursing staff towards supplemental oxygen therapy in ELBW infants. The final chapter
of Part I is Chapter 7. In this chapter the hazards in the process of supplemental oxygen
therapy in preterm infants are evaluated prospectively to meet objective I.VI. The hazards
were analysed by a multidisciplinary team using the Failure Mode and Effects Analysis
(FMEA)-tool.
Part II of the thesis is about IV therapy in newborn infants and starts with Chapter 8 where
a literature review is discussed to meet objective II.I. Chapter 9 describes an in-vitro study
to the effect of three different types of check valves on the flow characteristics of a low-flow
multi-infusion set. This study was set up to meet objective II.II.
Chapter 10 provides the general discussion about both supplemental oxygen therapy and
IV therapy. The main conclusions and the research approach of the work performed in this
thesis are presented together with recommendations for future work.
It should be noted that Chapters 3 to 9 are written as separate papers, consequently, there
is a certain amount of overlapping information within these chapters. Furthermore, in this
thesis the masculine form is used for all healthcare professionals and patients, merely to
simplify the text. No discrimination is intended.
02-08-12 11:36
26
CHAPTER 1
1.10 References
1. Schaffer AJ, Avery ME. Disease of the newborn.
Philadelphia: WB Saunders Co; 1960.
2. Philip AGS. The evolution of neonatology.
Pediatric Research. 2005;58(4):799-815.
3. Saigal S, Doyle LW. Preterm birth 3: An overview of
mortality and sequelae of preterm birth from infancy
to adulthood. The Lancet. 2008;371(9608):261269.
4. Craig E, Thompson J, Mitchell E. Socioeconomic
status and preterm birth: New Zealand trends,
1980 to 1999. Arch Dis Child Fetal Neonatal Ed.
2002;86(3):F142-146.
5. Branum AM, Schoendorf KC. Changing patterns of
low birthweight and preterm birth in the United
States, 198198. Paediatric and Perinatal Epidemiology. 2002;16(1):8-15.
6. Alexander GR, Slay M. Prematurity at birth:
trends, racial disparities, and epidemiology.
Mental Retardation and Developmental Disabilities
Research Reviews. 2002;8(4):215-220.
7. Goldenberg RL, Culhane JF, Iams JD, Romero R.
Preterm birth 1: Epidemiology and causes of preterm
birth. The Lancet. 2008;371(9606):75-84.
8. Copper RL, Goldenberg RL, Das A, Elder N, Swain M,
Norman G, et al. The preterm prediction study:
maternal stress is associated with spontaneous
preterm birth at less than thirty-five weeks
gestation. American Journal of Obstetrics and Gynecology. 1996;175(5):1286-1292.
9. Fiscella K. Race, perinatal outcome, and amniotic infection. Obstetrical & Gynecological Survey.
1996;51(1):60-66.
10. Goldenberg RL, Cliver SP, Mulvihill FX, Hickey CA,
Hoffman HJ, Klerman LV, et al. Medical, psychosocial, and behavioral risk factors do not explain the
increased risk for low birth weight among black
women. American Journal of Obstetrics and
Gynecology. 1996;175(5):1317-1324.
11. Neggers Y, Goldenberg RL. Some thoughts on body
mass index, micronutrient intakes and pregnancy outcome. Journal of Nutrition. 2003;133(5):1737S-1740S.
12. Conde-Agudelo A, Rosas-Bermdez A, KafuryGoeta AC. Birth spacing and risk of adverse
perinatal outcomes: a meta-analysis. JAMA.
2006;295(15):1809-1823.
13. Brett KM, Strogatz DS, Savitz DA. Employment,
job strain, and preterm delivery among women in
North Carolina. American Journal of Public Health.
1997;87(2):199-204.
14. Thompson J, Irgens LM, Rasmussen S, Daltveit AK.

Secular trends in socio economic status and the implications for preterm birth. Paediatric and Perinatal
Epidemiology. 2006;20(3):182-187.
15. Andres RL, Day MC. Perinatal complications
associated with maternal tobacco use. Seminars
in Neonatology. 2000;5(3):231-241.
16. Cnattingius S. The epidemiology of smoking during
pregnancy: smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine & Tobacco
Research. 2004;6(Suppl 2):S125-140.
17. Mullally A, Cleary BJ, Barry J, Fahey TP, Murphy DJ.
Prevalence, predictors and perinatal outcomes of
peri-conceptional alcohol exposure - retrospective
cohort study in an urban obstetric population in
Ireland. BMC Pregnancy and Childbirth. 2011;11(1):27.
18. Jackson RA, Gibson KA, Wu YW, Croughan MS.
Perinatal outcomes in singletons following in vitro
fertilization: A meta-analysis. Obstetrics &
Gynecology. 2004;103(3):551-563.
19. Malloy MH, Hoffman HJ. Prematurity, sudden
infant death syndrome, and age of death. Pediatrics.
1995;96(3):464-471.
20. Robertson CMT, Howarth TM, Bork DLR, Dinu IA.
Permanent bilateral sensory and neural hearing loss
of children after neonatal intensive care because of
extreme prematurity: A thirty-year study. Pediatrics.
2009;123(5):e797-e807.
21. Hille E, Weisglas-Kuperus N, Van Goudoever JB,
Jacobusse GW, Ens-Dokkum MH, de Groot L, et
al. Functional outcomes and participation in young
adulthood for very preterm and very low birth weight
infants: The Dutch project on preterm and small for
gestational age infants at 19 years of age. Pediatrics.
2007;120(3):e587-595.
22. Frank L. Developmental aspects of experimental
pulmonary oxygen toxicity. Free Radical Biology and
Medicine. 1991;11(5):463-494.
23. Maltepe E, Saugstad OD. Oxygen in health and
disease: regulation of oxygen homeostasis-clinical
implications. Pediatric Research. 2009;65(3):261-268.
24. Vento M. Titrating oxygen needs in the very preterm
newborn in the delivery room. Journal of NeonatalPerinatal Medicine. 2010;3(3):161-169.
25. Friel JK, Friesen RW, Harding SV, Roberts LJ. Evidence
of oxidative stress in full-term healthy infants.
02-08-12 11:36
26. Saugstad OD. Oxidative stress in the newborna 30-year perspective. Biology of the Neonate.
2005;88(3):228-236.
27. Saugstad OD. Oxygen and oxidative stress in
bronchopulmonary dysplasia. Journal of Perinatal
Medicine. 2010;38(6):571-577.
28. Buonocore G, Perrone S, Longini M, Vezzosi P,
Marzocchi B, Paffetti P, et al. Oxidative stress in
preterm neonates at birth and on the seventh day
of life. Pediatric Research. 2002;52(1):46-49.
29. Saugstad OD. Hypoxanthine as an indicator of hypoxia:
Its role in health and disease through free radical
production. Pediatric Research. 1988;23(2):143-150.
30. Saugstad OD. Oxygen radical disease in neonatology.
Seminars in Neonatology. 1998;3(3):229-238.
31. Saugstad OD. The oxygen radical disease of neonatology. Indian J Paediatrics. 1989;56:585-593.
32. Baraldi E, Carraro S, Filippone M. Bronchopulmonary
dysplasia: Definitions and long-term respiratory outcome. Early Human Development. 2009;85(10):S1-3.
33. Northway Jr WH, Rosan RC, Porter DY. Pulmonary
disease following respirator therapy of hyaline-membrane disease. Bronchopulmonary dysplasia. New
England Journal of Medicine. 1967;276(7):357-368.
34. Miller NE. Techniques of early respiratory management of very low and extremely low birth weight
infants. Neonatal Netw. 2010;29(3):153-160.
35. Fanaroff AA, Stoll BJ, Wright LL, Carlo WA,
Ehrenkranz RA, Stark AR, et al. Trends in neonatal
morbidity and mortality for very low birthweight
infants. American Journal of Obstetrics and Gynecology. 2007;196(2):147.e141-e148.
36. Donn SM, Sinha SK. Invasive and noninvasive
neonatal mechanical ventilation. Respiratory Care.
2003;48(4):426-439.
37. Jobe AJ. The new BPD: an arrest of lung development. Pediatric Research. 1999;46(6):641-643.
38. Schroeder P, Gortner L. Predicting bronchopulmonary
dysplasia. Intensive Care Medicine. 1999;25(2):241.
39. Saugstad OD. Chronic lung disease: Oxygen dogma
revisited. Acta Paediatr. 2001;90(2):113-115.
40. Rodriguez RJ. Management of respiratory distress syndrome: An update. Respiratory Care.
2003;48(3):279-287.
41. Smith G. The pharmacology of the ductus arteriosus.
Pharmacological Reviews. 1998;50(1):35-58.
27
42. Koch J, Hensley G, Roy L, Brown S, Ramaciotti C,

Rosenfeld CR. Prevalence of spontaneous closure of
the ductus arteriosus in neonates at a birth weight of
1000 grams or less. Pediatrics. 2006;117(4):1113-1121.
43. Hagan R, Minutillo C, French N, Reese A, Landau L,
LeSouef P. Neonatal chronic lung disease, oxygen
dependency, and a family history of asthma. Pediatric
Pulmonology. 1995;20(5):277-283.
44. Clyman R, Cassady G, Kirklin JK, Collins M, Philips
III JB. The role of patent ductus arteriosus ligation
in bronchopulmonary dysplasia: Reexamining a
randomized controlled trial. Journal of Pediatrics.
2009;154(6):873-876.
45. Hellstrom A, Perruzzi C, Ju M, Engstrom E, Hard AL,
Liu JL, et al. Low IGF-I suppresses VEGF-survival
signaling in retinal endothelial cells: Direct correlation
with clinical retinopathy of prematurity. Proc Natl
Acad Sci USA. 2001;98(10):5804-5808.
46. Cunningham S, Fleck BW, Elton RA, McIntosh N.
Transcutaneous oxygen levels in retinopathy of
prematurity. The Lancet. 1995;346(8988):1464-1465.
47. Steinkuller PG, Du L, Gilbert C, Foster A, Collins ML,
Coats DK. Childhood blindness. Journal of AAPOS.
1999;3(1):26-32.
48. Jegatheesan P, Ianus V, Buchh B, Yoon G, Chorne N,
Ewig A, et al. Increased indomethacin dosing for persistent patent ductus arteriosus in preterm Infants:
A multicenter, randomized, controlled trial.
Journal of Pediatrics. 2008;153(2):183-189.
49. Rekha S, Battu RR. Retinopathy of prematurity: incidence and risk factors. Indian Pediatrics.
1996;33(12):999-1003.
50. Holmstrm G, Broberger U, Thomassen P.
Neonatal risk factors for retinopathy of prematurity
a population based study. Acta Ophthalmologica
Scandinavica. 1998;76(2):204-207.
51. Liu PM, Fang PC, Huang CB, Kou HK, Chung MY,
Yang YH, et al. Risk factors of retinopathy of prematurity in premature infants weighing less than 1600 g.
American J Perinatol. 2005;22(2):115-120.
52. Tehrani FT, Bazar AR. A feedback controller for
supplemental oxygen treatment of newborn infants:
A simulation study. Medical Engineering and Physics.
1994;16(4):329-333.
02-08-12 11:36
28
CHAPTER 1
53. Seiberth V, Linderkamp O, Seiberth P, fr Augenheilkunde K, Osnabrck M. Risk factors in retinopathy of prematurity; A multivariate statistical analysis.
Ophthalmologica. 2000;214(2):131-135.
54. Tin W, Milligan DWA, Pennefather P, Hey E. Pulse
oximetry, severe retinopathy, and outcome at one
year in babies of less than 28 weeks gestation. British
Medical Journal. 2001;84(2):F106-110.
55. Chen M, itil A, McCabe F, Leicht KM, Fiascone J,
Dammann CEL, et al. Infection, oxygen, and immaturity: interacting risk factors for retinopathy of
prematurity. Neonatology. 2010;99(2):125-132.
56. Kotanchek ME, Helferty JP, Murray WB, Palmer C.
Application of data fusion to neonate oxygenation
control. In: Multisensor Fusion and Integration for
Intelligent Systems IEEE International Conference on
MFI; 1994; 1994. p. 509-514.
57. Anderson CG, Benitz WE, Madan A. Retinopathy
of prematurity and pulse oximetry: A national survey
of recent practices. J Perinatol. 2004;24:164-168.
58. Brooks SE, Marcus DM, Gillis D, Pirie E, Johnson
CMH, Bhatia J. The effect of blood transfusion protocol on retinopathy of prematurity: A prospective,
randomized study. Pediatrics. 1999;104(3):514-518.
59. Filho JBF, Bonomo PP, Maia M, Procianoy RS.
Weight gain measured at 6 weeks after birth as a
predictor for severe retinopathy of prematurity:
Study with 317 very low birth weight preterm babies.
Graefes Archive for Clinical and Experimental
Ophthalmology. 2009;247(6):831-836.
60. Hellstrom A, Hard AL, Engstrom E, Niklasson A,
Andersson E, Smith L, et al. Early weight gain
predicts retinopathy in preterm infants: new,
simple, efficient approach to screening. Pediatrics.
2009;123(4):e638-645.
61. Gunn TR, Easdown J, Outerbridge EW, Aranda JV.
Risk factors in retrolental fibroplasia. Pediatrics.
1980;65(6):1096-1100.
62. Sacks LM, Schaffer DB, Anday EK, Peckham GJ,
Delivoria-Papadopoulos M. Retrolental fibroplasia
and blood transfusion in very low-birth-weight
infants. Pediatrics. 1981;68(6):770-774.
63. Shohat M, Reisner SH, Krikler R, Nissenkorn I, Yassur
Y, Ben-Sira I. Retinopathy of prematurity: Incidence
and risk factors. Pediatrics. 1983;72(2):159-163.
64. Hagadorn JI, Richardson DK, Schmid CH, Cole CH.
Cumulative illness severity and progression from
moderate to severe retinopathy of prematurity.
J Perinatol. 2007;27(8):502-509.
65. Palmer EA, Flynn JT, Hardy RJ, Phelps DL, Phillips
CL, Schaffer DB, et al. Incidence and early course
of retinopathy of prematurity. The Cryotherapy for
Retinopathy of Prematurity Cooperative Group.
Ophthalmology 1991;98(11):1628-40.
66. Hussain N, Clive J, Bhandari V. Current incidence of
retinopathy of prematurity, 1989-1997. Pediatrics.
1999;104(3):e26.
67. Gilbert C, Rahi J, Eckstein M, OSullivan J, Foster A.
Retinopathy of prematurity in middle-income
countries. The Lancet. 1997;350(9070):12-14.
68. Terry TL. Fibroblastic overgrowth of persistent
tunica vasculosa lentis in infants born prematurely: II.
Report of cases - Clinical aspects. Transactions of the
American Ophthalmological Society. 1942;40:262-284.
69. Munoz B, West SK. Blindness and visual impairment
in the Americas and the Caribbean. British Journal of
Ophthalmology. 2002;86(5):498-504.
70. Sanchez ME, Andrews BJ, Karr D, Lansingh V,
Winthrop KL. The emergence of retinopathy of
prematurity in guatemala. Journal of pediatric ophthalmology and strabismus. 2010;47:e1-4.
71. Gilbert C, Fielder A, Gordillo L, Quinn G, Semiglia
R, Visintin P, et al. Characteristics of infants with
severe retinopathy of prematurity in countries with
low, moderate, and high levels of development:
implications for screening programs. Pediatrics.
2005;115(5):518-525.
72. Palmer EA, Flynn JT, Hardy RJ, Phelps DL, Phillips CL,
Schaffer DB, et al. Incidence and early course
of retinopathy of prematurity. The cryotherapy for
retinopathy of prematurity cooperative group.
Ophthalmology. 1991;98(11):1628-1640.
73. Good WV, Hardy RJ, Dobson V, Palmer EA, Phelps
DL, Quintos M, et al. The incidence and course of
retinopathy of prematurity: findings from the early
treatment for retinopathy of prematurity study.
Pediatrics. 2005;116(1):15-23.
74. Clark D, Mandal K. Treatment of retinopathy of prematurity. Early human development. 2008;84(2):95-99.
75. Silverman WA. A cautionary tale about supplemental
oxygen: The albatross of neonatal medicine.
Pediatrics. 2004;113(2):394-396.
76. Dennery PA, Phyllis AD. Oxygen administration in
the care of neonates: A double-edged sword. Chinese
Medical Journal. 2010;123(20):2938-2942.
77. Tin W. Oxygen therapy: 50 years of uncertainty.
Pediatrics. 2002;110(3):615-616.
02-08-12 11:36
78. Sternbach GL, Varon J. The discovery and rediscovery of oxygen. Journal of Emergency Medicine.
2005;28(2):221-224.
79. Campbell K. Intensive oxygen therapy as a possible
cause of retrolental fibroplasia; a clinical approach.
The Medical Journal of Australia. 1951;2(2):48-50.
80. Patz A, Hoeck LE, De La Cruz E. Studies on the
effect of high oxygen administration in retrolental
fibroplasia. I. Nursery observations. American Journal
of Ophthalmology. 1952;35(9):1248-1253.
81. Crosse V, Evans PJ. Prevention of retrolental fibroplasia. Archives of Ophthalmology. 1952;48(1):83-87.
82. Guy LP, Lanman JT, Dancis J. The possibility of total
elimination of retrolental fibroplasia by oxygen
restriction. Pediatrics. 1956;17(2):247-249.
83. Kinsey VE, Jacobus JT, Hemphill F. Retrolental fibroplasia: Cooperative study of retrolental fibroplasia
and the use of oxygen. Archives of Pediatrics and
Adolescent Medicine. 1956;92(4):395-397.
84. Alberman ED. Epidemiology of retinopathy of
prematurity. Contemporary Issues in Fetal and
Neonutal Medicine. 1985;2:249-266.
85. Avery M, Oppenheimer E. Recent increase in
mortality from byaline membrane disease*.
86. McDonald AD. The aetiology of spastic diplegia
a synthesis of epidemiological and pathological
evidence. Developmental Medicine & Child
Neurology. 1964;6(3):277-285.
87. Bolton DP, Cross KW. Further observations on cost
of preventing retrolental fibroplasia. The Lancet.
1974;303(7855):445-448.
88. Cross KW. Cost of preventing retrolental fibroplasia?
The Lancet. 1973;302(7835):954-956.
89. Usher RH. Clinical investigation of the respiratory
distress syndrome of prematurity. Interim report. New
York State Journal of Medicine. 1961;61:1677-1696.
90. Millam D. The history of intravenous therapy. Journal
of Infusion Nursing. 1996;19(1):5-14.
91. Rivera AM, Strauss KW, Van Zundert A, Mortier E.
The history of peripheral intravenous catheters:
How little plastic tubes revolutionized medicine.
Acta Anaesthesiologica Belgica. 2005;56(3):271-282.
92. Cosnett JE. Origins of intravenous fluid therapy.
The Lancet. 1989;1(8641):768-771.
93. Schulze KF, Graff M, Schimmel MS, Schenkman A,
Rohan P. Physiologic oscillations produced by an infusion pump. Journal of Pediatrics. 1983;103(5):796-798.
29
94. Klem S, Farrington J, Leff R. Influence of infusion

pump operation and flow rate on hemodynamic
stability during epinephrine infusion. Critical Care
Medicine. 1993;21(8):1213-1217.
95. Kern H, Kuring A, Redlich U, Dopfmer U, Sims N,
Spies C, et al. Downward movement of syringe
pumps reduces syringe output. British Journal of
Anaesthesia. 2001;86(6):828-831.
96. Evans A, Winslow E. Oxygen saturation and hemodynamic response in critically ill, mechanically ventilated
adults during intrahospital transport. American
Journal of Critical Care. 1995;4(2):106-111.
97. Hurlbut JC, Thompson S, Reed MD, Blumer JL,
Erenberg A, Leff RD. Influence of infusion pumps on
the pharmacologic response to nitroprusside. Critical
Care Medicine. 1991;19(1):98-101.
98. Neal D, Lin J. The effect of syringe size on reliability
and safety of low-flow infusions. Pediatric Critical
Care Medicine. 2009;10(5):592-596.
99. Capes DF, Dunster KR, Sunderland VB, McMillan D,
Colditz PB, McDonald C. Fluctuations in syringepump infusions: association with blood pressure
variations in infants. American Journal of HealthSystem Pharmacy. 1995;52(15):1646-1653.
100. Weiss M, Neff T, Gerber A, Fischer J. Impact of infusion line compliance on syringe pump performance.
Pediatric Anesthesia. 2000;10(6):595-599.
101. McCarroll C, McAtamney D, Taylor R. Alteration in
flow delivery with antisyphon devices. Anaesthesia.
2000;55(4):355-357.
102. Dimaguila MAVT, Di Fiore JM, Martin RJ, Miller MJ.
Characteristics of hypoxemic episodes in very low
birth weight infants on ventilatory support.
103. Tehrani FT. Automatic control of mechanical
ventilation and the inspired fraction of oxygen in the
premature infant: A simulation study. In: Proceedings
of the First Joint BMES/EMBS Conference; 1999;
Atlanta, GA, USA 1999. p. 339.
104. Shiao SYPK. Desaturation events in neonates
during mechanical ventilation. Critical Care Nursing
Quarterly. 2002;24(4):14-29.
105. Claure N, Ozdamar O, Bancalari E. A system for
automatic adjustment of the inspired oxygen in
mechanically ventilated premature infants. In:
Engineering in Medicine and Biology Society,
Proceedings of the 25th Annual International
Conference of the IEEE; 2003; 2003. p. 454-455.
02-08-12 11:36
30
CHAPTER 1
106. Urschitz MS, Horn W, Seyfang A, Hallenberger A,

Herberts T, Miksch S, et al. Automatic control of the
inspired oxygen fraction in preterm infants: A randomized crossover trial. American Journal of Respiratory and Critical Care Medicine. 2004;170(10):10951100.
107. Laptook AR, Salhab W, Allen J, Saha S, Walsh M.
Pulse oximetry in very low birth weight infants:
can oxygen saturation be maintained in the desired
range? J Perinatol. 2006;26:337-341.
108. Lawless ST. Crying wolf: False alarms in a
pediatric intensive care unit. Critical Care Medicine.
1994;22(6):981-985.
109. Beddis IR, Collins P, Levy NM, Godfrey S, Silverman
M. New technique for servo-control of arterial
oxygen tension in preterm infants. Arch Dis Child.
1979;54(4):278-280.
110. Collins P, Levy NM, Beddis IR, Godfrey S, Silverman
M. Apparatus for the servocontrol of arterial oxygen
tension in preterm infants. Medical and Biological
Engineering and Computing. 1979;17(4):449-452.
111. Sano A, Kikucki M. Adaptive control of arterial oxygen
pressure of newborn infants under incubator oxygen
treatments. In: Control Theory and Applications, IEE
Proceedings D; 1985; 1985. p. 205-211.
112. Dugdale RE, Cameron RG, Tealman GT. Closed-loop
control of the partial pressure of arterial oxygen in
neonates. Clinical Physics and Physiological
Measurement. 1988;9(4):291-305.
113. Taube J, Bhutani V. Automatic control of neonatal
fractional inspired oxygen. In: Engineering in Medicine
and Biology Society Proceedings of the Annual
International Conference of the IEEE; 1991; 1991. p.
2176-2177.
114. Tehrani FT, Bazar AR. An automatic control system
for oxygen therapy of newborn infants. In: Proceedings of the Annual International Conference
of the IEEE Engineering in Medicine and Biology
Society; 1991; 1991. p. 2180-2182.
115. Bhutani VK, Taube JC, Antunes MJ, DelivoriaPapadopoulos M. Adaptive control of inspired oxygen delivery to the neonate. Pediatric Pulmonology.
1992;14(2):110-117.
116. Morozoff PE, Evans RW, Smyth JA. Automatic control
of blood oxygen saturation in premature infants. In:
Second IEEE Conference on Control Applications
1993; Vancouver, BC , Canada 1993. p. 415-419.
117. Yu C, He WG, So JM, Roy R, Kaufman H, Newell

JC. Improvement in arteral oxygen control using
multiple-model adaptive control procedures. In:
Biomedical Engineering, IEEE Transactions on; 2007;
2007. p. 567-564.
118. Claure N, Gerhardt T, Everett R, Musante G, Herrera
C, Bancalari E. Closed-loop controlled Inspired oxygen concentration for mechanically ventilated very
low birth weight Infants with frequent episodes of
hypoxemia. Am Acad Pediatrics. 2001;107(5):1120-1124.
119. Tehrani FT. A control system for oxygen therapy of
premature infants. In: Proceedings of the 23rd Annual
International Conference of the IEEE Engineering in
Medicine and Biology Society; 2001; 2001.
p. 2059-2062.
120. Tehrani F, Rogers M, Lo T, Malinowski T, Afuwape S,
Lum M, et al. A dual closed-loop control system for
mechanical ventilation. Journal of Clinical Monitoring
and Computing. 2004;18(2):111-129.
121. Morozoff PE, Evans RW. Closed-loop control of
SaO2 in the neonate. Biomed Instrum Technol.
1992;26(2):117-127.
122. Sun Y, Kohane IS, Stark AR. Computer-assisted
adjustment of inspired oxygen concentration
improves control of oxygen saturation in newborn
infants requiring mechanical ventilation. Journal of
Pediatrics. 1997;131(5):754-756.
123. Morozoff EP, Smyth JA. Evaluation of three automatic
oxygen therapy control algorithms on ventilated low
birth weight neonates. In: 31st Annual International
Conference of the IEEE, EMBS.
Minneapolis, Minnesota, USA; 2009. p. 3079-3082.
124. Claure N, DUgard C, Bancalari E. Automated adjustment of inspired oxygen in preterm infants with
frequent fluctuations in oxygenation: A pilot clinical
trial. Journal of Pediatrics. 2009;155(5):640-645.
125. Claure N, Bancalari E, DUgard C, Nelin L, Stein M,
Ramanathan R, et al. Multicenter crossover study of
automated control of inspired oxygen in ventilated
preterm infants. Pediatrics. 2011;127(1):e76-83.
126. Zhang L, Cameron RG. A real-time expert control
strategy for blood gas management inneonates
under ventilation treatment. In: IEEE Colloquium on
exploiting the knowledge base: Applications of rule
based control; 1989; London, UK; 1989. p. 4/1-4/7.
02-08-12 11:36
31
127. Azhar N, Karim U. Automatic feedback control of

oxygen therapy using pulse oximetry. In: Engineering
in Medicine and Biology Society, Proceedings of the
Annual International Conference of the IEEE; 1991;
1991. p. 1614-1615.
128. Sun Y, Kohane I, Stark AR. Fuzzy logic assisted control
of inspired oxygen in ventilated newborn infants. In:
Proceedings of the Annual Symposium on Computer
Application in Medical Care; 1994; 1994. p. 757-761.
129. ODonnell CP. Automated adjustment of oxygen in
ventilated preterm infants: turn on, tune in, ROP out?
130. Wilkinson DJ, Andersen CC. Bedside algorithms
for managing desaturation in ventilated preterm
infants: A randomised crossover trial. Neonatology.
2008;95(4):306-310.
131. Chow LC, Wright KW, Sola A. Can changes in clinical
practice decrease the incidence of severe retinopathy of prematurity in very low birth weight infants?
Pediatrics. 2003;111(2):339-345.
132. Stowe CD, Storgion SA, Lee KR, Phelps SJ. Hemodynamic response to intentionally altered flow continuity of dobutamine and dopamine by an infusion
pump in infants. Pharmacotherapy. 1996;16(6):10181023.
133. Cunningham S, Deere S, McIntosh N. Cyclical
variation of blood pressure and heart rate in
neonates. Arch Dis Child. 1993;69(1 Spec No):64-67.
134. Rooke GA, Bowdle A. Syringe pumps for
infusion of vasoactive drugs. Anesthesia & Analgesia.
1994;78(1):150-156.
135. Lnnqvist PA, Lfqvist B. Design flaw can convert
commercially available continuous syringe pumps
to intermittent bolus injectors. Intensive Care
Medicine. 1997;23(9):998-1001.
136. Southern DA, Read MS. Lesson of the Week: Overdosage of opiate from patient controlled analgesia
devices. British Medical Journal. 1994;309(6960):1002.
137. Levi DS, Peterson N, Shah SD, Rakholia B, Haught A,
Carman G. Connecting multiple low-flow intravenous
infusions in the newborn: Problems and possible solutions. Pediatric Critical Care Medicine. 2010;11(2):275281.
138. Weiss M, Fischer J, Neff T, Schulz G, Bnziger O. Do
antisiphon valves reduce flow irregularities during
vertical displacement of infusion pump systems?
Anaesthesia and Intensive Care. 2000;28(6):680-683.
02-08-12 11:36
02-08-12 11:36
Part 1
supplemental
oxygen therapy
02-08-12 11:36
34
02-08-12 11:36
CHAPTER 2
Oxygenation in
preterm infants:
background, target
ranges & monitoring
techniques
A.C. van der Eijk, J. Dankelman, B.J. Smit
02-08-12 11:36
36
CHAPTER 2
It is inherent to their immaturity that preterm infants need some form of physiological
monitoring during their stay on the neonatal intensive care unit. In current neonatal
intensive care at least the heart rate, respiration rate, skin temperature, and oxygen
saturation are monitored continuously. In the past 60 years, various methods to monitor
blood oxygen levels have been developed. In this chapter, background information about
the oxygenation of the human body, and especially that of preterm infants, is provided.
Subsequently, the working principles and (dis)advantages of methods that are available
for monitoring oxygenation are elaborated on. Thereafter, an overview of literature on
target ranges for blood oxygen levels in (preterm) newborn infants is provided. Finally, a
future perspective of the needs for oxygen monitoring in preterm infants is discussed.
OBJECTIVE To obtain background information regarding oxygenation of the human
body, to get an overview of literature on target ranges for blood oxygen levels in newborn infants, and to evaluate methods for monitoring oxygenation in neonatology.
02-08-12 11:36
Oxygenation in preterm infants
37
2.1 INTRODUCTION
Approximately five to ten percent of all newborn infants need active resuscitation immediately after birth to survive.1 The majority of these infants is born preterm. Due to the immaturity of organs like the lungs and brain of these preterm infants, respiratory support and
supplemental oxygen therapy are necessary to reach and maintain adequate oxygenation.
In supplemental oxygen therapy, a gas mixture with >21% of oxygen is supplied to the
patient via mechanical ventilation. In preterm infants this therapy is not only used immediately after birth, but also in the first weeks after birth. Unfortunately, supplemental
oxygen therapy is not without risks. Both too low and too high levels of oxygen in the
tissues may have severe consequences for the development and outcome of preterm
infants.2 Therefore, to prevent the negative effects of supplemental oxygen therapy, blood
oxygen levels of preterm infants need to be monitored closely during their hospitalization
on the neonatal intensive care unit (NICU).
In the last decades, several methods and sensors to monitor oxygenation in preterm infants
have been developed. Regrettably, these monitoring systems did not always serve the specific needs for the patients, the family, and/or the healthcare professionals in the optimal
way. The aims of this chapter are to provide background information about oxygenation of
the human body, to provide an overview of the techniques (that were) available for monitoring oxygenation in preterm infants, and to discuss target ranges for blood oxygen levels
in (preterm) newborn infants. Finally, the needs and future perspectives for monitoring
oxygenation in (preterm) newborn infants are discussed.
To prevent the negative effects of supplemental

oxygen therapy, blood oxygen levels of
preterm infants need to be monitored closely.
02-08-12 11:36
38
CHAPTER 2
2.2 Oxygen in the human body

Cells in the human body can only function properly when the amount of oxygen delivered
to the cells is adequate to meet the demands of the cells.3 Because oxygen cannot be held
in stock in the cells, a constant delivery of oxygen is needed.4 In literature, the vaguely defined term oxygenation is often used as a measure for the amount of oxygen in the body.
However, an adequate oxygenation depends on the systemic oxygen transport (delivery),
the oxygen consumption (demand), and the mixed venous saturation (reserve).5 The oxygen
delivery, the demand, and the reserve depend on, and are maintained by, multiple parameters. Thus, when examining whether the oxygenation is adequate or not, interpretation
of multiple parameters is required.
In humans, the inhaled oxygen diffuses in the lungs from the alveoli to the pulmonary capillary blood (Figure 2.1). Most of the diffused oxygen is bound to haemoglobin, an intracellular protein, within the erythrocytes (i.e. red blood cells). A small proportion of oxygen is
dissolved in blood plasma. The sum of the amount of oxygen bound to haemoglobin and
the oxygen dissolved in the plasma is the oxygen content of blood (O2ct) (equation I).6
O2ct = (k1 Hb SO2) + (k2 PO2)
Where
Hb
SO2
PO2
k1
k2
(I)
= haemoglobin concentration (grams litre1)

= oxygen saturation of the blood (see 2.2.1)
= partial pressure of oxygen (see 2.2.1)
= Hfners constant (in theory, each gram of Hb binds 1.39 ml of oxygen,
in practice it is less)
= solubility coefficient of oxygen at body temperature (0.23 ml litre-1 kPa-1)
The oxygen rich erythrocytes are transported through the body via blood vessels to the
tissues. The oxygen delivery (DO2) is the amount of oxygen transported from the lungs to
the peripheral tissues, and depends on the oxygen content of arterial blood (aO2ct) and the
cardiac output (Q) (equation II en III)
DO2 = Q aO2ct
(II)
Q = heart rate stroke volume
(III)
At the places were oxygen is demanded, the oxygen dissociates from haemoglobin and diffuses into the cells. After oxygen is released, the oxygen-poor erythrocytes are transported
via blood vessels back to the heart and lungs. The difference in aO2ct and the oxygen content of the venous blood (vO2ct) is the amount of oxygen delivered to the tissues. The total
oxygen consumption (VO2) depends on Q and the amount of oxygen delivered to the tissues (see the Fick equation, IV).
VO2 = Q (aO2ct - vO2ct)
(IV)
02-08-12 11:36
39
Alveoli
Trachea
Alveoli
airspace
CO2
O2
Capillary
Left Lung
Red
blood cell
Figure 2.1 A schematic illustration of the lungs, the alveoli, and the diffusion of O2 and CO2
between the lungs and the blood. Figure adapted from [7].
2.2.1 The oxygen dissociation curve

As mentioned in the previous paragraph, oxygen binds to haemoglobin in the erythrocytes.
The percentage of haemoglobin bound to oxygen divided by the sum of all the available
haemoglobin is defined as oxygen saturation (SO2, equation V). For SO2 in the arterial
blood the abbreviation SaO2 is used. The actual amount of oxygen bound to haemoglobin
depends on the ability of haemoglobin to bind or release oxygen, i.e. the oxygen affinity.3
SO2 = [HbO2] / ([HbO2] + [Hb])
(V)
Where
[Hb]
= concentration of deoxyhemoglobin in blood (can bind to oxygen)
[HbO2] = concentration oxyhemoglobin of blood (bound to oxygen)
The oxygen affinity can be described by an S-shaped graph, the Oxygen Dissociation-curve
(OD-curve, Figure 2.2). The OD-curve represents the relation between the SO2 and the partial pressure of oxygen in the blood (PO2). The PO2, or the oxygen tension is the amount
of oxygen that has diffused across the alveolar capillary membrane and is dissolved in the
plasma of the blood. The PO2 is an important factor in the exchange of O2 and carbon dioxide (CO2) from the blood to the tissues and vice versa. In the lungs, high levels of oxygen
tension in arterial blood (PaO2) encourage oxygen to bind to haemoglobin.8, 9 As can be
seen in the OD-curve, at low PO2 levels the curve is steep. At these lower levels oxygen is
encouraged to perfuse into the tissues.
02-08-12 11:36
40
CHAPTER 2
The relative position of the OD-curve on the x-axis equals the oxygen affinity, and is influenced by several factors, amongst others the pH, CO2, 2,3-diphosphoglycerate (2,3-DPG),
temperature and type of haemoglobin. A change in the oxygen affinity influences both the
amount of oxygen that binds to the haemoglobin when passing the alveoli, and the amount
of oxygen that is released in the tissues. It has been shown that the OD-curve can shift very
quickly in preterm infants.10,11
Oxygen saturation (SO2) [%]
100
left shifted by:
pH
2.3-DPG
Temperature
75
right shifted by:

pH
2.3-DPG
Temperature
50
25
25
50
75
100
Partial pressure of oxygen (PO2) [mmHg]

(7.5 mmHg = 1 kPa)
Figure 2.2 A schematic illustration of the Oxygen Dissociation-curve (OD-curve). The OD-curve links
the oxygen saturation (SO2) to the partial pressure of oxygen in the blood (PO2). The relative position
of the curve on the x-axis equals the oxygen affinity, and is influenced by several factors. Graph is based
on [9].
Approximately 80% of all extremely low birth weight infants (1000 g.; ELBW) infants receive one or more blood transfusions during their first weeks of life. When a preterm infant
receives a blood transfusions, the transfused blood comes from human adults.12, 13 Most of
the haemoglobin present in the erythrocytes of healthy adults is adult haemoglobin (HbA).
In preterm infants, however, most of the haemoglobin in the blood is of a different type,
fetal haemoglobin (HbF). This HbF plays an important role in the oxygenation of the foetus. Compared to HbA, HbF binds 2,3-DPG poorly. 2,3-DPG is an organophosphate, which
is created in the erythrocytes during glycolysis. High levels of 2,3-DPG shift the curve to the
right, while low levels of 2,3-DPG cause a leftward shift. Thus, because HbF binds 2,3-DPG
02-08-12 11:36
41
poorly, the level of 2,3-DPG in the blood of foetuses is relatively low. Consequently, the ODcurve is shifted leftward compared to the HbA-rich blood of the mother. The leftward shift
of the OD-curve represents a higher oxygen affinity compared to the mother. The higher
oxygen affinity of the blood in foetuses accounts for a better transfer of oxygen in the blood
from the mother to the foetus and for optimal oxygenation of the foetus under the relatively hypoxic conditions in utero. Thus, in general, the higher the percentage of HbF in the
blood, the higher the SO2 level for a certain PO2 value, and vice versa.14, 15
In healthy newborns, HbF is broken down and replaced by HbA in three to six months after
birth. Nevertheless, this slow and regulated change of haemoglobin is not seen in preterm
infants in the NICU because of the blood transfusions with HbA. A transfusion with HbA
can speed up the natural process of haemoglobin change to only several hours, consequently causing a rapid right shift in the OD-curve (i.e. a decrease of oxygen affinity). Therefore,
it is advised to transfuse blood with HbA slowly, and to monitor blood oxygen levels closely
during and after the blood transfusion.16-19
To keep both SaO2 and PaO2 within a certain range, accurate monitoring is required. In the
next part of this chapter the (dis)advantages of methods developed to monitor oxygenation
in preterm infants are discussed.
2.3 Monitoring of oxygenation

The aim of monitoring was very well described by Murkovic:20 The primary aim of monitoring is to ensure that appropriate care or therapy can be given prior to the onset of complications. Thus, information obtained by monitoring equipment is important to alert for a
change in condition, and an aid for healthcare professionals in the decision making.
The monitoring of preterm infants is challenging.20 Preterm infants lie in an incubator
with an air temperature around 35C and with a high humidity. Furthermore, the sensors
used for monitoring have to be small and user-friendly for the patient, their family and the
healthcare professionals. In neonatal care, this implicates that the sensors should, at least,
not hamper the development of the preterm infant.
Before the techniques for continuous monitoring became available, healthcare professionals
could only use physical assessment of the skin of the patient to determine their oxygenation.21
In the recent decades, several methods and sensors have been developed, with varying techniques and functions to determine whether the total oxygen delivery achieves the total oxygen demand in preterm infants. These techniques are discussed in the following paragraphs.
02-08-12 11:36
42
CHAPTER 2
2.3.1 Physical assessment

For healthcare professionals, changes in the skin colour of patients have always been an
important indication for changes in the oxygenation. Nowadays, this method is still used
to determine oxygen requirement in preterm infants. However, physical assessment of the
skin is a subjective method, and not very reliable in detecting both hypoxia and hyperoxia
in preterm infants.22-24 Therefore, the possibility to perform blood gas analysis was a huge
improvement in the assessment of oxygen requirement.
2.3.2 Blood gas analysis
In blood gas analysis, a small sample of blood, either arterial, venous or capillary (mixed)
blood, is taken from the patient. This blood sample is analysed to determine the adequacy
of ventilation, pulmonary gas exchange, and the acid-base status of the patient.25 Blood
gas analysis from arterial blood provides multiple parameters like the SaO2, PaO2, arterial
carbon dioxide tension (PaCO2), pH, base excess, and fractions of different types of haemoglobin. Until the 1970s blood gas analysis was the only reliable method to determine
oxygen levels in the blood. Today, blood gas analysis is still the golden standard in medical
practice.
Unfortunately blood gas analysis is not without drawbacks. Firstly, there is the loss of blood
each time a sample is needed (0.3 ml). In preterm infants each drop of blood is of high value
for the patient, thus the number of blood samples that is collected should be minimised.
Secondly, the percutaneous puncture can lead to agitation. Regrettably, the specially developed indwelling catheters to reduce percutaneous punctures are associated with vascular
complications and infection.26 The third disadvantage is the fact that the method provides
only information about the moment the sample was taken. Thus, due to the intermittent
sampling, fast fluctuations in oxygenation could be missed.27, 28 To overcome the lack of
information about oxygenation in between blood sampling, since the 1960s several groups
worked on the development of continuous measurement of blood oxygen levels.
2.3.3 Continuous intra-arterial blood gas monitoring
Continuous measurement of blood oxygen levels became possible because of the development of catheters for continuous intra-arterial blood gas monitoring (CIBM). These CIBM
systems were designed to determine SaO2, PaO2, PCO2, and/or pH continuously and invasively. Initially, the CIBM systems were thought to be able to give insight in the rapid
changes in blood oxygen levels, and reduce the need for blood sampling. These advantages
would enhance therapeutic decision making, and reduce the need for blood transfusions.
Consequently, this would presumably lead to a reduction in the risks of infection, in workload of the caregivers, and in hospital costs.
The presumed advantages of CIBM systems were tested in animals,29 adults,30 infants,31-37
and neonates.38-42 Various interesting review papers discuss the results of the developed
CIBM systems in detail.28, 43, 44 Two of the most recent developed devices are the Paratrend,
and the Neotrend (Figure 2.3).29, 33-35, 41
02-08-12 11:36
43
PO2
PCO2
0.50 mm
pH
Temperature
23 mm
Figure 2.3 Schematic drawing of the the Paratrend (Diametrics Medical, High Wycombe, UK). The sensors for pH, PaCO2, PaO2 and the temperature are shown. This system was commercially available in the
1990s. Drawing is based on [41].
The Paratrend consists of a hybrid probe (diameter 0.5 mm, length 4 cm) with thermocouple,
an electrochemical PO2 sensor, and absorbance sensors for pH and PCO2. The Neotrend is
based on the Paratrend, and especially designed for neonates.
Although the Paratrend, the Neotrend, and some other devices were commercially
available, none of them were widespread in clinical use. This was due to, amongst others,
the risk of infection, and the deposition of plasma proteins on the device. This deposition
leads to platelet activation, adhesion and thrombus formation.42, 45-49 Another problem was
the wall effect, a phenomenon where the PaO2 value suddenly drops. This drop is caused
by the fact that the sensor is touching the arterial wall and measures the gas values of the
tissue instead of the blood.28, 43, 45 Above that, the sensors of the CIBM systems need frequent
recalibration,37, 39 and are highly fragile.28, 43, 45
As far as we know, today, sensors for intra-arterial measurement are not used in daily care
in NICUs anymore. However, in the future, the development in materials, and the improvement and miniaturization of techniques may increase new possibilities for CIBM systems.
2.3.4 Transcutaneous oxygen measurement
Simultaneously with the development of CIBM systems, transcutaneous oxygen measurement became available. In 1956 L.C. Clark invented a polarographic membrane-covered
oxygen electrode.50 When this Clark electrode is positioned on the chest of a patient it can
measure the diffusion of oxygen through the skin into the sensor. Based on this diffusion
the transcutaneous partial pressure of oxygen (tcPO2) and the transcutaneous partial pressure of carbon dioxide (tcPCO2) can be determined.51 This technique was the first, and still
the only, possibility to monitor PO2 and PCO2 continuous and transcutaneous.
02-08-12 11:36
44
CHAPTER 2
In preterm infants, tcPO2 correlates well with the PaO2. Therefore transcutaneous oxygen
monitoring was rapidly accepted for routine use in neonatal intensive care.52 However, soon
after the introduction in clinical practice, it became clear that the accuracy of the measurement declined when PaO2 increased, or when skin perfusion was decreased.53, 54 An adequate
skin perfusion is required to obtain optimal diffusion of oxygen. To achieve adequate skin
perfusion, the skin of the preterm infant needs to be heated to approximately 43 - 44C. Unfortunately, heating up the weak skin of preterm infants can lead to erythema. To prevent
this side effect, the sensor needs to be repositioned and recalibrated every 3 to 4 hours. This
repositioning and recalibration causes a high workload for the nursing staff.52, 55, 56
The necessity for heating of the skin, the high workload of the technique, the relative large
and heavy sensor, and the development of a newer technique pulse oximetry are the reason
that transcutaneous oxygen monitoring is not used very often anymore in the NICU.56, 57
Pulse oximetry was invented in Japan in the 1970s by T. Aoyagi.58 The technique was first
introduced in perioperative care, but it soon expanded into (neonatal) intensive care units.
Pulse oximetry was adopted quickly, because it was continuous, non-invasive, easy to apply,
and did not require calibrating or heating of the skin.59, 60
Pulse oximetry estimates SaO2 noninvasively with the help of (at least) two rays of (near)
infrared light (wavelengths of ~660 nm and ~940 nm). A probe is placed around an extremity, e.g., a hand or foot, of a preterm infant. Figure 2.4 shows an example of an pulse
oximeter probe. From one side of the probe, the two rays of light are emitted through
the tissue. From the other side of the probe, a sensor measures the light that is not
absorbed by the tissue. The two rays of light are absorbed differently by oxygenated and
deoxygenated blood. The sensor of the probe measures the ratio of the absorption of the
two wavelengths.49, 60 The variation in the absorbance rate of light is described by the law
of Beer-Lambert law (equation VI).61 The intensity (I) of light traveling through a medium
decreases exponentially with distance:
I = I0 * e()cd
(VI)
Where:
I0 = intensity of light at the start
()
= extinction coefficient [L/mmol/cm]
c
= (constant) concentration of the absorbing substance of the medium [mmol/L]
d
= optical path length through the medium
The oxygen saturation of the blood is determined by the amount of light that is (not) absorbed by the extremity (Figure 2.5). Based on SaO2 values validated in test subjects, the
absorbance rate can be converted to the pulse oximetry saturation value SpO2.
02-08-12 11:36
45
LED
Sensor
Figure 2.4 This figure shows a disposable pulse oximeter probe especially designed for patients with a
weight <3kg. The probe is positioned around a hand or foot of the patient.
One
cardiac circle
Pulsatile
arterial blood
LIGHT ABSORBANCE
Non-pulsatile
arterial blood
Venous &
capillary blood
Other tissue
TIME
Figure 2.5 Schematic illustration of the absorbance of light through the extremity of the patient.
Drawing based on [60, 61].
02-08-12 11:36
46
CHAPTER 2
As mentioned in 2.2.1 the SaO2 and the PaO2 are related by the OD-curve. Ideally, the
SpO2 value could provide healthcare professionals PaO2 values continuously. However, due
to the horizontal shifting and the sigmoid shape of the OD-curve, for high values of SpO2
(>95%), the corresponding PaO2 can lie within a wide range. Consequently, pulse oximeters
are relatively poor in detecting hyperoxia (high levels of PaO2).62-64
To prevent too low as well as too high values for SaO2 and PaO2, pulse oximeters provide
acoustic alarms to warn physicians and nurses when the measured SpO2 is outside a certain
target range. One of the main disadvantages of pulse oximetry is the high rate of false, or
non critical, alarms.65 These alarms are present because pulse oximetry is influenced by,
amongst others, low blood perfusion, ambient light, skin pigmentation and poor sensor
placement. Furthermore, the majority of the alarms is caused by motion artefacts.60, 66-70 The
high (false) alarm rate can cause a burden on patients, family, and healthcare professionals,
and should therefore be reduced.71-73
To reduce motion artefacts, in the last two decades several brands have developed motionresistant pulse oximeters. The performance of both these new-generation pulse oximetry
devices and the conventional ones have been evaluated in preterm neonates (see Chapter 3).
Although used methods and conclusions varied between studies, the clinical performance of
new generation pulse oximeters is shown to be better than conventional pulse oximeters.24
2.3.6 Near Infrared Spectroscopy
In addition to pulse oximetry, there is another method for monitoring oxygenation based
on the absorbance of light. This method, Near InfraRed Spectroscopy (NIRS) is a technique
for continuous monitoring of tissue oxygenation and hemodynamics. In NIRS, two soft
plastic probes are attached to, for instance, the head of a preterm infant. A beam of light is
sent through tissue and (partly) absorbed and scattered. This light can penetrate up to 6 - 8
cm into the tissue. The light that is scattered back is measured by a sensor. The advantages
of NIRS are the non-invasiveness, the inexpensiveness, and the availability of continuous
and repeatable measurements.74-84
NIRS was used for the first time for clinical research in newborns in 1985.85 Today the
technique is not progressed to daily clinical practice yet, and still remains mainly used for
research purposes. Reasons for this situation have probably to do with difficulties with
quantification, precision, wide intersubject variability, sensitivity to movement artefacts
and external light.49, 50, 83, 86-89
2.3.7 Capnography
Besides using invasive monitoring techniques or measurement via the skin, it is also
possible to monitor the exhaled air. With this capnography method, the partial pressure
of the end-tidal CO2 (PetCO2) is measured. Because the PetCO2 correlates well with the
PaCO2, capnography can serve as a trend monitor for the adequacy of ventilation.90 Compared to transcutaneous monitoring with a Clark electrode (see 2.3.4), capnography has
some clear advantages. The response time is faster, and there is no risk for burns of the skin.
02-08-12 11:36
47
Capnography can also be used to detect apnoea, is helpful during weaning, and to identify incorrect tube placement or occlusions. Another advantage is the potential decrease of
blood loss and complications associated with arterial catheters.91-98
Capnography can be performed by sidestream or mainstream gas sampling. For sidestream
capnography, the correlation with PaCO2 is relatively low. For main stream capnography
several studies confirmed the reliability in preterm infants.70, 96-100 However, technical limitations still hamper the measurement of PetCO2. One of the main limitations in capnography in preterm neonates is the need for relative high sample flow-rates (~50 ml/min) to
maintain accuracy.95
2.4 Reference values for blood oxygen levels

Since the moment that supplemental oxygen therapy was associated with negative effects
on the outcome of patients (see 1.4), multiple studies tried to define optimal levels for SaO2
and PaO2 in (preterm) newborn infants. In the next paragraphs the literature on the target
range for blood oxygen levels is discussed.
2.4.1 Reference values for oxygen saturation
In utero, SaO2 values of foetuses are between 70 and 80%.101 After birth, in healthy neonates, SpO2 levels rise slowly to levels above 95% in the first 15 minutes. After this initial
rise, SpO2 levels stabilise.102-105
Levesque et al.106 studied SpO2 levels in 718 healthy newborn infants at admission, at 24
hours, and at discharge from their nursery ((mean SD) postmenstrual age (PMA) = 39.3
1.6 weeks, birth weight (BW) = 3370 550 g). They found a mean SpO2 value of 97.2% (2
SD = 94 - 100%), and concluded that mean SpO2 values increase to a small degree with increasing postnatal age (PNA).
Wilkinson et al.,107 measured the actual SaO2 levels continuously with an intra-arterial
catheter. The subjects were breathing spontaneously, lying quietly and undisturbed (preterm) infants (mean (range) PMA = 32 (27 - 39) weeks, BW = 1.84 (1.05 - 3.3 kg)). This study
demonstrated that SaO2 levels change rapidly and transient between 85 - 95%. Another
study, by Mok et al.108 showed that mean SpO2 levels in 28 awake preterm infants with
a PMA between 29 and 34 weeks without lung disease were 92.9% (range = 86.6 - 96.3%)
during the first months of life.
2.4.2 Reference values for partial pressure of oxygen
Just as the SaO2 level, the PaO2 levels in foetuses and (preterm) neonates are lower than in
adults. Where the normal levels in adults lie between 8 - 10.7 kPa (60 - 80 mmHg), the PaO2
in the umbilical vein of foetuses is about 4.27 kPa (32 mmHg). In the descending aorta, the
PaO2 is about 2.9 kPa (22 mmHg).109
02-08-12 11:36
48
CHAPTER 2
Various studies were performed to determine normal levels for PaO2 in (preterm) newborn
infants.37, 110-113 For example, Orzalesi et al.113 determined PaO2 levels in 40 neonates (PMA =
28 - 40 weeks, BW = 1260 - 2480 grams). They showed that approximately 24 hours after
birth, PaO2 lies somewhere between 9.3 - 10.6 kPa (70 - 80 mmHg). In the study from Mok
et al. already mentioned above,108 they showed that the mean (range) tcPO2 was 9 kPa (4.7
- 10.9 kPa) (67.5 (35 - 82) mmHg) during the first months of life. Conway et al.37 showed how
variable the PaO2 of a spontaneously breathing infant with a respiratory illness can be. In a
female preterm infant (1.867 gram, GA = 31 weeks, PNA = 4 days) the PaO2 level varied from
6 kPa (45 mmHg) to more than 17.3 kPa (130 mmHg) during a 20-minute period.
Today, a PaO2 range of 6.7 - 10 kPa (50 - 75 mmHg) is considered as normal in healthy neonates.114 In preterm infants, PaO2 levels below 6 kPa (<45 mmHg) are thought to be too low.
PaO2 values above 12 kPa (>90 mmHg) are considered as hyperoxia.62
2.4.3 Target ranges & outcome
Various publications studied the effect of SpO2 and/or tcPCO2 levels on differences in (long
term) outcome. In 1992, Flynn et al.115 studied the effect of tcPO2 levels on retinopathy of
prematurity (ROP, see 1.4.4) in 101 preterm infants (BW = 500 - 1300 g.). They showed that
ROP occurred more often when tcPO2 exceeded 10.7 kPa (80 mmHg ) in the first weeks of
life. Brouillette et al.116 state that PaO2 values between 6.65 - 10.7 kPa (50 - 80 mmHg) minimise the risk of ROP and lower the FiO2 and airway pressure requirements. Today it is also
suggested that not only the absolute level of PaO2 should be taken into consideration. The
variability of PaO2 is also associated with the incidence and severity of ROP.117-120
The first randomised trial focusing on target levels for SpO2 was the STOP-ROP trial.121 In
this study a target range of 89 to 94% was compared with a target range of 96 to 99% in 649
preterm infants (mean PMA = 24.5 weeks). In the 96 - 99% group, there was a significant
increase of oxygen dependency, use of diuretics, and hospitalisation at three months corrected age. There was no difference in mortality, growth or visual outcome. The authors
suggested that 89 - 94% is a better target range than 96 - 99%.
Since the STOP-ROP trial various other studies (observational, randomised trials,
and/or surveys) were performed.122-125 One study investigated whether individualised
limits for SpO2 are more effective than fixed limits. In this study, Gupta et al.126 recalculated
the shape of the OD-curve for each premature infant individually. They advised not to use
individualised alarm limits in clinical practice as this will cause some practical problems
(calculating, communicating, and confusion).
02-08-12 11:36
49
Recently, a study (BOOST II) was started to compare target ranges for SpO2 of 91 to 95% and
85 to 89%. This study was terminated early because the survival rates at a PMA of 36 weeks
was larger for preterm infants (born with a GA <28 weeks) in the target range of 91 to 95%
than for 85 to 89%.127 Currently, a prospective meta-analysis of all the neonatal oxygen trials
is performed. Until the results of this meta-analysis are reported, and longer-term data on
survival and morbidity are available, Stenson et al. suggested not to use a target range for
SpO2 of 85 to 89% in infants born with a GA <28 weeks.127
Based on the current evidence, the latest general advise is to aim the SpO2 about 85 to 93%
with alarm limits set 1 to 2% above and/or below these targets.59 It has been shown that both
low and high PaO2 occurs very rarely in neonates breathing supplemental oxygen when
their SpO2 values are 85 to 93%.128
2.5 Monitoring oxygenation in preterm infants:

future perspectives
Based on the previous paragraphs about the working principles and (dis)advantages of
methods for monitoring oxygenation in preterm neonates, it is clear that none of these
monitoring techniques are optimal. Current available techniques are not accurate enough,
cumbersome, and/or not user- and patient-friendly. Above that, the current techniques
for continuous monitoring, like transcutaneous measurements and pulse oximetry, provide information about oxygen delivery only: the oxygen demand and reserve remain unknown.3, 129 Consequently, techniques currently used in clinical practice cannot provide a
complete overview of the oxygenation. Hirschl et al.130 even state that the evaluation of the
overall physiologic effect of mechanical ventilation is hampered because of the inability to
evaluate the cardiac output and oxygen consumption.
An ideal monitoring technique for oxygenation produces real-time, accurate, quantifiable
and reproducible data about the oxygen delivery and consumption for specific tissues. Furthermore, the technique should at least make a distinction between arterial and venous
blood.131, 132 However, before manufacturers and researchers start working on the development of new and improved techniques for monitoring oxygenation and other parameters
in (preterm) newborn infants, it should be clear which parameters are actually necessary
to monitor. Simultaneously, it should be questioned whether monitoring more and more
parameters will actually improve care. Certainly, more parameters will provide healthcare
professionals with more information, but more information may hamper the interpretation, and may confuse caregivers. Consequently, the aim of monitoring, to ensure that
appropriate care or therapy can be given prior to the onset of complications,20 may be missed.
Hence, monitoring multiple parameters is only useful when the interpretation of the
parameters is easy and unambiguous.
02-08-12 11:36
50
CHAPTER 2
In many medical devices the user can decide how he wants the information presented. Yet,
the user is often not the expert in defining the optimal user-interface. Therefore, we suggest
that it is the task of the researchers and manufacturers to develop systems that are easy to
use and interpret. They should define and advise in which form the presented information
is the most useful for healthcare professionals. Secondly, manufacturers should maintain
their focus on the reduction of false or non critical alarms.
From a patient point of view, improvements in neonatal monitoring should focus on the development of techniques and sensors which are not noticed by the patient or even by their
family. Ideally, the patient and their family should experience the hospital environment as
a friendly place, comparable with the home situation. This invisible monitoring may also
help to protect the fragile skin of the patients, and the effectiveness of phototherapy. In
phototherapy, the skin of patients is illuminated by fluorescent light to reduce bilirubin
levels in the blood. The larger the surface that can be illuminated, the more effective the
therapy can be. Because of the fragile skin, and the need for phototherapy, the adhesiveness
and the total area of the skin covered by the sensors to monitor the patient should be as
small as possible.
Although at first sight it would be logical to develop non-invasive monitoring techniques
for preterm infants, it may be interesting to focus on combining monitoring techniques
with other devices that are present with the preterm infant anyway, like the nasogastric
feeding tube, the catheter for supply of medication, the diaper, the blanket, or the tube
of the mechanical ventilator. Kyriacou et al.133 for instance, performed a pilot study to the
suitability of the oesophagus as a new measuring site for SpO2 in infants and neonates.
They concluded that the oesophagus can be used for monitoring SpO2.
In conclusion, to obtain a complete view of the oxygenation of the human body, it is
necessary to obtain information about multiple parameters. However, especially in neonatal intensive care, it is often not possible to measure the required parameters in a
reliable, and user-friendly way. Current available techniques for monitoring oxygenation in
preterm neonates are suboptimal. Improvements in monitoring techniques should focus
on continuous, reliable, accurate, real-time, and user-friendly methods, keeping in mind
the special and vulnerable group of patients hospitalised in a NICU. Focussing on invisible
monitoring may lead to techniques that do not hamper the development of the patient.
02-08-12 11:36
51
2.6 References
1. Christenson M. International population reports
WP/02, global population profile: 2002.
In: Bureau UC, editor. Washington DC:
US Government Printing Office; 2004.
2. Askie L, Henderson-Smart DJ, Ko H. Cochrane
review: Restricted versus liberal oxygen exposure for
preventing morbidity and mortality in preterm or low
birth weight infants. Evidence-Based Child Health:
A Cochrane Review Journal. 2010;5(1):371-413.
3. Dudell G, Cornish JD, Bartlett RH. What constitutes
adequate oxygenation? Pediatrics. 1990;85(1):39-41.
4. Huang YCT. Monitoring oxygen delivery in the
critically ill. Chest. 2005;128(5 suppl 2):554S-560S.
5. Whyte RK. Mixed venous oxygen saturation in
the newborn. Can we and should we measure it?
Scandinavian Journal of Clinical & Laboratory
Investigation. 1990;50(S203):203-211.
6. McLellan S, Walsh T. Oxygen delivery and
haemoglobin. Continuing Education in Anaesthesia,
Critical Care & Pain. 2004;4(4):123-126.
7. www.patient.co.uk. Published 2012.
8. Wyka KA, Mathews PJ, Clark WF. Foundations of
Respiratory Care. Melbourne: Delmar Thomson
Learning; 2002.
9. Widnaier EP, Raff H, Strang KT. Vander, Sherman and
Lucianos Human Physiology: The Mechanisms of
Body Function. ninth ed. Boston: McGraw-Hill; 2004.
10. Sola A, Rogido MR, Deulofeut R. Oxygen as a neonatal health hazard: call for detente in clinical practice.
Acta Paediatr. 2007;96(6):801-812.
11. Wilkinson AR, Phibbs RH, Heilbron DC, Gregory
GA, Versmold HT. In Vivo Oxygen Dissociation
Curves in Transfused and Untransfused Newborns
With Cardiopulmonary Disease. Am Rev Respir Dis
1980;122(4):629-634.
12. Widness JA, Seward VJ, Kromer IJ, Burmeister LF,
Bell EF, Strauss RG. Changing patterns of red blood
cell transfusion in very low birth weight infants.
13. Bishara N, Ohls RK. Current controversies in the
management of the anemia of prematurity.
Seminars in Perinatology. 2009;33(1):29-34.
14. Bard H, Teasdale F. Red cell oxygen affinity,
hemoglobin type, 2, 3-diphosphoglycerate, and
pH as a function of fetal development. Pediatrics.
1979;64(4):483-487.
15. Wimberley PD. Oxygen monitoring in the newborn.
Scandinavian Journal of Clinical & Laboratory
Investigation. 1993;214:127-130.
16. Garby L, Sjolin S, Vuille JC. The relative rate of synthesis of haemoglobin F and haemoglobin A during the
first months of life. Acta Paediatr. 1962;51(2):245-254.
17. Delivoria-Papadopoulos M, Morrow G, Oski FA.
Exchange transfusion in the newborn infant with
fresh and old blood: the role of storage on 2,
3-diphosphoglycerate, hemoglobin-oxygen affinity,
and oxygen release. J Pediatr. 1971;79(6):898-903.
18. Emond D, Lachance C, Gagnon J, Bard H. Arterial
partial pressure of oxygen required to achieve 90%
saturation of hemoglobin in very low birth weight
newborns. Pediatrics. 1993;91(3):602-604.
19. Halleux VD, Truttmann A, Gagnon C, Bard H.
The effect of blood transfusion on the hemoglobin
oxygen dissociation curve of very early preterm
infants during the first week of life. Seminars in
Perinatology. 2002;26(6):411-415.
20. Murkovic I. Sensors in neonatal monitoring: Current
practice and future trends. Technology and Health
Care. 2003;11(6):399-412.
21. Giuliano KK, Higgins TL. New-Generation Pulse
Oximetry in the Care of Critically Ill Patients. In:
AACN; 2005. p. 26-37.
22. Stern L. Oxygen toxicity in premature infants.
Graefes Archive for Clinical and Experimental
Ophthalmology. 1975;195(2):71-76.
23. Wright J. Introduction to pulse oximetry. British
Journal of Perioperative Nursing. 2003;13(11):456-460.
24. Giuliano KK, Higgins TL. New-generation pulse
oximetry in the care of critically ill patients.
American Journal of Critical Care. 2005;14(1):26-37.
25. Woolley A, Hickling K. Errors in measuring blood
gases in the intensive care unit: Effect of delay in
estimation. Journal of Critical Care. 2003;18(1):31-37.
26. Ganesan V, Malbon S, Jones PW, Spencer SA. Blood
gas monitoring in very low birth weight infants
receiving respiratory support: a survey of current
practice in the UK and Ireland. Journal of Clinical
Excellence. 2002;4(4):403-410.
27. Thorson SH. Variability of arterial blood gas values in
stable patients in the ICU. Chest. 1983;84(1):14-18.
28. Mahutte CK. On-line arterial blood gas analysis
with optodes: current status. Clinical Biochemistry.
1998;31(3):119-130.
29. Clutton-Brock TH, Fink S, Markele D, Luthra AJ,
Hendry SP. The evaluation of a new intravascular
blood gas monitoring system in the pig. Journal of
Clinical Monitoring and Computing. 1994;10(6):387391.
02-08-12 11:36
52
CHAPTER 2
30. Haller M, Kilger E, Briegel J, Forst H, Peter K. Continuous intra-arterial blood gas and pH monitoring
in critically ill patients with severe respiratory failure:
A prospective, criterion standard study. Critical Care
Medicine. 1994;22(4):580-587.
31. Tobias JD. Monitoring of pH and PCO2 in children
using the Paratrend 7 in a peripheral vein. Canadian
Journal of Anesthesia. 1998;45(1):81-83.
32. Huch A, Huch R, Neumayer E, Rooth G. Continuous
intra-arterial PO2 measurements in infants. Acta
Paediatr. 1972;61(6):722-723.
33. Tobias JD, Connors D, Strauser L, Johnson T. Continuous pH and PCO2 monitoring during respiratory
failure in children with the Paratrend 7 inserted into
the peripheral venous system. Journal of Pediatrics.
2000;136(5):623-627.
34. Weiss IK, Fink S, Edmunds S, Harrison R, Donnelly
K. Continuous arterial gas monitoring: initial experience with the Paratrend 7 in children. Intensive Care
Med. 1996;22(12):1414-1417.
35. Coule LW, Truemper EJ, Steinhart CM, Lutin WA.
Accuracy and utility of a continuous intra-arterial
blood gas monitoring system in pediatric patients.
Critical Care Medicine. 2001;29(2):420-426.
36. Weiss IK, Fink S, Harrison R, Feldman JD, Brill JE.
Clinical use of continuous arterial blood gas monitoring in the pediatric intensive care unit. Pediatrics.
1999;103(2):440-445.
37. Conway M, Durbin GM, Ingram D, McIntosh N,
Parker D, Reynolds EO, et al. Continuous monitoring
of arterial oxygen tension using a cathetertip polarographic electrode in infants. Pediatrics.
1976;57(2):244-250.
38. Rais-Bahrami K, Rivera O, Mikesell GT, Short
BL. Continuous blood gas monitoring using an
in-dwelling optode method: clinical evaluation of
the Neotrend sensor using a luer stub adaptor to
access the umbilical artery catheter. J Perinatol.
2002;22(5):367-369.
39. Goddard P, Keith I, Marcovitch H, Roberton NRC,
Rolfe P, Scopes JW. Use of a continuously recording
intravascular oxygen electrode in the newborn. Arch
Dis Child. 1974;49(11):853-860.
40. Strauss AW, Escobedo M, Goldring D. Continuous
monitoring of arterial oxygen tension in the newborn infant. J Pediatr. 1974;85(2):254-261.
41. Morgan C, Newell SJ, Ducker DA, Hodgkinson J,
White DK, Morley CJ, et al. Continuous neonatal
blood gas monitoring using a multiparameter intraarterial sensor. Arch Dis Child Fetal Neonatal Ed.
1999;80(2):93.
42. Pollitzer MJ, Soutter LP, Reynolds EO. Continuous

monitoring of arterial oxygen tension in infants: four
years of experience with an intravascular oxygen
electrode. Pediatrics. 1980;66(1):31-36.
43. Ganter M, Zollinger A. Continuous intravascular
blood gas monitoring: development, current techniques, and clinical use of a commercial device. Br J
Anaesth. 2003;91(3):397-407.
44. Venkatesh B, Hendry SP. Continuous intra-arterial
blood gas monitoring. Intensive Care Medicine.
1996;22(8):818-828.
45. Menzel M, Henze D, Soukup J, Engelbrecht K,
Senderreck M, Clausen T, et al. Experiences with
continuous intra-arterial blood gas monitoring.
Minerva Anestesiol. 2001;67(4):325-331.
46. Vailas GN, Brouillette RT, Scott JP, Shkolnik A,
Conway J, Wiringa K. Neonatal aortic thrombosis:
recent experience. J Pediatr. 1986;109(1):101-108.
47. Spencer A, Rolfe P, Murphy A, Wright T, McKenna
S. In vivo assessment of catheter-tip PO2 sensor:
sampling lumen fabrication. Medical and Biological
48. Mermel LA. Prevention of intravascular catheterrelated Infections. Annals of Internal Medicine.
2000;132(5):391-402.
49. Nicklin S, Wickramasinghe YA, Andrew Spencer S.
Neonatal intensive care monitoring. Current
Paediatrics. 2004;14(1):1-7.
50. Clark Jr LC. Monitor and control of blood and tissue
oxygen tensions. ASAIO Journal. 1956;2(1):41.
51. Eberhard P, Mindt W, Jann F, Hammacher K.
Continuous PO2 monitoring in the neonate by skin
electrodes. Medical and Biological Engineering and
Computing. 1975;13(3):436-442.
52. Mik V, Krauss AN, Ross GS. Doctors and the health
industry: A case study of transcutaneous oxygen
monitoring in neonatal intensive care. Social Science
& Medicine. 1996;42(9):1247-1258.
53. Fanconi S, Doherty P, Edmonds JF, Barker GA,
Bohn DJ. Pulse oximetry in pediatric intensive care:
comparison with measured saturations and transcutaneous oxygen tension. Journal of Pediatrics.
1985;107(3):362.
54. Krauss AN, Waldman S, Frayer WW, Auld PA.
Non-invasive estimation of arterial oxygenation in
newborn infants. Journal of Pediatrics. 1978;93(2):275.
55. Fanconi S, Tschupp A, Molinari L. Long-term transcutaneous monitoring of oxygen tension and carbon
dioxide at 42 C in critically ill neonates: improved
performance of the tcPO2 monitor with topical
metabolic inhibition. European Journal of Pediatrics.
1996;155(12):1043-1046.
02-08-12 11:36
56. Saugstad OD. Is oxygen more toxic than currently

believed? Pediatrics. 2001;108(5):1203-1205.
57. Mike V, Krauss AN, Ross GS. Reflections on a medical Innovation: Transcutaneous oxygen monitoring
in neonatal intensive care. Technology and Culture.
1993;34(4):894-922.
58. Aoyagi T, Miyasaka K. Pulse oximetry: its invention,
contribution to medicine, and future tasks.
Anesthesia & Analgesia2002;94(1 Suppl):S1-3.
59. Finer N, Leone T. Oxygen saturation monitoring for
the preterm infant: The evidence basis for current
practice. Pediatric Research. 2009;65(4):375-380.
60. Salyer JW. Neonatal and pediatric pulse oximetry.
Respiratory Care. 2003;48(4):386-396.
61. Wieben O. Light absorbance in pulse oximetry.
In: Webster JG, editor. Design of Pulse Oximeters:
Taylor & Francis; 1997.
62. Sola A, Chow L, Rogido M. Pulse oximetry in
neonatal care in 2005. A comprehensive state of
the art review. Anales de Pediatra (Barcelona).
2005;62(3):266-280.
63. Paky F, Koeck CM. Pulse oximetry in ventilated
preterm newborns: reliability of detection of hyperoxaemia and hypoxaemia, and feasibility of alarm
settings. Acta Paediatr. 1995;84(6):613-617.
64. Bucher HU, Fanconi S, Baeckert P, Duc G. Hyperoxemia in newborn infants: detection by pulse
oximetry. Pediatrics. 1989;84(2):226-230.
1994;22(6):981-985.
66. Trivedi NS, Ghouri AF, Shah NK, Lai E. Effects
of motion, ambient light, and hypoperfusion on
pulse oximeter function. Survey of Anesthesiology.
1998;42(3):168.
67. Severinghaus JW, Spellman Jr MJ. Pulse oximeter
failure thresholds in hypotension and vasoconstriction. Anesthesiology. 1990;73(3):532.
68. Sahni R, Gupta A, Ohira-Kist K, Rosen TS. Motion
resistant pulse oximetry in neonates. British Medical
Journal. 2003;88(6):F505508.
69. Belal SY, Taktak AFG, Nevill AJ, Spencer SA. A fuzzy
system for detecting distorted plethysmogram
pulses in neonates and paediatric patients.
Physiological Measurement. 2001;22(2):397-412.
70. Rozycki HJ, Sysyn GD, Marshall MK, Malloy R,
Wiswell TE. Mainstream end-tidal carbon dioxide
monitoring in the neonatal intensive care unit.
Pediatrics. 1998;101(4):648-653.
53
71. Chambrin MC. Alarms in the intensive care unit:

how can the number of false alarms be reduced.
Critical Care. 2001;5(4):184-188.
72. Aaron JN, Carlisle CC, Carskadon MA, Meyer TJ, Hill
NS, Millman RP. Environmental noise as a cause of
sleep disruption in an intermediate Respiratory Care
unit. Sleep. 1996;19(9):707-710.
73. Sabar R, Zmora E. Nurses response to alarms from
monitoring systems in NICU. Pediatric Research.
1997;41(4 Part 2):174.
74. Wyatt JS, Deply DT, Cope M, Wray S, Reynolds
EOR. Quantification of cerebral oxygenation
and haemodynamics in sick newborn infants by
near infrared spectrophotometry. The Lancet.
1986;2(8515):1063-1066.
75. Brazy JE. Cerebral oxygen monitoring with near
infrared spectroscopy: Clinical application to
neonates. Journal of Clinical Monitoring and
Computing. 1991;7(4):325-334.
76. Faris F, Thorniley M, Wickramasinghe Y, Houston
R, Rolf P, Livera N, et al. Noninvasive in vivo nearinfrared optical measurement of the penetration
depth in the neonatal head. Clin Phys Physiol Meas.
1991;12:353358
77. Livera LN, Spencer SA, Thorniley MS, Wickrmasinghe YA, Rolfe P. Effects of hypoxaemia and
bradycardia on neonatal cerebral haemodynamics.
Arch Dis Child. 1991;66(4):376-80.
78. Von Siebenthal K, Bernert G, Casaer P. Nearinfrared spectroscopy in newborn infants. Brain &
development. 1992;14(3):135-143.
79. Wickramasinghe YA, Rolfe PJ, Palmer K, Watkins S,
Spencer SA, Doyle M, et al. Development and clinical evaluation of noninvasive near-infrared monitoring of cerebral oxygenation. In: Wolfbeis OS, editor.
Biochemical and medical sesnors; 1993; Budapest,
Hungary: Proc. SPIE; 1993. p. 29.
80. Wahr JA, Tremper KK, Samra S, Delpy DT.
Near-infrared spectroscopy: theory and
applications. J Cardiothorac Vasc Anesth.
1996;10(3):406-18.
81. Adcock LM, Wafelman LS, Hegemier S, Moise AA,
Speer ME, Contant CF, et al. Neonatal intensive care
applications of near-infrared spectroscopy. Clinics in
Perinatology. 1999;26(4):893-903.
82. Tsuji M, Saul JP, du Plessis A, Eichenwald E, Sobh J,
Crocker R, et al. Cerebral intravascular oxygenation
correlates with mean arterial pressure in critically ill
premature infants. Pediatrics. 2000;106(4):625-632.
02-08-12 11:36
54
CHAPTER 2
83. Naulaers G, Morren G, Van Huffel S, Casaer P,

Devlieger H. Cerebral tissue oxygenation index
in very premature infants. Arch Dis Child Fetal
Neonatal Ed. 2002;87(3):189-192.
84. Reynolds EOR, Wyatt JS, Azzopardi D, Delpy DT,
Cady EB, Cope M, et al. New non-invasive methods
for assessing brain oxygenation and haemodynamics.
Br Med Bull. 2004;44(4):1052-1075.
85. Brazy JE, Lewis DV, Mitnick MH, Jobsis van der Vliet
FF. Noninvasive monitoring of cerebral oxygenation
in preterm infants: preliminary observations.
Pediatrics. 1985;75(2):217-225.
86. Nicklin SE, Hassan IAA, Wickramasinghe YA,
Spencer SA. The light still shines, but not that
brightly? The current status of perinatal near
infrared spectroscopy. Arch Dis Child Fetal Neonatal
Ed. 2003;88(4):263-268.
87. Toet MC, Lemmers PMA, van Schelven LJ, van Bel
F. Cerebral oxygenation and electrical activity after
birth asphyxia: Their relation to outcome. Pediatrics
2006;117(2):333-339.
88. Bernal NP, Hoffman GM, Ghanayem NS, Arca MJ.
Cerebral and somatic near-infrared spectroscopy
in normal newborns. Journal of Pediatric Surgery.
2010;45(6):1306-1310.
89. Greisen G. Is near-infrared spectroscopy living up
to its promises? Seminars in Fetal and Neonatal
Medicine. 2006;11(6):498-502.
90. McDonald MJ, Montgomery VL, Cerrito PB, Parrish
CJ, Boland KA, Sullivan JE. Comparison of end-tidal
CO2 and PaCO2 in children receiving mechanical ventilation. Pediatric Critical Care Medicine.
2002;3(3):244-249.
91. Hand IL, Sheppard EK, Krauss AN, Auld PAM. Discrepancies between transcutaneous and end-tidal
carbon dioxide monitoring in the critically ill neonate
with respiratory distress syndrome. Critical Care
Medicine. 1989;17(6):556.
92. Geven W, Nagler E, de Boo T, Lemmens W.
Combined transcutaneous oxygen, carbon dioxide
tensions and end-expired CO2 levels in severely ill
newborns. Advances in experimental medicine and
biology. 1987;220:115.
93. Tobias JD, Meyer DJ. Noninvasive monitoring of
carbon dioxide during respiratory failure in toddlers
and infants: end-tidal versus transcutaneous carbon
dioxide. Anesthesia & Analgesia. 1997;85(1):55.
94. Wyllie J, Carlo WA. The role of carbon dioxide detectors for confirmation of endotracheal tube position.
Clinics in perinatology. 2006;33(1):111-119.
95. Hammer J. Ventilator strategies--what monitoring
is helpful? Paediatric respiratory reviews.
2006;7(Supp. 1):S183-185.
96. Singh SA, Singhal N. Does end-tidal carbon dioxide

measurement correlate with arterial carbon dioxide
in extremely low birth weight infants in the first
week of life? Indian Pediatrics. 2006;43(1):20-25.
97. Wu CH, Chou HC, Hsieh WS, Chen WK, Huang PY,
Tsao PN. Good estimation of arterial carbon dioxide
by end tidal carbon dioxide monitoring in the
neonatal intensive care unit. Pediatric pulmonology.
2003;35(4):292-295.
98. Nangia S, Saili A, Dutta AK. End tidal carbon dioxide
monitoringits reliability in neonates. Indian journal
of pediatrics. 1997;64(3):389-394.
99. Curley MA, Thompson JE. End-tidal CO2 monitoring
in critically ill infants and children. Pediatric Nursing.
1990;16(4):397-403.
100. Aliwalas LLD, Noble L, Nesbitt K, Fallah S, Shah
V, Shah PS. Agreement of carbon dioxide levels
measured by arterial, transcutaneous and end tidal
methods in preterm infants 28 weeks gestation. J
Perinatol. 2004;25(1):26-29.
101. Nicolini U, Nicolaidis P, Fisk NM, Vaughan JI, Fusi L,
Gleeson R, et al. Limited role of fetal blood sampling
in prediction of outcome in intrauterine growth
retardation. The Lancet. 1990;336(8718):768-772.
102. Kamlin COF, ODonnell CPF, Davis PG, Morley CJ.
Oxygen saturation in healthy infants immediately
after birth. Journal of Pediatrics.
2006;148(5):585-589.
103. Dimich I. Evaluation of oxygen saturation monitoring
by pulse oximetry in neonates in the delivery system.
Canadian Journal of Anesthesia. 1991;38(8):985-988.
104. Mariani G, Dik PB, Ezquer A, Aguirre A, Esteban
ML, Perez C, et al. Pre-ductal and post-ductal O2
saturation in healthy term neonates after birth.
105. Toth B, Becker A, Seelbach-Gbel B. Oxygen
saturation in healthy newborn infants immediately
after birth measured by pulse oximetry. Archives of
Gynecology and Obstetrics. 2002;266(2):105-107.
106. Levesque BM, Pollack P, Griffin BE, Nielsen HC.
Pulse oximetry: Whats normal in the newborn
nursery? Pediatric pulmonology. 2000;30(5):406-412.
107. Wilkinson AR, Phibbs RH, Gregory GA. Continuous in vivo oxygen saturation in newborn infants
with pulmonary disease: A new fiberoptic catheter
oximeter. Crit Care Med 1979;7(5):232-236.
108. Mok JYQ, Hak H, McLaughlin FJ, Pintar M, Canny
GJ, Levison H. Effect of age and state of wakefulness on transcutaneous oxygen values in preterm
infants: a longitudinal study. Journal of Pediatrics.
1988;113(4):706-709.
02-08-12 11:36
109. Weinberger B, Laskin DL, Heck DE, Laskin JD.

Oxygen toxicity in premature infants. Toxicology
and Applied Pharmacology. 2002;181(1):60-67.
110. Weisbrot I, James L, Prince C, Holaday D, Apgar
V. Acid-base homeostasis of the newborn infant
during the first 24 hours of life. Journal of Pediatrics.
1958;52(4):395-403.
111. Graham BD, Koeff ST, Tsao MU, Sloan C, Wilson JL.
Direct plasma oxygen tension determination in infants and adults in air and in modified atmospheres.
Amer J Dis Child. 1959;98:593.
112. Oliver Jr TK, Demis JA, Bates GD. Serial blood
gas tensions and acid base balance during the
first hour of life in human infants. Acta Paediatr.
1959;48(4):346-360.
113. Orzalesi MM, Mendicini M, Bucci G, Scalamandre A,
Savignoni PG. Arterial oxygen studies in premature
newborns with and without mild respiratory disorders. Arch Dis Child. 1967;42(222):174-180.
114. Askin DF. Interpretation of neonatal blood gases,
Part II: Disorders of acid-base balance. Neonatal
Netw. 1997;16(6):23-29.
115. Flynn JT, Bancalari E, Snyder ES, Goldberg RN, Feuer
W, Cassady J, et al. A cohort study of transcutaneous oxygen tension and the incidence and severity
of retinopathy of prematurity. New England Journal
of Medicine. 1992;326(16):1050-1054.
116. Brouillette RT, Waxman DH. Evaluation of the
newborns blood gas status. Clinical Chemistry.
1997;43(1):215-221.
117. Saito Y, Omoto T, Cho Y, Hatsukawa Y, Fujimura M,
Takeuchi T. The progression of retinopathy of prematurity and fluctuation in blood gas tension. Graefes
Arch Clin Exp Ophthalmol. 1993;231(3):151-156.
118. Cunningham S, Fleck BW, Elton RA, McIntosh N.
Transcutaneous oxygen levels in retinopathy of
prematurity. The Lancet. 1995;346(8988):1464-1465.
119. York JR, Landers S, PhD RSK, PhD PGA, PhD JSP.
Arterial oxygen fluctuation and retinopathy of
prematurity in very-low-birth-weight infants.
J Perinatol. 2004;24:82-87.
120. McColm JR, Cunningham S, Wade J, Sedowofia K,
Gellen B, Sharma T, et al. Hypoxic oxygen fluctuations produce less severe retinopathy than hyperoxic
fluctuations in a rat model of retinopathy of prematurity. Pediatric Research. 2004;55(1):107-113.
121. Phelps DL, Lindblad A, Bradford D. Supplemental
therapeutic oxygen for prethreshold retinopathy of
prematurity (STOPROP), a randomized, controlled
trial. I: Primary outcomes. The STOP-ROP Multicenter Study Group. Pediatrics. 2000;105:295-310.
55

year in babies of less than 28 weeks gestation.
British Medical Journal. 2001;84(2):F106-110.
123. Askie LM, Henderson-Smart DJ, Irwig L, Simpson
JM. Oxygen-saturation targets and outcomes
in extremely preterm infants. N Engl J Med.
2003;349(10):959-967.
124. Anderson CG, Benitz WE, Madan A. Retinopathy of
prematurity and pulse oximetry: A national survey of
recent practices. J Perinatol. 2004;24:164-168.
125. SUPPORT Study Group of the Eunice Kennedy
Shriver NICHD. Target ranges of oxygen saturation
in extremely preterm infants. The New England
Journal of Medicine. 2010;362(21):1959-1969.
126. Gupta R, Yoxall CW, Subhedar N, Shaw NJ.
Individualised pulse oximetry limits in neonatal
intensive care. Arch Dis Child Fetal Neonatal Ed.
1999;81(3):194-196.
127. Stenson B, Brocklehurst P, Tarnow-Mordi W.
Increased 36-week survival with high oxygen
saturation target in extremely preterm infants. New
128. Castillo A, Sola A, Baquero H, Neira F, Alvis R, Deulofeut R, et al. Pulse oxygen saturation levels and
arterial oxygen tension values in newborns receiving
oxygen therapy in the neonatal intensive care unit:
is 85% to 93% an acceptable range? Pediatrics.
2008;121(5):882-889.
129. Lister G, Moreau G, Moss M, Talner NS. Effects of
alterations of oxygen transport on the neonate.
Seminars in Perinatology. 1984;8(3):192-204.
130. Hirschl RB, Palmer P, Heiss KF, Hultquist K,
Fazzalari F, Bartlett RH. Evaluation of the right atrial
venous oxygen saturation as a physiologic monitor
in a neonatal model. Journal of Pediatric Surgery.
1993;28(7):901-905.
131. Ince C, Sinaasappel M. Microcirculatory oxygenation
and shunting in sepsis and shock. Critical Care
Medicine. 1999;27(7):1369-1377.
132. Ellis C. Microcirculatory flows, microcirculatory
responsiveness, microcirculatory and regional
(arteriolar/venular) 02 saturations. In: Sibbald WJ,
Messmer K, Fink MP, editors. Tissue oxygenation in
acute medicine. Berlin Heide: Springer; 2002. p. 204.
133. Kyriacou PA, Jones DP, Langford RM, Petros AJ. A pilot
study of neonatal and pediatric esophageal pulse oximetry. Anesthesia & Analgesia. 2008;107(3):905-908.
02-08-12 11:36
56
02-08-12 11:36
CHAPTER 3
New-generation
pulse oximeters in
extremely low birth
weight infants: how
do they perform in
clinical practice?
A.C. van der Eijk, S. Horsch, P.H.C. Eilers, J. Dankelman, B.J. Smit
02-08-12 11:36
58
CHAPTER 3
Pulse oximetry is widely used in medical care. The oxygen saturation measured by
a pulse oximeter (SpO2 ) is by some even considered as the fifth vital sign (next to
body temperature, heart rate, blood pressure, and respiratory rate).1 Despite the
numerous advantages, a large disadvantage of pulse oximetry is the sensitivity for
motion artifacts. In the end of the 20th century, new types of motion resistant
pulse oximeters were introduced. This chapter discusses the performance of these
new-generation pulse oximeters in extremely low birth weight infants. The contents of this chapter are based on the article New-generation pulse oximeters in
extremely low birth weight infants: how do they perform in clinical practice?
(published in The Journal of Perinatal and Neonatal Nursing, Vol 26:2, 2012).
OBJECTIVE To evaluate the performance of new-generation pulse oximeters of three
different brands in extremely low birth weight infants.
02-08-12 11:36
New-generation pulse oximeters in ELBW infants
59
3.1 INTRODUCTION
In extremely low birth weight (1000 g.; ELBW) infants, oxygenation needs to be monitored
closely because of its narrow therapeutic range.2 Pulse oximetry is a widely used method for
monitoring arterial oxygen saturation continuously (SpO2).
Although pulse oximetry is a patient-friendly and non-invasive method, there are some
well known disadvantages of this technique. The accuracy of pulse oximetry is easily influenced by movement artefacts,3 low blood perfusion,4 and ambient light.5 Due to these limitations, pulse oximetry is notorious for its high alarm rate.6, 7 Furthermore, the technique
was not designed for detecting hyperoxia,8, 9 and cannot counteract for the presence of fetal
heamoglobin.10-12 The latter two limitations are especially of importance in preterm infants
as neonatal hyperoxia is associated with injury to a.o. the developing brains,13 lung,14 and
retina.15, 16
In the last two decades, several brands have introduced motion-resistant pulse oximeters.
The performance of both these new-generation pulse oximetry devices and the conventional ones have been evaluated extensively in (preterm) neonates.3, 8, 9, 17-24 Although conclusions varied between studies, Giuliano et al.25 concluded in a review that the clinical performance of new generation pulse oximeters is better than conventional pulse oximeters. They
did not find evidence for differences in performances between brands.
In the years that pulse oximetry technology has advanced, survival rates of ELBW infants
have increased remarkably.26, 27 Amongst survivors, the incidence of oxidative stress related
diseases is considerable.28 Hence, reliable monitoring of oxygenation is essential in this
most vulnerable group of patients.
In our department, physicians and nurses had diverse impressions of the performance of
pulse oximeter brands when applied to ELBW infants. Therefore, the aim of this study was
to evaluate the performance of new-generation pulse oximeters in ELBW infants. To objectively evaluate the performance, a prospective crossover observational study was designed
where pulse oximeter measurements of either equal or different brands were compared by
simultaneous dual measurement.
Pulse oximetry is a widely used

method for monitoring arterial oxygen
saturation continuously.
02-08-12 11:36
60
CHAPTER 3
3.2 Methods
The prospective crossover observational study was performed at a level-III-c29 neonatal
intensive care unit (NICU). Approval was obtained from the ethical review board of the
Erasmus Medical Centre - Sophia Childrens Hospital, Rotterdam, the Netherlands. Informed written consent by the parents was given for all included infants prior to inclusion.
3.2.1 Patients
Infants were eligible for inclusion when the gestational age (GA) was <30 weeks, the birth
weight (BW) was 1000 gram, and the postnatal age (PNA) of the infant was <14 days. Data
recording started when it was possible to place a disposable pulse oximeter probe on both
feet (for study purpose) and on the left hand (for clinical purpose). Infants were excluded
when the status of the skin was poor (e.g., hematoma or skin lesions), or when infants were,
according to the responsible medical staff member, too instable to handle.
3.2.2 Study set-up
Two, brand-new, disposable new-generation pulse oximeter probes from one of the three
brands used in the study, were placed around either foot by the nurse taking care of the
infant. The devices used to collect the data, and the characteristics for each pulse oximeter
brand are shown in Table 3.1. The researcher told the nurse which probe should be placed
on which foot. For example, probe I of brand A (IA) was placed on the left foot, and probe
II of brand A (IIA) on the right foot. To prevent interference of light, and thus erroneous
measurements, all probes used during this study were covered by an extra opaque wrap,
as suggested by Ralston et al.30 and Fouzas et al.31
Table 3.1 Pulse oximeter characteristics, and specifications of the data recording
Pulse oximeter characteristics
Brand
Software
Masimo
Experimental set-up
Disposable
probe
Monitoring
device
SET
LNOP
NeoPt-L
<1 kg
Philips
FAST
Nellcor
Oximax
Averaging
time
[s]
Recording
software
Recording
frequency
[Hz]
Recorded
SpO2 value
MP50,
M1020B
A03 &
M3001 A
Trendface
~1
One
decimal
behind
the dot
M1133A
<3 kg
MP50,
M1020B
A01 &
M3001 A
Trendface
~1
One
decimal
behind
the dot
Softcare SCPR <1.5kg
N600x
2-4
Tera Term
Integers
02-08-12 11:36
61
After a minimum of 50 minutes of dual SpO2-monitoring, the probes were switched over
to the other foot. Table 3.2 illustrates the study design, and especially which probe was positioned on which foot. Thus, following the example, probe IA was placed on the right foot,
and probe IIA on the left foot. After the second hour, one of the probes was removed and
replaced by a pulse oximeter from a different brand: probe IA stayed on the right foot, probe
IIA was removed from the infant, and probe I of brand B (IB) was positioned on the left
foot. After the third hour, the pulse oximeter probes were switched again to the other foot.
Thus, probe IB was now positioned on the right foot, and the probe IA on the left foot. After
the fourth hour, the recordings ended, and both the pulse oximeter probes were removed
from the infant. To prevent that NICU personnel would be influenced by the two extra
pulse oximeter measurements in their daily care, SpO2 values obtained by the two additional pulse oximeters were blinded for the nursing staff.
Table 3.2 Study design
Number of
subjects
Dual SpO2
measurement by
Pulse oximeter
probe left foot
Pulse oximeter
probe right foot
Duration
Masimo SET Masimo SET
Masimo probe 1
Masimo probe 2
~ 1 hour
Masimo probe 2
Masimo probe 1
~ 1 hour
Nellcor probe 1
Masimo probe 1
~ 1 hour
Masimo probe 1
Nellcor probe 1
~ 1 hour
Nellcor probe 1
Nellcor probe 2
~ 1 hour
Nellcor probe 2
Nellcor probe 1
~ 1 hour
Philips probe 1
Nellcor probe 1
~ 1 hour
Nellcor probe 1
Philips probe 1
~ 1 hour
Philips probe 1
Philips probe 2
~ 1 hour
3
Masimo SET Nellcor Oximax
Nellcor Oximax Nellcor Oximax
3
Nellcor Oximax Philips FAST
Philips FAST Philips FAST
3
Philips FAST Masimo SET
Philips probe 2
Philips probe 2
~ 1 hour
Masimo probe 1
Philips probe 1
~ 1 hour
Philips probe 1
Masimo probe 1
~ 1 hour
To compare each of the three brands with each other, three different tests were performed. Each test
was performed in three subjects (column 1). Two pulse oximeter probes of either equal or different brand
were placed on the feet of the subject for approximately four hours (column 2). During these four hours,
pulse oximeter probes were switched between feet every hour (column 3 & 4).

A multi-parameter patient monitor was used for monitoring SpO2 with Masimo or Philips
probes (MP50, modules: M1020B A01 & A03, MMS: M3001 A, Philips Medical Systems, Boeblingen, Germany). Dual SpO2 monitoring by Nellcor was performed by two stand-alone
pulse oximeters (model: N600x). The averaging time of the software of both Masimo and
02-08-12 11:36
62
CHAPTER 3
Philips was 5 seconds. The Nellcor-software could not be set to 5 seconds, but required a
range for the averaging time. The range closest to 5 seconds was chosen: subsequently, the
averaging time was 2 to 4 seconds. A laptop was used to record SpO2-measurements (software: Tera Term, open source, and Trendface from Ixellence GmbH, Wildau, Germany).
3.2.4 Data analysis
To prevent artefacts in SpO2 values caused by (re)positioning of the probes, and to
make sure that the data used for analysis was exactly the same length, time periods of 45
minutes were analysed. The data obtained in the first minute after the moment that both
pulse oximeters provided an SpO2 value, were excluded from analysis. The next 2nd up to 46th
minute of collected data were used for analysis. All SpO2 data was interpolated to 1 Hz and
rounded to integers. The absolute difference between simultaneously obtained SpO2 values
was defined as SpO2 . When one or both pulse oximeters did not provide an SpO2 value,
these time periods were defined as drop outs, and excluded for calculation of SpO2 .
There was no gold standard to compare the pulse oximeter measurements with: SpO2
always is the difference between two sensors (either of equal or different brands). So it is
not possible to detect any bias. We investigated whether there is a difference in variance.
The variance of a difference of the two pulse oximeters is the sum of the variance of the
individual variances. The study set-up was designed such that all possible combinations
of brands were tested, thus a linear regression model, for the observed variance of the
difference, could be developed. The independent variables are indicators of which two
brands were compared, and the dependent variable is the observed variance.
3.3 Results
Nine infants, all Caucasian, with a mean SD gestational age of 26 3/7 1 4/7 weeks,
birth weight of 825 136 gram, and postnatal age of 7 4 days were included. The patient
characteristics are shown in Table 3.3. In total, 36 periods of 45 minutes (i.e. 27 hours) of
dual SpO2 measurements were analysed. During the study, infants were either on invasive
ventilation (9 hours) or CPAP (18 hours).
02-08-12 11:36
63
Table 3.3 Patient characteristics
Ventilation modes
during data
recording
Patient
number
Gender
GA
[weeks]
Birth weight
[grams]
PNA at day
of study
[days]
26 4 / 7
710
SIMV
25 3 / 7
800
CPAP
26 5 / 7
1000
SIMV
23 6 / 7
720
11
NIV
26 4 / 7
700
11
NIV
27 3 / 7
705
HFO
24 5 / 7
795
12
CPAP
29 4 / 7
995
NIV
26 6 / 7
1000
NIV
SpO2 without drop-out

median*
(min-Q1Q3-max)
94
(68-94-97-100)
90
(71-87-93-100)
96
(75-95-97-100)
92
(47-87-95-100)
89
(75-87-92-99)
92
(52-89-94-100)
94
(65-92-96-98)
89
(74-87-91-97)
92
(48-90-93-100)
SpO2
median
(min-Q1Q3-max)
1
(0-1-2-19)
(0-1-3-13)
(0-1-2-14)
(0-2-5-14)
(0-1-3-26)
(0-0-2-8)
(0-2-4-16)
(0-1-2-13)
(0-1-2-7)
* The median SpO2 (column 7) was based on the measurements of the research probe that was
positioned on the patient for the complete recording time (Q1 = 25th percentile, Q3 = 75th percentile).
3.3.1 SpO2 values

For each infant the median (min - Q1 - Q3 - max) SpO2 was determined, using the SpO2
measurements of the research probe placed on the infant for the complete 4 x 45 minutes
of recording (the 2nd till 46th minute from each measurement period, see Table 3.3). In the
27 hours of dual SpO2 recording, 6 drop-outs were observed in total, in 4 different infants
(4 times Philips, once Masimo, and once Nellcor). These drop outs lasted between 2 to 21
seconds (in total 0.5% of the recording time). Never more than one pulse oximeters dropped
out at the same time.
Desaturations <70% showed by at least one pulse oximeter occured in 34 moments in 5 different infants. In 11 (32%) of the 34 occasions, the second pulse oximeter showed a desaturation <70% as well. Hence, in 23 (67%) of the 34 occasions only one of the pulse oximeters
showed a desaturation <70%. Of these 23 times, in 6 times (26%) the desaturation lasted for
more than 20 seconds.
02-08-12 11:36
64
CHAPTER 3
Figures 3.1 to 3.4 show examples of typical situations that occurred during the recording,
and corresponding frequency histograms of the SpO2 measurements. In Figure 3.1 the
SpO2 values of the right foot are continuously higher than SpO2 values of the left foot.
After switching the probes to the other foot, the measurements from the right foot are still
higher than the left foot. Figure 3.2 shows that SpO2 varies over time in one infant. After
changing the probes to the other foot, SpO2 is smaller. Figure 3.3 shows two examples of a
desaturation <80% by only one of the pulse oximeters. Figure 3.4 shows two desaturations
<50% by two pulse oximeters at the same time.
Median (range) SpO2 was 2 (0 to 26)% and SpO2 varied over time. In 9% of the time
SpO2 was 5% (drop-out time excluded). In 31 of the 36 time periods, the standard deviation of the difference was between 1.6 and 3.8%. In the remaining five time periods the
standard deviation of the difference was higher (range 6.0 to 9.4%). These relatively large
differences were related to deep desaturations in only one pulse oximeter (e.g., Figure 3.3),
and/or due to the fact that there was a short time difference between the pulse oximeter
measurements. These situations are present in clinical practice, but not representative for
normal operation. Their impact on the linear regression model was large. Therefore, it was
decided to exclude these exceptional situations in the calculations, resulting in variances
of 2.0, 2.9 and 3.3 for the three brands, with standard errors (SE) around 0.7. The SE of the
differences between the three brands (equation VII) were around 1.0.
SE of the differences between brands = ((SE brand A)2 + (SE brand B)2)
(VII)
Dividing the differences of the means by the SE of the differences results in 0.8, 0.5 and
1.3. Even in case of a normal distribution, the results would have to be larger than 1.96
to indicate a significant difference (P <0.05). Thus, no significant differences between the
variances of the difference of the three brands were found. It is more convenient to express
variances as standard deviations, by taking square roots. We obtained the values 1.4, 1.7 and
1.8 in units of SpO2 .
02-08-12 11:36
65
100
100
95
95
SpO2 [%]
SpO2 [%]
90
85
80
75
0
100
200
300
time [s]
400
500
600
20
100
200
300
time [s]
16
Frequency [%]
12
10
8
8
6
2
0
70 72
74
76
78
80
82
84
86
88
90
92
94
96
0
70 72
98 100
74
76
78
80
82
100
100
95
95
90
85
80
100
200
300
time [s]
400
500
90
92
94
96
98 10
90
85
600
20
100
200
300
time [s]
400
500
60
24
22
Patient 4; Pulse oximeter 1, left foot

Patient 4; Pulse oximeter 2, right foot
20
Frequency [%]
12
10
8
16
14
12
10
8
2
0
70 72

18
14
Frequency [%]
88
75
0
2
74
76
78
80
82
84 86 88
SpO2 [%]
90
92
94
96
98 100
0
70 72
74
76
78
80
82
84
86
SpO2 [%]
Figure 3.1A-D Figure 3.1A&C show two sets of 10 minutes of dual SpO2-measurement (different
brands) in one patient. In Figure 3.1A the SpO2 value from the right foot is 5 to 10% higher than SpO2
from the left foot. After changing the probes from feet (Figure 3.1C), the value from the right foot is still
5 to 10% higher than SpO2 from the left foot. Figure 3.1B&D show the frequency histograms for proportion of time spent at each saturation value for the complete 45 minutes including the 10 minutes of
measurement shown in Figure 3.1A&C.
86
80
75
84
SpO2 [%]
SpO2 [%]
SpO2 [%]
SpO2 [%]
16
60
18
500

10
14
400
12
18
Frequency [%]
85
80
75
90
02-08-12 11:36
88
90
92
94
96
98 10
CHAPTER 3
100
95
95
90
85
80
SpO2 [%]
100
95
SpO2 [%]
100
90
85
75
100
200
300
time [s]
400
500
600
20
100
200
300
time [s]
400
500
600
12
16

10
12
10
8
Frequency [%]
14
Frequency [%]
18
6
4
6
4
2
0
70 72
74
76
78
80
82
84
86
88
90
92
94
96
0
70 72
98 100
85
75
75
0
90
80
80
74
76
78
80
82
SpO2 [%]
84
86
88
90
92
94
96
98 100
34
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70
100
95
95
90
85
80
85
200
300
time [s]
400
500
100
200
300
time [s]
400
500
20

10
8
16
Frequency [%]
12
14
12
10
8
2
74
76
78
80
82
84 86 88
SpO2 [%]
Patient 7; Pulse oximeter 1,

left foot
right foot
72
74
76
78
80
90
92
94
96
98 100
0
70 72
74
76
78
80
82
84
86
88
90
92
94
96
98 100
SpO2 [%]
84 86 88
SpO2 [%]
100
200
300
time [s]
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70 72 74 76 78 80 82 84 86 88
SpO2 [%]
Figure 3.2A-D Figure 3.2A&C show two sets of 10 minutes of dual SpO2-measurement (equal brand) in
one patient. In Figure 3.2A the SpO2 varies over time. After switching the probes between feet (Figure
3.2C), SpO2 is smaller. Figure 3.2B&D show the frequency histograms for proportion of time spent at
each saturation value for the complete 45 minutes including the 10 minutes of measurement shown in
Figure 3.2A&C.
82
85
18
14
90
600
24
22
300
time [s]
75
0
600

0
70 72
200
80
75
100
20
16
90
80
75
18
SpO2 [%]
100
95
SpO2 [%]
100
100
SpO2 [%]
Frequency [%]
SpO2 [%]
Frequency [%]
SpO2 [%]
66
02-08-12 11:36
100
90
85
80
75
300
time [s]
400
500
600
80
82
84
86
88
90
92
94
96
Frequency [%]
98 100
34
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70
100
200
300
time [s]
400
500
600

left foot
right foot
Frequency [%]
200
72
74
76
78
80
SpO2 [%]
82
84 86 88
SpO2 [%]
90
92
94
96
98 100
100
SpO2 [%]
SpO2 [%]
95
90
85
80
75
300
time [s]
400
500
600
80
82
84
86
88
90
92
94
96
98 100
SpO2 [%]
Frequency [%]
100
200
300
time [s]
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70 72 74 76 78 80 82 84 86 88
SpO2 [%]
400
500
600
90
85
80
75
70
65
60
55
50
45
40
100
200
400
36
34
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70 72 74 76 78 80 82 84 86 88
SpO2 [%]
100
95
90
85
80
75
70
65
60
55
50
100
200
12
300
time [s]
90
500
92
400
94
96
500
600
98 100
600

10
8
6
4
2
90
92
94
96
98 100
0
70 72
74
76
78
80
82
84
86
SpO2 [%]
Figure 3.3A-D Figure 3.3A&C show two sets of 10 minutes of dual SpO2-measurement in one patient.
Both figures show short desaturations <80% by only one of the pulse oximeters. Figure 3.3A show SpO2
values obtained by equal brands; Figure 3.3C show SpO2 values obtained by different brands. Figure
3.3B&D show the frequency histograms for proportion of time spent at each saturation value for the
complete 45 minutes including the 10 minutes of measurement shown in Figure 3.3A&C.
300
time [s]
14
Frequency [%]
200
oximeter 1, left foot

oximeter 2, right foot
78
SpO2 [%]
SpO2 [%]
95
oximeter 1, left foot

oximeter 2, right foot
78
67
02-08-12 11:36
88
90
92
94
96
98 100
68
CHAPTER 3
100
90
SpO2 [%]
SpO2 [%]
95
85
80
75
34
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70
100
200
300
time [s]
400
500
600

left foot
right foot
Frequency [%]
Frequency [%]
72
74
76
78
80
82
84 86 88
SpO2 [%]
90
92
94
96
98 100
100
SpO2 [%]
SpO2 [%]
95
90
85
80
75
100
200
300
time [s]
400
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70 72 74 76 78 80 82 84 86 88
SpO2 [%]
500
600
100
200
300
time [s]
400
36
34
32
30
28
26
24
22
20
18
16
14
12
10
8
6
4
2
0
70 72 74 76 78 80 82 84 86 88
SpO2 [%]
100
95
90
85
80
75
70
65
60
55
50
100
200
300
time [s]
90
500
92
400
94
96
500
600
98 100
600
14
12

10
Frequency [%]
Frequency [%]
90
85
80
75
70
65
60
55
50
45
40
8
6
4
2
90
92
94
96
98 100
0
70 72
74
76
78
80
82
84
86
88
90
92
94
96
98 100
SpO2 [%]
Figure 3.4A-D Figure 3.4A&C show two sets of 10 minutes of dual SpO2-measurement in two different patients. Both figures show large desaturations registered by both pulse oximeters. Figure 3.4A
show SpO2 values obtained by equal brands; Figure 3.4C show SpO2 values obtained by different brands.
Figure 3.4B&D show the frequency histograms for proportion of time spent at each saturation value for
the complete 45 minutes including the 10 minutes of measurement shown in Figure 3.4A&C.
02-08-12 11:36
69
3.4 Discussion
Simultaneously obtained SpO2 measurements with new-generation motion resistant
pulse oximeters from both feet of nine ELBW infants differ from each other (median (range)
SpO2 was 2.0% (0 to 26%)), both for equal and different brands. The differences found between the pulse oximeter measurements varied over time within infants for all three tested
brands. These differences may have an effect on the outcome of extremely preterm infants
because the oxygenation of this vulnerable patients has a narrow therapeutic range. The
variances of the fluctuations of the three pulse oximeter brands were not significantly different. Since it is unlikely that the SpO2 values in both feet of these stable infants are really
different, our results suggest the handling of the pulse oximeter in clinical practice influences the performance of the pulse oximeter the most.
When studying pulse oximetry in clinical practice there is a lack of continuous knowledge
about the real SaO2 values of the patient. Because the actual SaO2 value was not known,
there was no gold standard. Consequently, it is impossible to tell whether both, one, or no
pulse oximeters provided SpO2 values that were matching the actual SaO2 values. Thus,
based on the set-up of this study, it is impossible to judge which of the three brands performed best in terms of correct measurement, or other characteristics (e.g., easy to handle,
time to first measurement). However, finding the best performing pulse oximeter was not
the goal of the study. An earlier performed review by Giuliano et al.25 already mentioned
that they did not find evidence for differences in performances between pulse oximeter
brands. Our results confirm their conclusion by showing that the variances of the difference of the three pulse oximeter brands did not significantly differ. We aimed to evaluate
the performance of new-generation pulse oximeters in ELBW infants in clinical practice.
Knowledge about the usefulness of pulse oximeters in ELBW infants helps us to improve
strategies for ventilation, and patient care in general.
A limitation of this study was the different averaging time between pulse oximeter brands.
It has been shown that the averaging time of pulse oximeter software influences the saturations values on the monitor.23, 32 A shorter averaging time makes the pulse oximeter more
sensitive for changes in SpO2 , but increases the chance of a false alarm due to artefacts.
However, the averaging time was comparable (between 2 to 5 seconds) for all pulse oximeters. Above that, this limitation reflects a problem that healthcare professionals are confronted with in clinical practice.
Variation in pulse oximetry readings is inherent to pulse oximetry, and has been described
before.3, 8, 9, 17-23 However, especially in ELBW infants these (large) non-consistent deviations
are, regarding the risks for hypoxia and hyperoxia, undesirable and should be minimised.
The causes for the deviations as seen in this study are in literature often indicated as multifactorial, and could find their origin either in the patient, the software of the pulse oximeter, and/or the probe position and placement. The difficulty in clinical practice is that the
actual cause of the deviations cannot always be identified.
02-08-12 11:36
70
CHAPTER 3
In theory, SaO2 levels in both feet of a healthy infant are similar, because both feet are in
the post-ductal region of the body.33-36 Therefore, the origin of differences in simultaneously
obtained SpO2 values must be found in the pulse oximetry measurement. It is conceivable
that there are differences between extremities of the infant such as movement and low
blood perfusion. The differences between extremities could be the reason for the deviating
SpO2 values in Figure 3.1. The well known motion artefacts in pulse oximetry are mainly
due to irregular venous blood flow during motion. Pulse oximeters assume that there is
only pulsating blood flow in the arterial compartment; the irregular flow of the venous
blood influences the SpO2 measurement. A low blood perfusion may influence the SpO2
measurement due to a poor signal quality caused by a lack of clear pulsations.
Poor positioning of the probes can also cause differences between pulse oximeter measurements, which may be the case in Figure 3.2. False readings due to inappropriate probe
placement were first mentioned in the 1990s.37, 38 Although manufacturers since then have
been working on the improvement of the software to recognise sensor displacement, appropriate probe placement is still a requirement to obtain reliable pulse oximeter readings. In
this study the probes were brand-new and positioned carefully, to obtain high performance.
In daily practice, it is possible that deviations of pulse oximeters may even be larger due to
poor positioning of the probes, and because of using probes longer than recommended.39
To prevent movement of the probe, the probe should be positioned firmly and close to the
skin, preferably with an extra cover to prevent influence of ambient light.30, 31 However, the
weak skin of ELBW infants often hampers the firm positioning of probes.
A challenge for pulse oximeter manufacturers is to deal with moments of movements, low
blood perfusion, and poor sensor placement leading to unreliable measurement. On the
one hand manufacturers want to provide a SpO2 value as often as possible, on the other
hand there are moments where the measurement is not reliable enough to provide SpO2
values. Because the software used by the manufacturers differs, there are moments that
pulse oximeters of one brand do not provide SpO2 values while others do, and vice versa.
The challenges in the development of pulse oximetry has led to a relative wide accuracy
range. Technical specifications for pulse oximeters used for neonates suggest an accuracy of
2 or 3 digits (1 SD) between 70 and 100% saturation. The accuracy under 70% saturation is
often unspecified. The 2 or 3 digits (1 SD) means that the standard deviation is equal to 2
or 3 digits. In other words, in case of a standard deviation of 2 digits, an SpO2-value of 90%
is with 0.68 certainty within 90 2 digits (i.e. 88 to 92%), and with 0.95 certainty within 90
4 digits (i.e. 86 to 94%), see Figure 3.5. The latter range is of the same magnitude as the latest
advised allowed range for the SpO2 in ELBW infants (i.e. 85 to 93%).40 In case of a standard
deviation of 3%, the accuracy range of pulse oximeters is even wider. This relative wide accuracy range is, regarding the risk for hypoxia and hyperoxia in ELBW infants, undesirable,
and should be improved.
02-08-12 11:36
71
SpO2 =
88%
Frequency (rate of occurrence) of each value
86%
90%
- 1 SD
94%
+ 1 SD
+ 2 SD
2 digits
2 digits
- 2 SD
92%
mean
68%
95%
Figure 3.5 Accuracy range of pulse oximetry with a standard deviation (SD) of 2 digits.
Besides the possible improvements that should be made in pulse oximetry (probes) by
the manufacturers, healthcare professionals themselves can play an important role in
optimizing pulse oximetry. They are the users of the technique and have to interpret pulse
oximeter measurements. Although the basics of pulse oximetry are described extensively in
literature, it has been shown that knowledge about the working principles, the limitations,
and the correct method for interpretation of pulse oximetry among healthcare professionals still needs improvement.41-48 Therefore, the working principles of pulse oximetry,
and the interpretation of the SpO2 values should be educated extensively and frequently
among physicians and nurses.
To conclude, in clinical practice, simultaneously obtained pulse oximeter measurements
from both feet of ELBW infants differ from each other, both for equal and different brands.
No systematic deviations between pulse oximeter measurements were found in all three
tested brands. Our results suggest that not the brand, but the handling of the pulse oximeter
in clinical practice, like the place and positioning of the probe, influences the performance
of the pulse oximeter the most. Although the introduction of pulse oximetry into the NICU
has led to a major improvement in patient monitoring, the limitations of pulse oximetry
emphasise the need for the improvement and/or the development of new techniques for
both reliable and user-friendly oxygenation measurements in ELBW infants.
02-08-12 11:36
72
CHAPTER 3
3.5 References
1. Sola A, Rogido MR, Deulofeut R. Oxygen as a
neonatal health hazard: call for detente in clinical
practice. Acta Paediatr. 2007;96(6):801-812.
3. Sahni R, Gupta A, Ohira-Kist K, Rosen TS. Motion
resistant pulse oximetry in neonates. Arch Dis Child
Fetal Neonatal Ed. 2003;88(6):F505F508.
4. Webb RK, Ralston AC, Runciman WB. Potential errors
in pulse oximetry II. Effects of changes in saturation
and signal quality*. Anaesthesia. 1991;46(3):207-212.
5. Trivedi NS, Ghouri AF, Shah NK, Lai E, Barker SJ.
Effects of motion, ambient light, and hypoperfusion
on pulse oximeter function. Journal of Clinical
Anesthesia. 1997;9(3):179-183.
6. Lawless ST. Crying wolf: false alarms in a pediatric intensive care unit. Critical Care Medicine.
1994;22(6):981-985.
7. Trivedi NS, Ghouri AF, Shah NK, Lai E. Effects
of motion, ambient light, and hypoperfusion on
pulse oximeter function. Survey of Anesthesiology.
1998;42(3):168.
8. Brockway J, Hay Jr WW. Prediction of arterial partial
pressure of oxygen with pulse oxygen saturation
measurements. Journal of Pediatrics. 1998;133(1):63-66.
9. Paky F, Koeck CM. Pulse oximetry in ventilated preterm newborns: reliability of detection of hyperoxaemia and hypoxaemia, and feasibility of alarm settings.
Acta Paediatr. 1995;84(6):613-617.
10. Shiao SYPK. Effects of Fetal Hemoglobin on Accurate Measurements of Oxygen Saturation in
Neonates. Journal of Perinatal & Neonatal Nursing.
2005;19(4):348-361.
11. Dawson J, Davis P, ODonnell C, Kamlin C, Morley C.
Pulse oximetry for monitoring infants in the delivery
room: a review. Arch Dis Child Fetal Neonatal Ed.
2007;92(1):F4-7.
12. Zijlstra W, Buursma A, Meeuwsen-Van Der Roest
W. Absorption spectra of human fetal and adult
oxyhemoglobin, de-oxyhemoglobin, carboxyhemoglobin, and methemoglobin. Clinical Chemistry.
1991;37(9):1633-1638.
13. Gressens P, Rogido M, Paindaveine B, Sola A.
The impact of neonatal intensive care
practices on the developing brain.
14. Saugstad O. Bronchopulmonary dysplasia-oxidative

stress and antioxidants. Seminars in Neonatology.
2003;8(1):39-49.
British Medical Journal. 2001;84(2):F106-F110.
16. Saugstad O. Oxidative stress in the newborn-a
30-year perspective. Biology of the Neonate.
2005;88(3):228-236.
17. Grieve S, McIntosh N, Laing I. Comparison of two
different pulse oximeters in monitoring preterm
infants. Critical Care Medicine. 1997;25(12):2051-2054.
18. Bohnhorst B, Peter CS, Poets CF. Pulse oximeters
reliability in detecting hypoxemia and bradycardia: comparison between a conventional and two
new generation oximeters. Critical Care Medicine.
2000;28(5):1565-1568.
19. Bohnhorst B, Peter CS, Poets CF. Detection of
hyperoxaemia in neonates: data from three new
pulse oximeters. Arch Dis Child Fetal Neonatal Ed.
2002;87(3):F217-219.
20. Hay W, Rodden D, Collins S, Melara D, Hale K,
Fashaw L. Reliability of conventional and new
pulse oximetry in neonatal patients. J Perinatol.
2002;22(5):360-366.
21. Gerstmann D, Berg R, Haskell R, Brower C, Wood K,
Yoder B, et al. Operational evaluation of pulse
oximetry in NICU patients with arterial access.
J Perinatol. 2003;23(5):378-383.
22. Workie F, Rais-Bahrami K, Short B. Clinical use of
new-generation pulse oximeters in the neonatal intensive care unit. American J Perinatol. 2005;22(7):357-360.
23. Ahmed S, Rich W, Finer N. The Effect of Averaging
Time on Oximetry Values in the Premature Infant.
Pediatrics. 2010;125(1):e115-e121.
24. Johnston E, Boyle B, Juszczak E, King A, Brocklehurst
P, Stenson B. Oxygen targeting in preterm infants
using the Masimo SET Radical pulse oximeter. Arch
Dis Child Fetal Neonatal Ed. 2011;96(6):F429-433.
Oximetry in the Care of Critically Ill Patients.
In: AACN; 2005. p. 26-37.
26. Philip AGS. The evolution of neonatology.
27. Norman M. Incidence of and Risk Factors for
Neonatal Morbidity after Active Perinatal Care:
Extremely Preterm Infants Study in Sweden
(EXPRESS). Acta Paediatr. 2010;99(7):978-992.
02-08-12 11:36
28. Fellman V, Hellstrm-Westas L, Norman M, Westgren

M, Klln K, Lagercrantz H, et al. One-year survival
of extremely preterm infants after active perinatal care in Sweden. Obstetric Anesthesia Digest.
2010;30(1):22-23.
29. Stark AR. Levels of neonatal care. Pediatrics.
2004;114(5):1341-1347.
30. Ralston A, Webb R, Runciman W. Potential errors
in pulse oximetry III: Effects of interference, dyes,
dyshaemoglobins and other pigments*. Anaesthesia.
1991;46(4):291-295.
31. Fouzas S, Priftis KN, Anthracopoulos MB.
Pulse Oximetry in Pediatric Practice. Pediatrics.
2011;128(4):740-752.
Oximetry in the Care of Critically Ill Patients.
American Journal of Critical Care. 2005;14(1):26-37.
33. Baquero H, Alviz R, Castillo A, Neira F, Sola A.
Avoiding hyperoxemia during neonatal resuscitation:
time to response of different SpO2 monitors.
Acta Paediatr. 2011;100(4):515-518.
34. Mariani G, Dik PB, Ezquer A, Aguirre A, Esteban ML,
Perez C, et al. Pre-ductal and Post-ductal O2
Saturation in Healthy Term Neonates after Birth.
35. Meier-Stauss P, Bucher H, Hrlimann R, Knig V,
Huch R. Pulse oximetry used for documenting
oxygen saturation and right-to-left shunting immediately after birth. European Journal of Pediatrics.
1990;149(12):851-855.
36. Vento M. Titrating oxygen needs in the very preterm
newborn in the delivery room. Journal of NeonatalPerinatal Medicine. 2010;3(3):161-169.
37. Southall D, Samuels M. Inappropriate sensor application in pulse oximetry.
The Lancet. 1992;340(8817):481-482.
38. Poets CF, Seidenberg J, Von Der Hardt H. Failure of
pulse oximeter to detect sensor detachment.
The Lancet. 1993;341(8839):244.
39. Erler T, Avenarius S, Wischniewski E, Schmidt K,
Klber H. Longevity of Masimo and Nellcor pulse
oximeter sensors in the care of infants. J Perinatol.
2003;23(2):133-135.
40. Finer N, Leone T. Oxygen Saturation Monitoring for
the Preterm Infant: The Evidence basis for Current
Practice. Pediatric Research. 2009;65(4):375-380.
73
41. Elliott M, Tate R, Page K. Do clinicians know how to

use pulse oximetry? A literature review and clinical
implications. Australian Critical Care.
2006;19(4):139-144.
42. Popovich DM, Richiuso N, Danek G. Pediatric
health care providers knowledge of pulse oximetry.
Pediatric Nursing. 2004;30(1):14-20.
43. Rodriguez LR, Kotin N, Lowenthal D, Kattan M. A
study of pediatric house staffs knowledge of pulse
44. Attin M, Cardin S, Dee V, Doering L, Dunn D,
Ellstrom K, et al. An educational project to improve
knowledge related to pulse oximetry. American
Journal of Critical Care. 2002;11(6):529-534.
45. Fouzas S, Politis P, Skylogianni E, Syriopoulou T,
Priftis KN, Chatzimichael A, et al. Knowledge on
Pulse Oximetry Among Pediatric Health Care
Professionals: A Multicenter Survey. Pediatrics.
2010;126(3):e657-662.
46. Stoneham M, Saville G, Wilson I. Knowledge about
pulse oximetry among medical and nursing staff.
The Lancet. 1994;344(8933):1339-1342.
47. Kruger P, Longden P. A study of a hospital staffs
knowledge of pulse oximetry. Anaesthesia and
Intensive care. 1997;25(1):38-41.
48. Solberg MT, Hansen TWR, Bjrk IT. Nursing
assessment during oxygen administration in
ventilated preterm infants. Acta Paediatr.
2010;100(2):193-197.
02-08-12 11:36
74
02-08-12 11:36
CHAPTER 4
Manual adjustments
of the inspired
oxygen fraction
in extremely low
birth weight infants
A.C. van der Eijk, J. Dankelman, S. Schutte, H.J. Simonsz, B.J. Smit
02-08-12 11:36
76
CHAPTER 4
In current neonatal intensive care, oxygenation of preterm infants is controlled manually by the healthcare professionals. This chapter evaluates the control behaviour of
physicians and nurses, and is based on the article An observational study to quantify
manual adjustments of the inspired oxygen fraction in extremely low birth weight infants
(published in Acta Paediatrica, Vol 101:3, 2012). The focus lies on the quantification of
manual FiO2 adjustments, and the relation between these adjustments with SpO2 values
and bedside care in extremely low birth weight infants. The findings of this research help
us to identify differences between manual and (semi-)automatic control of oxygenation,
and to improve manual FiO2 adjustment.
OBJECTIVE To quantify manual adjustments in the FiO2 performed by neonatal intensive care unit personnel in extremely low birth weight infants, in relation to SpO2 and
bedside care.
02-08-12 11:36
Manual adjustments of the inspired oxygen fraction in ELBW infants
77
4.1 INTRODUCTION
To prevent tissue damage due to hypoxemia or hyperoxia, oxygenation of preterm infants
needs to be monitored closely.1 Although there are multiple parameters influencing oxygenation, one of the leading parameters for assessment of oxygenation is the oxygen saturation measured by pulse oximetry (SpO2 ).2 When oxygenation of preterm infants is controlled manually by the nursing staff, SpO2 is frequently outside the target range.3-5
To improve the outcome of preterm infants while oxygenation is controlled manually, two
studies focused on the development of protocols to standardise when, why, and how FiO2
should be adjusted.6, 7 The protocol described by one of these studies7 was implemented
in several neonatal intensive care units (NICUs) worldwide, and actually showed a
reduction in the incidence of retinopathy of prematurity.
Several groups have been working on the development of (semi-)automatic control
devices for oxygenation, to prevent too high and too low blood oxygen levels. These devices
adjust the fraction of inspired oxygen (FiO2 ) supplied to the patient automatically when
SpO2 deviates from the target.8-14 The results of the developed devices are promising, showing an increased stability of SpO2 . However, these devices are not widely used, and long
term outcome effects are not known yet.15
None of the studies mentioned above quantified the step size of the performed manual
FiO2 adjustments. Detailed knowledge about the actual behaviour of healthcare professionals with respect to manual control of oxygenation is still lacking.
Therefore, the aim of this study was to quantify manual FiO2 adjustments performed by
physicians and nurses in extremely low birth weight (1000 g.; ELBW) infants, in relation
to SpO2 , and bedside care. We expected that the step size of FiO2 adjustments chosen by
the physicians and nurse depends on SpO2 values. A single-centre observational study was
designed and manual FiO2 adjustments were related to SpO2 levels and bedside care.
To prevent tissue damage due to hypoxemia

or hyperoxia, oxygenation of preterm infants
needs to be monitored closely.
02-08-12 11:36
78
CHAPTER 4
4.2 Methods
This single-centre study was performed at the level-III16 NICU of the Erasmus Medical
Centre - Sophia Childrens Hospital in Rotterdam, the Netherlands. The NICU has a
capacity for 30 neonates divided over three rooms. Approval was obtained from the ethical
review board of the Erasmus Medical Centre. Informed written consent by the parents was
provided for all included patients prior to inclusion.
4.2.1 Patients
Patients were eligible for this study when gestational age (GA) was 28 weeks, birth weight
(BW) was 1000 g., and when there was a need for supplemental oxygen therapy in the first
two weeks of life. Patients were excluded if they were scheduled for transfer to another
unit, planned for surgery, or when death was expected within 2 days. For each patient in
the study, SpO2 was recorded continuously (1 Hz) for 3 days; simultaneous on-site video
recordings were made to document bedside care by the NICU staff, FiO2 , and preset alarm
limits.
The choice for using cameras had two advantages. Firstly, there was no researcher
present in the NICU to collect observations by hand. The presence of a researcher could
have disturbed the nursing staff, and their behaviour. Secondly, the cameras made it possible to make observations for 24 hours per day continuously. Recording started when FiO2
was >21% and the postnatal age of the patient was between 2 and 14 days. When the need
for supplemental oxygen ended, the recordings still were continued. During the recordings,
the nurse to patient ratio varied between 1:2 and 1:3.
According to standard practice, all patients were cared for in an incubator and connected
to a multi-parameter patient monitor (Hewlett Packard Viridia monitor, CMS M1046A).
SpO2 was obtained by a pulse oximeter (disposable probe: Nellcor OxiMax, MAX-N <3 kg
- module: Hewlett Packard, M1020A) placed on a foot or the left hand of the patient. The
experimental set-up consisted of three small cameras with motion detection (AXIS communications, type: 221 Day & Night Network Camera) and two laptops (Figure 4.1). The
cameras were attached to a column next to the incubator with magnets. One camera was
aimed at the environment of the incubator to record bedside care by the NICU personnel.
Because the ventilators did not have a digital output for FiO2-data collection, a second
camera recorded the ventilator to monitor manual adjustments in FiO2 . The third camera
focused on the patient monitor to obtain alarm limits for SpO2 . Motion detection was used
to activate cameras when motion around the incubator was detected. The PCs recorded the
videos and the SpO2 from the patient monitor (software: Dataplore from Ixellence GmbH,
Wildau, Germany). Prior to the study, the personnel of the NICU were informed about the
video recording without discussion of the study objectives.
02-08-12 11:36
Nurse
SpO2
ECG
PC with
PDM
software
79
Researcher
HP VIRIDIA
PATIENT
MONITOR
Camera 3
Laptop 1:
Dataplore
software
Any
ventilator
Camera 2
Laptop 1:
Camera
software
Camera 1
Figure 4.1 Schematic overview of the experimental set-up. The left part (outlined by dashed light
line) consisted of the standard equipment for neonatal intensive care. The right part of the experimental set-up (outlined by dashed dark line) consisted of equipment necessary for study purposes.
The laptops and cameras were placed such that NICU personnel were not disturbed.
4.2.3 Data collection

Before recording was started, time settings of the laptops were synchronised (one-second
accuracy). Parameters were obtained by Dataplore and exported to .txt files. There was
only a start time and sample time provided (sample time = 1.032 sec). Prior to analysis all
data was interpolated to 1 Hz. The recorded videos were calibrated manually with the collected data using the timestamp in the video screens. After calibration, the videos were
annotated manually by the researchers. For each second, FiO2 , alarm limits, and detailed
bedside care were registered. Observations were discussed with neonatologists to interpret
the performed actions. Obtained data were used for descriptive analysis.
4.2.4 Target levels for SpO2
During the study, the protocol for preterm infants with a postmenstrual age (PMA) <32
weeks prescribed a target SpO2 of 88 to 94% when supplemental oxygen was supplied (FiO2
>21%) or 88 to 100% for patients in room air (FiO2 = 21%). An alarm sounded when SpO2
was outside these limits for more than 20 seconds. An extra loud low saturation alarm
was set for SpO2 <80%. The nursing staffs task was to keep SpO2 within the target range.
There was, besides their normal education, neither a protocol or specific training for nurses
describing when and how FiO2 should be adjusted.
02-08-12 11:36
80
CHAPTER 4
When SpO2 levels were <88% or >94% for >20 seconds, these time periods were defined as
Lowsats and Highsats respectively. Lowsats with a minimum SpO2 <70% were studied in
more detail and referred to as deep Lowsats. It was expected that the alarm limits for SpO2
would be kept reasonably well. However, in the course of the study adherence to the protocol
for alarm limits was found to be low. Motivations for deviating from the protocol for alarm
limits were not written in patient records. Therefore, it proved very difficult to determine
why doctors or nurses had deviated from the alarm limit protocol, and whether they had, in
doing so, pursued a target SpO2 outside of the 88 to 94% range. An unknown target SpO2
complicates interpretation the adjusting behaviour of the physicians and nurses. Therefore,
the relation between step size of FiO2 adjustments and corresponding SpO2 values were
evaluated in detail only when alarm limits were set according to protocol.
4.2.5 FiO2 adjustments
FiO2 adjustments with a step size 15% (e.g., FiO2 increased from 25 to 50%) were defined
as large adjustments. FiO2 adjustments were considered single when there were no
other adjustments made in the two minutes before and after the adjustment. Furthermore,
FiO2 adjustments were divided in closed incubator (closed-I), i.e. the incubator was completely closed two minutes before and after the adjustment, or open-I, i.e. the incubator
was open(ed) in the two minutes before and/or after the adjustment. To determine the
reaction of the patient on FiO2 adjustments, the delta SpO2 was determined. This delta
SpO2 was calculated by subtracting the median SpO2 calculated over the period from the
30th until the 60th second after an FiO2 adjustment from the median SpO2 of the final
30 seconds before the adjustment.
4.3 Results
During a 7-month period, 833 hours of monitored parameters and video recordings were
collected.
4.3.1 Patients
In total 12 patients were included with a median (range) GA of 26 2/7 (24 2/7 - 28) weeks, BW
of 760 (545 - 935) grams, and PNA of 4 (2 - 12) days. Detailed information about the patients
is shown in Appendix A. Recruitment failed in 46 patients because they did not need supplemental oxygen therapy in the first two weeks of life (27), parental refusal (8), imminent
death (6), or the experimental set-up was unavailable (5). In one patient (#11), data recording
was stopped after 48 hours because of transfer to another unit.
4.3.2 Manual FiO2 adjustments
Due to changes in position of equipment of interest and incomplete video recordings, FiO2
was available for analysis during 726 hours. Of these 726 hours, patients were either on invasive
ventilation (572 hours) or CPAP (155 hours). Overall, supplemental oxygen therapy was provided in 634 hours; hence, patients were on room air in 92 hours. The staff, predominantly nurses,
adjusted FiO2 2160 times (Figure 4.2). FiO2 was increased 851 times and decreased 1309 times.
02-08-12 11:36
81
The time between two successive FiO2 adjustments varied from a few seconds to more
than seven hours (median = 4.7 min). The average number of adjustments per hour for each
infant separately varied between 0.7 and 5.2 (Appendix A). In 46% of the FiO2 adjustments,
the adjustment was single, i.e. there was no other adjustment in the two minutes before and
after the adjustment. Nine percent of the FiO2 adjustments were made within 20 seconds
after the previous FiO2 adjustment, i.e. within the time it takes for a ventilator to change
the FiO2 level in the gas mixture supplied to the patient.
21
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
1
0
0
0
0
0
1
2
0
2
2
1
1
2
5
4
2
3
16
0
27
15
6
8
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
2
1
4
0
1
0
1
2
1
3
2
2
8
5
5
3
5
10
3
9
7
4
0
38
33
10
13
21
4
3
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
1
2
2
1
0
3
3
0
0
0
2
2
0
2
0
0
18
9
1
4
1
4
0
0
0
1
30
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
1
0
0
1
1
0
0
0
2
4
1
0
3
2
7
2
3
0
4
14
4
2
3
4
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
1
1
3
1
0
1
6
5
1
0
6
5
14
3
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
2
0
2
0
2
0
1
4
2
9
2
0
1
0
0
1
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
2
0
0
0
0
1
0
0
0
0
3
0
1
1
0
2
1
1
0
2
15
4
2
4
1
0
3
12
15
7
19
5
3
1
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
1
3
0
2
0
0
9
4
4
3
4
8
3
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
1
1
0
0
1
0
0
0
1
2
0
0
2
9
2
1
3
0
9
2
2
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
1
0
2
1
0
0
0
2
0
0
1
4
6
0
4
3
1
4
3
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5
0
0
3
0
1
0
2
0
0
4
0
1
3
5
7
0
3
1
1
1
2
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
1
0
5
1
0
0
1
3
2
1
2
1
0
6
15
4
5
23
1
1
4
3
8
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
1
0
0
0
0
0
0
0
0
2
1
1
0
0
0
1
1
0
0
1
5
0
1
4
5
1
0
0
1
0
1
0
0
0
0
0
0
0
0
40
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
2
0
0
0
0
6
0
1
2
1
2
1
0
0
0
1
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
0
0
0
1
2
2
1
0
1
0
0
0
0
0
2
3
1
1
0
2
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
1
0
1
1
0
0
1
3
0
0
0
3
1
5
1
1
3
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
4
1
1
2
0
0
2
4
3
2
10
4
0
0
1
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
1
0
1
2
0
1
0
0
2
2
1
0
2
0
0
0
0
0
0
1
1
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
1
0
0
2
0
1
0
0
7
1
1
2
1
1
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
1
1
0
0
0
0
1
0
1
0
1
2
0
2
4
5
2
0
0
0
3
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
3
0
2
0
1
1
0
3
1
1
2
1
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
1
0
3
1
0
1
1
5
0
0
0
1
0
2
6
2
1
4
4
1
1
3
8
1
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
1
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
1
1
0
0
0
1
0
0
1
0
2
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
3
1
1
0
0
0
0
0
0
2
0
1
0
0
1
1
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
0
0
0
1
3
0
2
0
0
0
1
0
0
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
2
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
4
0
0
1
0
0
1
2
3
1
4
1
0
2
0
1
0
1
0
0
0
0
0
0
0
1
0
0
0
0
1
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
1
0
0
0
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
1
0
0
0
0
1
1
0
0
0
0
0
2
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
3
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
1
1
1
0
0
0
0
0
1
0
1
1
0
1
1
0
0
1
3
2
0
6
0
0
0
0
1
3
2
2
0
0
1
1
0
0
1
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
1
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
60
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
4
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
1
0
0
0
1
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
1
0
1
2
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
70
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
0
0
1
0
0
1
0
2
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
4
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
0
0
0
0
0
1
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Occurrence
= 1-5 times
= 6-10 times
= 11-15 times
= >15 times
80
90
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
1
3
0
3
1
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
2
0
1
0
0
0
0
1
4
2
1
1
1
5
0
2
4
5
10
7
0
19
19
12
5
1
0
6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
1
0
0
0
1
0
0
1
4
1
0
3
5
12
5
6
0
28
18
2
2
1
4
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
1
1
1
1
2
2
14
4
6
8
0
39
7
8
3
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1
0
0
0
0
0
0
2
0
3
3
1
19
6
6
15
7
0
40
19
8
9
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
2
1
0
0
0
0
0
0
0
2
0
0
0
0
1
2
5
5
1
5
5
8
0
26
11
7
21
30
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
2
1
0
0
0
1
0
7
3
2
4
6
12
9
0
20
23

40
50
60
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
0
1
0
1
0
2
0
0
4
2
7
4
9
0
19
70
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0
0
0
1
0
0
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
0
4
6
2
9
4
7
12
4
17
9
0
FiO2 [%] after adjustment
80
90
100
In 24% of the increases in FiO2 , and in 47% of the decreases, the FiO2 adjustment was
closed-I, i.e. the incubator was closed for at least two minutes before and after the adjustment. In 11% of the increases, and in 6% of the decreases, the incubator was closed before
the adjustment, but opened within two minutes after the adjustment. In 62% of the increases, and in 43% of the decreases, the incubator was open(ed) in the two minutes before
the adjustment. In 3% of the increases, and in 4% of the decreases, it was not possible to
determine from the video recordings whether the incubator was open or closed.
100
FiO2
before adjustment
adjustment
O2
[%][%]before
Figure 4.2 Overview of all FiO2 adjustments, represented by a square. The intensity of the colour
corresponds to the number of occurrences. The horizontal axis indicates FiO2 before the adjustment;
the vertical axis indicates the FiO2 after the adjustment. This figure shows the wide variety in step
sizes used for manual FiO2 adjustments in relation to the FiO2 levels before and after the adjustments.
Both for high and low FiO2 levels, large and small step sizes are used. Downwards adjustments
(below the diagonal) occur more often, with smaller step sizes than upwards adjustments.
02-08-12 11:36
82
CHAPTER 4
The median (range) for the step size of all increases was 5% (1 to 65%). For all decreases
the step size was -3% (-1 to -65%). Figure 4.3 shows that the large variation in the step size
of FiO2 adjustments was present in all infants. For the closed-I adjustments, the median
(range) for increases, and respectively decreases was 2% (1 to 20%) and -2% (-1 to -32%). For
the open-I adjustments, the step size was 5% (1 to 65%) for increases, and -4% (-1 to -65%) for
decreases. These open-I adjustments differed significantly from the closed-I adjustments
(Mann-Whitney U-test p <<1, = 0.05).
step size FiO increase all babies
Step size
[%]
step
sizeFiO2
FiO increase
increase
Patient
number all babies
step size FiO
decrease
Step
size FiO2
[%]
stepsize
FiO decrease
decrease
Patient number
Figure 4.3A&B Distribution of step size for manual FiO2 adjustments plotted for each infant
separately. Figure 4.3A presents the increases in FiO2, Figure 4.3B plots represent the decreases.
The boxes are built-up of the median (central line), and the 25th and 75th percentiles (the edges).
The outliers are plotted individually (+). These figures illustrate that the wide variety in step sizes used
for FiO2 adjustments occurs in all included infants.
02-08-12 11:36
83
4.3.2.1 Large manual FiO2 adjustments

For 136 of the 2160 FiO2 adjustments, the step size was 15%. These relatively large
adjustments were seen in all subjects and were performed by multiple nurses and doctors.
In 6 cases, the person who made the adjustment could not be identified; the other 130
adjustments were performed by 47 different staff members.
The 136 large adjustments included 88 increases and 48 decreases. Increases were
typically performed before or during medical or nursing procedures of the infant, whereas
decreases were mainly used to return the FiO2 to the baseline (FiO2 level before the medical or nursing procedure). Forty large adjustments (28 increases, 12 decreases) were single,
i.e. without any other adjustment in the two minutes before and after the adjustment.
Ten of the 136 large adjustments (4 increases, 6 decreases) were defined as closed-I.
In 10 of the 88 increases, the median SpO2 of the 30 seconds prior to the adjustment was
below 60%; in 59 increases the SpO2 was between 60 and 80%. Hence, in 19 increases, the
SpO2 value did not seem to be the motive for the large step size. Of all 48 downward FiO2
adjustments with a large step size, median SpO2 was >94% in 19 cases, and between 80 and
94% in 20 situations. In nine cases SpO2 was below 80%.
4.3.3 SpO2 levels
In the 726 hours where FiO2 and SpO2 were known, 177 deep Lowsats, i.e. SpO2 levels
<88% for >20 seconds with a minimum SpO2 <70%, were seen. The median (range) duration
of these deep desaturations was 268 (23 - 915) seconds. In 6 of these deep Lowsats, the video
recordings of the incubator were not clear enough for analysis. In 111 deep Lowsats the desaturation started while the incubator was open(ed) in the two minutes before the start of
the desaturation, hence in 60 deep Lowsats, the desaturation was not caused by medical or
nursing procedures. Of these 60 deep Lowsats, in 53 the incubator was opened during the
desaturation, thus a physician or nurse reacted on the desaturation. In 136 deep Lowsats
there was at least one FiO2 adjustment performed, hence 41 deep Lowsats recovered without any FiO2 adjustments.
Each delta SpO2 , i.e. the median SpO2 over the period from the 30th until the 60th second
after an FiO2 adjustment subtracted from the median SpO2 of the final 30 seconds before
the adjustment, is plotted against the step size of the corresponding FiO2 adjustment in
Figure 4.4. This figure shows that delta SpO2 varies for equal step sizes. Because of this
variation, and because manual FiO2 adjustments are often accompanied by other nursing
tasks or sequential adjustments in FiO2 , it was decided to focus on the closed-I and single
adjustments to study the reaction of the patient on the FiO2 adjustment. However, no relation between the step size of FiO2 adjustments and the delta SpO2 was found for these
adjustments either.
02-08-12 11:36
CHAPTER 4
Delta SpO2 [%]
84
Occurrence
= 1-5 times
= 6-10 times
= 11-15 times
= > 15 times
Step size FiO2 [%]
Figure 4.4 Overview of the delta SpO2 (median SpO2 over the period from the 30th until the 60th
second after an FiO2 adjustment subtracted from the median SpO2 of the final 30 seconds before the
adjustment) plotted against the step size of the corresponding FiO2 adjustments. This figure shows the
wide variation in reaction of the patient on FiO2 adjustments with an equal step size.
FiO2 , SpO2 , and alarm limits were available in 654 hours (FiO2 >21% in 565 hours). The
alarm limits were set in accordance to the protocol 64% (FiO2 >21%) and 84% (FiO2
= 21%) of the time. The relation between SpO2 levels before an FiO2 adjustment and
the step size of the FiO2 adjustment was only evaluated in detail for the time
period that patients received supplemental oxygen therapy and the alarm limits were set
according to protocol (64% of 565 hrs = 362 hrs). For similar median SpO2 values before an
FiO2 adjustment, the step sizes varied widely (e.g., for median SpO2 = 80%, FiO2 step size
varied from -30 to 20%) (Figure 4.5). No correlation could be found between step size and
preceding SpO2 values.
In the same 362 hours, 1818 Lowsats (SpO2 <88% for >20 s.) and 2784 Highsats (SpO2
>94% for >20 s.) were observed. The duration of the Lowsats and Highsats varied
widely; about 80% lasted for 2 minutes, while some of the Lowsats lasted up to 20 minutes.
Some Highsats even lasted up to one hour. The total time spent in Highsats varied
between infants from 7.6 to 41.3%. For Lowsats, the percentage was between 5.6 to
43.0% (Appendix A, Table A2). In total, SpO2 was outside of alarm limits for 46% of the time
(SpO2 was >94% in 30% of the time and SpO2 was <88% in 16% of the time).
02-08-12 11:36
Median SpO2 [%]
85
Occurrence
= 1-5 times
= 6-10 times
= 11-15 times
= > 15 times
Step size FiO2 [%]
Figure 4.5 Overview of all FiO2 adjustments (for time periods when SpO2 alarm limits were 88 and
94%; represented by the two horizontal lines). The horizontal axis represents the step size of the FiO2
adjustment; the vertical axis represents median SpO2 value of the final 30 seconds prior to the FiO2
adjustment. This figure shows that FiO2 adjustments often do not correspond to the SpO2 value. A
relevant part of the adjustments is performed while SpO2 is within the target range. Also, there are some
decreases in FiO2 while SpO2 is already below target and vice versa. This illustrates that, for healthcare
professionals, SpO2 is not the (only) parameter used to determine the step size.
Of all 1818 Lowsats, 23% had a minimum SpO2 <80%, whereas 22% of the 2784 Highsats had a maximum SpO2 of 100%. Most of the Lowsats (86%) and most of the Highsats
(88%) recovered without an FiO2 adjustment (Figure 4.6). When FiO2 was adjusted during
a Lowsat or Highsat, the median duration of the Lowsats preceding an FiO2 adjustment
was 79 s (max 11 minutes); the median duration of the Highsats preceding an FiO2
adjustment was 59 s (max 25 minutes).
Step size and frequency of manual FiO2 adjustments did not significantly differ for the time
periods that alarm limits were not set according to the protocol compared to the periods
when alarm limits were set according to the protocol (Mann-Withney U-test p=0.3070,
=0.05).
02-08-12 11:36
86
CHAPTER 4
700
Lowsats or Highsats without FiO2 adjustment
Lowsats or Highsats with FiO2 adjustment
600
Occurrence
500
400
300
200
100
0
45
50
55
60
65
70
75
80
85
90
95
100
Most extreme SpO2 [%] value during Lowsat or Highsat
Figure 4.6 Occurrence of all Lowsats and Highsats lasting >20 seconds (for time periods when FiO2
>21% and SpO2 alarm limits were 88 & 94%) with or without an FiO2 adjustment during the Lowsat,
respectively Highsat. This figure shows that most of the Lowsats and Highsats recover without an FiO2
adjustment. The further away the SpO2 deviates from the target range, the more often FiO2 is adjusted
to recover SpO2.
4.4 Discussion
In this single-centre observational study, manual FiO2 adjustments in ELBW infants were
quantified and examined in relation to SpO2 values and bedside care. Manual FiO2 adjustments varied widely in frequency and step size for equal levels of SpO2 in all subjects and
observed nurses. Alarm limits for SpO2 were set according to the protocol in 64% of the
time. During these periods, SpO2 values were outside the target range for 46% of the time.
The majority of the Lowsats (SpO2 <88%) and Highsats (SpO2 >94%) recovered without
any FiO2 adjustment, whereas the time needed for recovery was sometimes large. 126 of 136
FiO2 adjustments with a step size 15% and 111 of 171 desaturations <70% were associated
with medical or nursing procedures.
We were surprised to find that alarm limits for SpO2 deviated from the protocol in 36% of
the time. Within the remaining 64% of the time, SpO2 was within these alarm limits only
54% of the time. We wondered whether the protocol, prescribing limits of 88 and 94%, did
02-08-12 11:36
87
not fulfill the needs. However, the time that SpO2 was within target range is comparable
with, or better than, results of other studies. Laptook et al.4 compared two target ranges;
90 to 95% and 88 to 94%. The time SpO2 was within target range was (mean SD) 57.7
9.8% and 59.4 12.4%, respectively. Hagadorn et al.3 compared performance in 14 centres. In
these centres SpO2 was within target range 16 to 64% of the time. In the centre of Armbuster et al.17 SpO2 was within target range between 68 and 79% of the time.
The low compliance with SpO2 alarm limits has also been described before,17-20 and reveals
a tendency of healthcare professionals to accept higher SpO2 levels than the protocol prescribes. The reason for this behaviour is probably that physicians and nurses want to prevent
periods of hypoxia at the expense of hyperoxia. Other reasons for changing alarm limits,
whether justified or not, could be related to e.g., specific infant conditions, or to minimise
the number of alarms.21 Our results show that there is no significant difference in step size
of FiO2 adjustments when alarm limits were or were not set according to the protocol. This
could suggest that the actual alarm limits do not influence the manual control behaviour of
the nursing staff to a large extent.
The study was performed in one centre, and the question remains whether the results are
generalisable to other NICUs. Furthermore, because NICU personnel knew that they were
being recorded, it is not unlikely that they performed better than in normal practice.22 However, beside the fact that the time for SpO2 within target range was comparable with other
studies, there are multiple centres worldwide without a protocol for manual FiO2 adjustments. Exact numbers for the presence or absence of protocols for FiO2 adjustments are
unavailable, but Nghiem et al.20 showed that 32% of the centres in the US do not have a
protocol for SpO2 alarm limits. Which may also imply absence of protocols for FiO2 adjustment. Nghiem et al.20 also showed that the absence of protocols was associated with larger
variation in practice compared to centres with protocols.
It would be interesting to know which of the FiO2 adjustments performed by the nursing staff
were (in)sufficient to normalise SpO2 levels after hypoxemia or hyperoxia. Were, for instance,
larger and more timely adjustments more effective than smaller or delayed adjustments?
Unfortunately, due to the observational character of our study the relation between FiO2
adjustments and SpO2 levels is obscured by confounders: such as the nursing actions that were
performed simultaneously with the FiO2 adjustments. However, even for FiO2 adjustments
that were closed-I and single, there was a large variation in the SpO2 reaction on FiO2 adjustments. Moreover, most Lowsats and Highsats recovered without any FiO2 adjustment at all.
126 of the 136 FiO2 adjustments with a step size 15%, and 111 of the 171 desaturations <70%
were associated with medical or nursing procedures. Large FiO2 increases were mostly
performed before or during the procedure, while large decreases were performed after the
procedure. Physicians and nurses perform this preoxygenation to prevent hypoxia during
the care of the patient. When these large FiO2 adjustments in anticipation of procedures are
really necessary, it will be challenging to implement them in an automatic adjustment
device.
02-08-12 11:36
88
CHAPTER 4
Studies on automatic control of oxygenation show more frequent FiO2 adjustments and
smaller variability of SpO2 values for automatic control devices than in manual control.8-12, 23, 24 Yet, there are four major differences between automatic and manual control
(Figure 4.7). Firstly, nurses take into account information from various sources simultaneously. Automatic devices rely solely on a single input parameter (e.g., SpO2). Secondly,
nurses can perform multiple actions to stabilise oxygenation (e.g., suctioning), whereas automatic devices can only adjust FiO2. Thirdly, nurses take care for multiple patients,5, 25 whereas
automatic devices focus only on one patient. Finally, nurses are influenced by experience and
expertise. These characteristics are not present in automatic devices yet. To identify optimal
methods for controlling oxygenation either manually or (semi-)automatically, we recommend further development and comparative testing of protocols for manual control.
Human
factor
Patient
Other
Other
ventilator
settings
Suctio- Touchning
ing
FiO2
SpO2
Heartrate
Respiration
Skincolor
Environment
Other
Decision
making
Nurse on ward
Dedicated nurse
Automatic device
Figure
4.7 General
template
of differences
the control
the oxygenation
among
manual
control
Figure
4; General
template
of differences
in theincontrol
of theofoxygenation
among
manual
control
by
by nurses on the ward, manual control by a dedicated nurse only focussing on the oxygenation, and
nurses
on the
ward,based
manual
by value
a dedicated
automatic
control
oncontrol
the SpO2
only. nurse only focusing on the oxygenation, and automatic
control based on the SpO2 value only.
To conclude, in this single-centre study manual FiO2 adjustments varied widely in frequency
and step size for equal levels of SpO2 . When alarm limits for SpO2 were set according to the
protocol, SpO2 levels were outside of alarm limits for approximately half of the time. Deep
desaturations (SpO2 <70%), and FiO2 adjustments with a step size 15% were associated with
medical or nursing procedures. When these large step sizes for FiO2 are really necessary, it
will be challenging to implement them in an automatic adjustment device.
02-08-12 11:37
89
4.5 References
2. Solberg MT, Hansen TWR, Bjrk IT. Nursing assess-

ment during oxygen administration in ventilated

preterm infants. Acta Paediatr. 2010;100(2):193-197.
3. Hagadorn JI, Furey AM, Nghiem TH, Schmid CH,
Phelps DL, Pillers DAM, et al. Achieved versus
intended pulse oximeter saturation in infants born less
than 28 weeks gestation: The AVIOx Study. Pediatrics.
2006;118(4):1574-1582.
4. Laptook AR, Salhab W, Allen J, Saha S, Walsh M. Pulse
oximetry in very low birth weight infants: can oxygen
saturation be maintained in the desired range?
J Perinatol. 2006;26:337-341.
5. Sink DW, Hope SAE, Hagadorn JI. Nurse:patient
ration and chievement of oxygen saturation goals in
premature infants. Arch Dis Child Fetal Neonatal Ed.
2011;96(2):F93-98.
6. Wilkinson D, Andersen C. Bedside algorithms
infants: A randomised crossover Trial. Neonatology.
2008;95(4):306-310.
practice decrease the incidence of severe retinopathy
of prematurity in very low birth weight infants?

Pediatrics. 2003;111(2):339-345.
8. Bhutani VK, Taube JC, Antunes MJ, Delivoria-Papadopoulos M. Adaptive control of inspired oxygen delivery
to the neonate. Pediatric Pulmonology.
1992;14(2):110-117.
9. Sun Y, Kohane IS, Stark AR. Computer-assisted adjustment of inspired oxygen concentration improves
control of oxygen saturation in newborn infants
requiring mechanical ventilation. Journal of Pediatrics.
1997;131(5):754-756.
Herberts T, Miksch S, et al. Automatic control of
the inspired oxygen fraction in preterm Infants:
A randomized crossover Trial. American Journal of Respiratory and Critical Care Medicine.
2004;170(10):1095-1100.
11. Claure N, DUgard C, Bancalari E. Automated adjustment of inspired oxygen in preterm Infants with
trial. Journal of Pediatrics. 2009;155(5):640-645.

Conference of the IEEE, EMBS. Minneapolis, Minnesota, USA; 2009. p. 3079-3082.
13. Tehrani FT, Abbasi S. Evaluation of a computerized
system for mechanical ventilation of infants. Journal of
Clinical Monitoring and Computing. 2009;23(2):93-104.
14. Claure N, Bancalari E, DUgard C, Nelin L, Stein M,
Ramanathan R, et al. Multicenter crossover study of
automated control of inspired oxygen in ventilated
preterm infants. Pediatrics. 2011;127(1):e76-83.
15. ODonnell CP. Automated adjustment of oxygen in
ventilated preterm infants: turn on, tune in, ROP out?
2004;114(5):1341-1347.
17. Armbruster J, Schmidt B, Poets CF, Bassler D. Nurses
compliance with alarm limits for pulse oximetry:
qualitative study. J Perinatol. 2009;30:531-534.
18. Clucas L, Doyle LW, Dawson J, Donath S, Davis PG.
Compliance with alarm limits for pulse oximetry in
very preterm infants. Pediatrics. 2007;119(6):1056-1060.
19. Mills B, Davis P, Donath S, Clucas L, Doyle L. Improving compliance with pulse oximetry alarm limits for
very preterm infants? Journal of Paediatrics and Child
Health. 2010;46(5):255-258.
20. Nghiem TH, Hagadorn JI, Terrin N, Syke S,
MacKinnon B, Cole CH. Nurse opinions and pulse
oximeter saturation target limits for preterm infants.
Pediatrics. 2008;121(5):e1039-1046.
21. McIntyre J. Ergonomics: Anaesthetists use of auditory
alarms in the operating room. International Journal of
22. Holden JD. Hawthorne effects and research into
professional practice. Journal of Evaluation in Clinical
Practice. 2001;7(1):65-70.
23. Collins P, Levy NM, Beddis IR, Godfrey S, Silverman
M. Apparatus for the servocontrol of arterial oxygen
tension in preterm infants. Medical and Biological
24. Dugdale RE, Cameron RG, Tealman GT. Closed-loop
control of the partial pressure of arterial oxygen in
neonates. Clinical Physics and Physiological
Measurement. 1988;9(4):291-305.
25. Claure N, Gerhardt T, Everett R, Musante G, Herrera
C, Bancalari E. Closed-loop controlled inspired oxygen
concentration for mechanically ventilated very low
birth weight infants with frequent episodes of hypoxemia. Am Acad Pediatrics. 2001;107(5):1120-1124.
02-08-12 11:37
90
02-08-12 11:37
CHAPTER 5
Pulse oximetry alarm

limits in extremely low
birth weight infants:
when do deviations from
the protocol occur?
A.C. van der Eijk, D. Rook, J. Dankelman, S. Schutte, B.J. Smit, H.J. Simonsz
02-08-12 11:37
92
CHAPTER 5
Alarm management is an on-going issue in medical departments, and especially in neonatal intensive care units. The high false alarm rate of pulse oximeters may lead to a
reduced response of the staff and unnecessary stress for patients and their family. This
chapter is based on the article Pulse oximetry alarm limits in extremely low birth weight
infants: when do deviations from the protocol occur? (submitted), and evaluates the current practice of pulse oximetry alarm limits, by making use of the data obtained during
the observational study descriped in Chapter 4. The focus of this chapter is on the occurrence of the (non-)compliance to the protocol for pulse oximetry alarm limits and the
possible reasons and motivations for the (non-)compliance.
OBJECTIVE To study the compliance to the protocol for pulse oximetry alarm limits in
extremely low birth weight infants in relation to FiO2, SpO2, and bedside care.
02-08-12 11:37
Pulse oximetry alarm limits in ELBW infants
93
5.1 INTRODUCTION
In neonatal intensive care, arterial oxygen saturation of extremely low birth weight (1000
g.; ELBW) infants is monitored continuously by pulse oximetry (SpO2). To prevent disorders related to supplemental oxygen therapy, like retinopathy of prematurity (ROP) and
bronchopulmonary dysplasia (BPD), auditory alarms are used to warn the nursing staff when
SpO2 is outside of alarm limits.
In the past decades medical devices in neonatal care became more complex, and the number and variety of alarms have increased.1 Originally, the word alarm comes from Italian
allarme, meaning: get your weapons. Hence, an alarm should only sound at those moments
that an action is really needed. In current neonatal intensive care, the proportion of non
critical or false alarms is high. Lawless suggested that over 94% of the alarms in a pediatric
intensive care unit was not clinically important.2 Almost half of these alarm were initiated
by pulse oximetry. This high false alarm rate may lead to a reduced response of the staff and
unnecessary stress for patients and their family.1-14 Not surprisingly, alarm management is an
on-going issue in neonatology2, 15 and other medical departments.3,16-18
Optimal target levels for SpO2 in ELBW infants are still under debate, and the protocols
for target levels for SpO2 and the matching pulse oximetry alarm limits differ between
centres.19-37 Even worse, in daily practice the SpO2 alarm limits vary within one department.
These diverse alarm limit policies and practices may affect the outcome of ELBW infants.
None of the studies that focused on the pulse oximetry alarm limit settings collected corresponding SpO2, FiO2 and/or bedside care. Hence, no information is available about the
situations in which alarm limit adjustments are performed. Therefore, the aims of this study
were to evaluate the compliance to the protocol for pulse oximetry alarm limits in ELBW
infants and to identify the circumstances where deviations from the protocol for pulse oximetry alarm limits were performed. On ward video recordings, alarm limits, SpO2 levels,
and the fraction of inspired oxygen (FiO2) supplied to the ELBW infants obtained during an
observational study were used for descriptive analysis.
Originally, the word alarm comes from Italian

allarme, meaning: get your weapons.
02-08-12 11:37
94
CHAPTER 5
5.2 Methods
In an observational study on manual control of supplemental oxygen42 in the level-III-c43
neonatal intensive care unit (NICU) of the Erasmus Medical Centre - Sophia Childrens
Hospital in Rotterdam, the Netherlands, on-ward video and data recording was performed.
Patients were eligible for this observational study when gestational age (GA) was 28 weeks,
birth weight (BW) was 1000 g., postnatal age (PNA) was between 2 and 14 days, and there
was a need for supplemental oxygen therapy.
Alarm limits for pulse oximetry were collected together with SpO2, FiO2, and bedside care
of the NICU staff each second for 48 to 72 hours per included infant. These recorded data
and videos were used to evaluate the compliance to the protocol for pulse oximetry alarm
limits and to identify the circumstances under which deviations from the protocol for pulse
oximetry alarm limits were performed in ELBW infants.
5.2.1 Working situation
The NICU of the Sophia Childrens Hospital has capacity for 30 patients divided over
three rooms. During the data collection the nurse:patient ratio varied between 1:2 and 1:3.
Working shifts for the nursing staff last 8 hours; day shift (08.00 - 16.00), evening shift
(16.00 - 00.00), and night shift (00.00 - 08.00). During a 30-minute handover between shifts
nurses inform each other about the (clinical) situation of the patients.
5.2.2 Policy for pulse oximetry alarm limits
In the studys centre the protocol for pulse oximetry alarm limits was 88% (lower limit)
and 94% (upper limit) when FiO2 was >21%, or 88% and 100% when infants were on room
air (FiO2 = 21%) for all patients with a postmenstrual age (PMA) <32 weeks. An extra loud
desaturation alarm was defined for SpO2 <80%. The nursing staff was allowed to adjust
alarm limits in the multi parameter patient monitor (Hewlett Packard Viridia monitor,
CMS M1046A) according to the protocol. The medical staff was allowed to prescribe alarm
limits that deviate from the protocol. All adjustments in the alarm limits should be communicated between the responsible nurse and physician and should be noted in the digital
Patient Data Management Software (PDMS; Picis Inc., Wakefield, United Kingdom), a software program on the PC next to the incubator.
5.2.3 Data analysis
Data was suitable for analysis when SpO2, FiO2, and the alarm limits for pulse oximetry
were available. To define whether an adjustment of alarm limits from 88 - 94% to 88 - 100%
or vice versa, was or was not according to the protocol, FiO2 levels within five minutes before and after a change in the alarm limits were taken into account. The same five minutes
before and after the adjustment were analysed to study the bedside care situation and SpO2
levels. Alarm limit adjustments were divided in two categories: according to the protocol
and not according to the protocol. When alarm limits were not set according to the proto-
02-08-12 11:37
95
col, motivations for the deviations were looked for in the patient records. When no motivation was found, the adjustment was defined as undetermined deviation. Furthermore, the
adjustments were categorised in wider or narrower allowed range for SpO2.
5.3 RESULTS
Twelve ELBW infants were included (median (min - max) GA = 26 2/7 (24 2/7 - 28) weeks,
BW = 760 (545 - 935) gram, PNA = 4 (2 - 12) days). In Appendix A, detailed patient characteristics are presented. In total 654 hours of data was suitable for analysis. This data was obtained during day shifts (225 hours), evening shifts (234 hours), and night shifts (195 hours).
Four typical examples of the collected data are shown in Figure 5.1.
100
90
80
70
[%]
60
SpO2
50
FiO2
40
Alarm limit SpO2
30
20
10
0
247
247.5
248
248.5
249
249.5
250
250.5
251
Postnatal age [hours]
Figure 5.1A Four hours of recorded data from patient #2. Alarm limits are adjusted twice. Once from
88 - 96% to 88 - 99%, and once from 88 - 99% to 88 - 94%.
02-08-12 11:37
96
CHAPTER 5
100
90
80
70
60
[%]
50
40
30
SpO2
FiO2
20
Alarm limit SpO2

10
0
251
250.5
251.5
252
252.5
253
253.5
254
254.5
Figure 5.1B Four hours of recorded data from patient #4. Alarm limits are adjusted once from 88 - 94%
to 88 - 98%.
100
90
80
70
[%]
60
SpO2
FiO2
50
Alarm limit SpO2

40
30
20
10
0
96.5
97
97.5
98
98.5
99
99.5
100
100.5
101
101.5
Figure 5.1C Four hours of recorded data from patient #5. Alarm limits are adjusted five times: from
88 - 94% to 88 - 100% to 88 - 96% to 88 - 94% to 88 - 100% and to 88 - 94%.
02-08-12 11:37
97
100
90
80
70
[%]
60
50
40
30
SpO2
20
FiO2
10
Alarm limit SpO2
0
136
136.5
137
137.5
138
138.5
139
139.5
140
Figure 5.1D Four hours of recorded data from patient #6. Alarm limits are adjusted three times: from
88 - 99% to 88 - 96% to 88 - 94% and to 92 - 98%.
5.3.1 Occurrence of alarm limits

Supplemental oxygen therapy (FiO2 >21%) was provided in 563 hours. During these hours,
alarm limits were set according to the protocol in 64% of the time (67% for day shifts, 69%
for evening shifts, and 57% for night shifts). In the remaining time, nine other alarm limit
combinations were used (Table 5.1). When alarm limits deviated from the protocol, the upper limit was always increased; in 5% of the time the lower limit was increased as well. In
total, in 9% of the time the upper alarm limit was set to 100%, hence there was no alarm for
high SpO2 values while patients received supplemental oxygen therapy.
When patients were on room air (91 hours), alarm limits were set according to the protocol
in 84% of the time (82% for day shifts, 83% for evening shifts, and 88% for night shifts). In
the remaining time, five other alarm limit combinations were used with an upper limit set
<100% (Table 5.1).
The non-protocolled alarm limits were observed in eight patients; in two patients alarm
limits were set to 88 - 94% during the complete recording, and in two patients alarm limits
were changed between 88 - 94% and 88 - 100%.
02-08-12 11:37
98
CHAPTER 5
Table 5.1 Occurrence of alarm limits, and corresponding FiO2 levels in the included patients (n=12)
Occurrence of alarm limits [% of time]
Alarm limits
Number of patients
where limits occurred
For FiO2 >21%
For FiO2 =21%
Average number of
FiO2 adjustments
per hour
88-94
12
64
11
3.5
88-95
3.1
88-96
2.5
88-98
1.2
88-99
<1
3.2
88-100
84
1.6
90-96
<1
6.4
90-98
<1
<1
5.8
92-98
3.2
93-98
<< 1
The bold numbers represent the percentage of time that alarm limits were according to the protocol.
5.3.2 FiO2 levels

The average number of FiO2 adjustments was 3.7, 2.6, and 2.3 per hour for day shifts, evening shifts, and night shifts, respectively.
5.3.3 SpO2 levels
When supplemental oxygen was provided and alarm limits were set according to the protocol (88 and 94%), corresponding SpO2 values were within alarm limits for 54% of the time.
When the alarm limits were set wider than 88 to 94%, the time SpO2 spent within the alarm
limits increased up to 85% for alarm limits set to 88 and 100% (Figure 5.2).
5.3.4 Characteristics of the alarm limit adjustments
Alarm limits were adjusted 45 times (21 times during day shifts, 17 times during evening
shifts, and 7 during night shifts) by 22 different nurses and 2 physicians. In 27 of the 45 times
the adjustments were performed according to the protocol: alarms limits were changed to
88 - 94% because FiO2 was changed from room air to >21% in 17 cases, and alarm limits were
changed to 88 - 100% because FiO2 was changed from >21% to room air in 10 cases. FiO2 was
adjusted from >21% to room air or vice versa in 156 times. Hence, in 129 (82%) of the changes
from room air to >21% or vice versa, alarm limits were not adjusted, resulting in either too
much (i.e. unnecessary upper limit) or too less number (i.e. lack of upper limit) of alarms.
02-08-12 11:37
99
Figure 5.2 Total time [%] SpO2 was below, within or above the preset alarm limits for the seven most
frequent occurring alarm limit combinations for the time that FiO2 was >21%.
In 18 of the 45 cases the adjustments were not according to protocol. In 4 of these 18 adjustments an upper limit was set <100% while the patient was on room air. In the other 14 of
the 18 alarm limit adjustments the limits were set wider and/or higher than 88 - 94% while
FiO2 was >21%. In 3 of these 14 non-protocolled alarm limit adjustments, the adjustment
was noted in PDMS as on medical indication. These 3 adjustments were made in patient
#4 and #6, both of them were diagnosed with pulmonary hypertension of the neonate
(PPHN). The motivations for the remaining 11 deviations were undetermined. In 8 of these
11 adjustments the alarm limits were set wider. Thus, the allowed range for SpO2 was
larger than the previous alarm limit setting.
The details of the adjustments in alarm limits are shown in Table 5.2 and Figure 5.3. In
Appendix B the circumstances with respect to the FiO2, SpO2, and bedside care around
non-protocolled alarm limit adjustments are described in detail for each patient separately.
02-08-12 11:37
100
CHAPTER 5
Alarm limits before

adjustment
Alarm limits after

adjustment
FiO2 level 5 minuts

before & after
adjustment, minimum-maximum [%]
SpO2 level 5 minuts

before & after adjustment, minimummaximum [%]
Staff member
Ventilation at
moment of
adjustment
According to the
protocol
56
nurse 1
nurse 2
nurse 3
nurse 4
nurse 5
nurse 5
nurse 5
nurse 5
nurse 5
nurse 6
nurse 6
nurse 6
HFV
HFV
HFV
HFV
HFV
NIV
NIV
NIV
NIV
NIV
NIV
NIV
N0
N0
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
26 4/7
800
2nd of
twins
24
day
day
day
88 - 94
88 - 96
88 - 99
88 - 100
88 - 99
88- 94
21 - 21
21 - 26
23 -25
90 - 100
84 - 99
85 - 99
nurse 7
nurse 7
physician 1
HFO
HFO
HFO
Yes
No
Yes
12
60
evening
day
day
day
evening
88 - 94
88 - 100
88 - 94
88 - 100
88 - 94
88 - 100
88 - 94
88 - 100
88 - 94
88 - 100
21 - 34
30 - 30
21 - 23
27 - 52
22 - 22
88 - 100
81 - 96
89 -99
81 - 100
90 - 100
nurse 8
nurse 9
nurse 9
nurse 9
nurse 10
NIV
NIV
CPAP
CPAP
CPAP
No
Yes
Yes
Yes
N0
53
evening
day
88 - 96
88 - 94
88 - 94
88 - 98
50 - 50
85 - 85
90 -100
72 -98
nurse 11
nurse 12
SIMV
SIMV
Yes
N0
47
evening
evening
evening
night
night
day
day
day
day
day
day
88 - 94
88 - 100
88 - 94
88 - 100
88 - 94
88 - 100
88 - 96
88 - 94
88 - 100
88 - 94
88 - 100
88 - 100
88 - 94
88 - 100
88 - 94
88 - 100
88 - 96
88 - 94
88 - 100
88 - 94
88 - 100
88 - 94
21 - 25
24 - 27
21 - 22
22 - 25
21 -23
21 - 25
21 - 22
21 - 22
27 - 31
21 - 26
21 - 23
89 - 100
91 - 100
88 - 98
83 - 91
85 - 100
92 - 100
82 - 100
82 - 100
81 - 98
87 -97
87 - 97
nurse 13
nurse 13
nurse 14
nurse 14
nurse 14
physician 2
nurse 15
nurse 15
nurse 15
nurse 15
nurse 15
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
Yes
Yes
Yes
Yes
No
No
No
No
Yes
Yes
Yes
88 - 100
88 - 99
88 - 96
88 - 94
92 - 98
88 - 98
88 - 99
88 - 96
88 - 94
92 - 98
88 - 98
92 - 97
28 - 28
29 - 38
26 - 30
29 - 38
34 - 40
35 - 35
93 - 99
89 - 100
75 - 99
97 - 100
78 - 99
88 - 100
nurse 16
nurse 16
nurse 17
nurse 17
nurse 18
nurse 18
SIMV
SIMV
SIMV
SIMV
SIMV
SIMV
No
No
Yes
No
No
No
28
935
1st of
twins
27
545
27 3/7
885
2nd of
twins
Data suitable for

analysis [hours]
95 - 100
87 - 93
95 - 100
79 - 91
71 - 100
82 - 99
84 - 99
88 - 100
85 - 98
91 - 98
80 - 98
86 - 99
Multiple births
21 - 21
23 - 25
21 - 22
21 - 23
30 - 35
21 - 25
21 - 23
22 - 24
25 - 26
21 - 22
21 - 25
21 - 23
Caesarion section
615
90 - 98
88 - 96
88 - 100
88 - 94
88 - 94
88 - 100
88 - 94
88 - 100
88 - 94
88 - 100
88 - 94
88 - 100
Sexe
26 2/7
90 - 96
90 - 98
88 - 96
88 - 100
88 - 100
88 - 94
88 - 100
88 - 97
88 - 100
88 - 94
88 - 100
88 - 97
Gestational age
[weeks]
evening
evening
night
evening
evening
night
night
night
night
day
day
day
Patient
Working shift
Post natal age at

start recording [days]
Birth weight [grams]
Table 5.2 Characteristics of alarm limit adjustment for each infant separately
26 6/7
640
60
evening
evening
evening
evening
day
day
26 1/7
845
2nd of
twins
52
evening
day
88 - 96
88 - 94
88 - 96
88 - 94
25 - 26
25 - 26
91 - 99
91 - 99
nurse 19
nurse 20
SIMV
SIMV
No
Yes
26 1/7
805
1st of
twins
65
evening
evening
88 - 94
88 - 100
88 - 94
88 - 100
21 - 26
21 - 21
80 - 100
95 - 100
nurse 4
nurse 4
NIV
NIV
Yes
Yes
26 1/7
690
71
10
25 3/7
585
62
42
day
93 - 100
88 - 94
25 - 25
85 - 100
nurse 21
SIMV
Yes
62
day
88 - 95
88 - 95
29 - 29
84 - 98
nurse 22
HFO
No
11
24 2/7
720
triples,
7
1 life born
12
25 2/7
875
02-08-12 11:37
SpO2 level 5 minuts

before & after adjustment, minimummaximum [%]
Staff member
Ventilation at
moment of
adjustment
According to the
protocol
Motivations found in
patient records
Allowed range for

SpO2
Type of adjustmente
- 98
- 96
- 100
- 94
- 94
- 100
- 94
- 100
- 94
- 100
- 94
- 100
21 - 21
23 - 25
21 - 22
21 - 23
30 - 35
21 - 25
21 - 23
22 - 24
25 - 26
21 - 22
21 - 25
21 - 23
95 - 100
87 - 93
95 - 100
79 - 91
71 - 100
82 - 99
84 - 99
88 - 100
85 - 98
91 - 98
80 - 98
86 - 99
nurse 1
nurse 2
nurse 3
nurse 4
nurse 5
nurse 5
nurse 5
nurse 5
nurse 5
nurse 6
nurse 6
nurse 6
HFV
HFV
HFV
HFV
HFV
NIV
NIV
NIV
NIV
NIV
NIV
NIV
N0
N0
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
Yes
Yes
Undetermined deviation
-
Wider
Narrower
Wider
-
FiO2 = 21%, Needless upper limit, too high lower limit

FiO2 >21%, Too high upper limit, lower limit corrected
FiO2 >21%, No upper limit
-
- 100
- 99
94
21 - 21
21 - 26
23 -25
90 - 100
84 - 99
85 - 99
nurse 7
nurse 7
physician 1
HFO
HFO
HFO
Yes
No
Yes
-
Wider
-
FiO2 = 21%, Needless upper limit

-
- 100
- 94
- 100
- 94
- 100
21 - 34
30 - 30
21 - 23
27 - 52
22 - 22
88 - 100
81 - 96
89 -99
81 - 100
90 - 100
nurse 8
nurse 9
nurse 9
nurse 9
nurse 10
NIV
NIV
CPAP
CPAP
CPAP
No
Yes
Yes
Yes
N0
Wider
Wider
FiO2 = 21%, No upper limit

- 94
- 98
50 - 50
85 - 85
90 -100
72 -98
nurse 11
nurse 12
SIMV
SIMV
Yes
N0
Medical indication
Wider
- 100
- 94
- 100
- 94
- 100
- 96
- 94
- 100
- 94
- 100
- 94
21 - 25
24 - 27
21 - 22
22 - 25
21 -23
21 - 25
21 - 22
21 - 22
27 - 31
21 - 26
21 - 23
89 - 100
91 - 100
88 - 98
83 - 91
85 - 100
92 - 100
82 - 100
82 - 100
81 - 98
87 -97
87 - 97
nurse 13
nurse 13
nurse 14
nurse 14
nurse 14
physician 2
nurse 15
nurse 15
nurse 15
nurse 15
nurse 15
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
HFV
Yes
Yes
Yes
Yes
No
No
No
No
Yes
Yes
Yes
-
Wider
Narrower
Narrower
Wider
-

-
- 99
- 96
- 94
98
- 98
97
28 - 28
29 - 38
26 - 30
29 - 38
34 - 40
35 - 35
93 - 99
89 - 100
75 - 99
97 - 100
78 - 99
88 - 100
nurse 16
nurse 16
nurse 17
nurse 17
nurse 18
nurse 18
SIMV
SIMV
SIMV
SIMV
SIMV
SIMV
No
No
Yes
No
No
No
Medical indication
Medical indication
Narrower
Narrower
Equal
Wider
Narrower
FiO2 = 21%, Too high upper limit

FiO2 = 21%, Too high upper limit, lower limit corrected
-
- 96
- 94
25 - 26
25 - 26
91 - 99
91 - 99
nurse 19
nurse 20
SIMV
SIMV
No
Yes
-
Wider
-
- 94
- 100
21 - 26
21 - 21
80 - 100
95 - 100
nurse 4
nurse 4
NIV
NIV
Yes
Yes
- 94
25 - 25
85 - 100
nurse 21
SIMV
Yes
- 95
29 - 29
84 - 98
nurse 22
HFO
No
Wider
adjustment
FiO2 level 5 minuts

before & after
adjustment, minimum-maximum [%]
101
02-08-12 11:37
102
CHAPTER 5
Alarm limit adjustments

n = 45
Non-protocolled
n = 18
Protocolled
n = 27
Medical indication
n=0
FiO2 = 21%
n=4
FiO2 >21%
n = 14
Undetermined
n=4
Medical indication
n=3
Narrower
n=2
Wider
n=2
Narrower
n=1
Equal
n=1
Undetermined
n = 11
Wider
n=1
Wider
n=8
Narrower
n=3
Figure 5.3 The categorisation of the 45 alarm limits adjustment.
5.4 DISCUSSION
This single-centre study evaluated the compliance to the protocol for pulse oximetry alarm
limits in twelve ELBW infants and identified circumstances where deviations from the protocol for pulse oximetry alarm limits were performed. When supplemental oxygen therapy
was provided, alarm limits deviated from the protocol in 36% of the time. In 9% of the time
there was no alarm for high SpO2 values. Furthermore we demonstrated that in >80% of
the FiO2 adjustments from room air to >21% or vice versa, alarm limits were not changed,
resulting in either too much (i.e. unnecessary upper limit) or too less number (i.e. lack of upper limit) of alarms. No relation was found between alarm limit settings and the frequency
of adjustments in FiO2.
In the 563 hours of collected data, alarm limit were adjusted 45 times, of which 18 were not
according to the protocol. In 15 of these 18 adjustments, the motivation for the adjustment
remained undetermined. Consequently, it was not clear why adjustments were made, and
whether the adjustments were discussed with the medical staff. This unregistered variation
in alarm limits is undesirable because they may influence outcome of patients.
Literature showed that deviations from the protocol for alarm limits occur frequently in
other centres as well.38-41 For example, the results of a single centre study by Laptook et al.39
showed that alarm limits were set incorrectly in 26 15% (range 6 to 56%) of the time. After
a change in the protocol, alarm limits were set incorrectly in 23 16% (range 0 to 50%) of the
time. Clucas et al.38 presented data that showed that the lower and upper alarm limit were
02-08-12 11:37
103
not according to the protocol in respectively 8.9% and 76.5% of the time. Nghiem et al.40
illustrated that only 28% of the nurses of NICUs in the US accurately identified the upper
and lower limits of their centres policy for pulse oximetry alarm limits. Although these
studies present interesting information, they focussed on alarm limits for pulse oximetry
only. To our knowledge, this study is the first that described alarm limits for pulse oximetry
as well as the corresponding SpO2, FiO2 and bedside care. Thanks to the observational
character of our study, we were able to evaluate the effect of alarm limits as occurred in clinical practice without the possible introduction of bias due to study-related interventions.
When alarms limits are adjusted, they remain unchanged until the patient monitor is
turned on/off (reset), or until someone adjusts the alarm limits again. In an environment
with a high workload and work in shifts, such as a NICU, it is possible that the staff is not
fully aware of current alarm limits for each patient.44-46 Thus, in clinical practice there are
four different situations possible (Table 5.3): The alarm limits are or are not set according to
the protocol, in combination with the nursing staff being or not being aware of the alarm
limits. Ideally, the NICU personnel is fully aware of the alarm limits and all the alarms are
set according to the protocol (situation 1). However, both literature and our results show
that alarm limits are frequently in situation 3 or 4.
Table 5.3 The four situations that can occur when alarm settings are adjusted by the NICU personnel
Staff is aware of
current alarm limits
Staff is unaware of
current alarm limits
Alarm limits set according to

protocol
Situation 1:
Protocolled and aware
Situation 2:
Protocolled and unaware
Alarm limits not set according to

protocol
Situation 3:
Non-protocolled and aware
Situation 4:
Non-protocolled and unaware
Apparently caregivers agree (sub)consciously with the frequently deviating alarm limits for
certain patients. This agreement suggests that there could be well argued, but unspoken
and unexpressed reasons for (not) making alarm limit adjustments based on e.g., earlier
experiences. This happened, for instance, in the two patients where PPHN was diagnosed
and SpO2 was aimed higher than the protocol prescribed. Hence, it could be argued that
the current protocol of the studys centre, providing only two different alarm limit settings,
is invalid or incomplete for the patient group it is used for.
To improve alarm management, and make sure situation 1 is the only situation in the
NICU, the motivations for the non-protocolled alarm limits should be studied first. In 1985,
McIntyre showed that possible reasons for alarm limit changes are related to a reduction
of the number of alarms, the lack of consensus about target levels for SpO2, factors related
02-08-12 11:37
104
CHAPTER 5
to the individual nurse or physician, and the temporarily changes of patient care.47 When
motivations for deviations in the protocol indeed appear to be valid for certain patients
(e.g., when PPHN is diagnosed), the protocol should be expanded to allow the NICU staff
to make these deviations. The next step is to make sure that NICU personnel is aware of
the actual alarm limit settings. To achieve this, technical improvements in medical devices
may help. For instance, periodic reminders about the current alarm limits or the implementation of a digital protocol changing the alarm limits automatically could increase both
awareness and compliance to the protocol.
To conclude, deviations from the protocol for alarm limits for SpO2 occur frequently and
the motivations for deviating from the protocol were poorly registered. The unregistered
variation in clinical practice may influence outcome results without providing a clear insight in the causes for the outcome differences, and should therefore be minimised. However, the frequent noncompliance to the protocol raises questions about the actual validity of the protocol. To improve the (compliance of the) protocol in the near future, it will
be necessary to check alarm limits regularly and to register and evaluate (motivations for)
alarm limit adjustments.
02-08-12 11:37
105
5.5 References
1. Meredith C, Edworthy J. Are there too many alarms in
the intensive care unit? An overview of the problems.
Journal of Advanced Nursing. 1995;21(1):15-20.
1994;22(6):981-985.
3. Block F, Nuutinen L, Ballast B. Optimization of alarms:
a study on alarm limits, alarm sounds, and false alarms,
intended to reduce annoyance. Journal of Clinical
Monitoring and Computing. 1999;15(2):75-83.
4. Chambrin MC. Alarms in the intensive care unit: how
can the number of false alarms be reduced. Critical
Care. 2001;5(4):184-188.
5. Meyer TJ, Eveloff SE, Bauer MS, Schwartz WA, Hill
NS, Millman RP. Adverse environmental conditions
in the respiratory and medical ICU settings. Chest.
1994;105(4):1211-1216.
6. Balogh D, Kittinger E, Benzer A, Hackl JM. Noise in the
ICU. Intensive Care Medicine. 1993;19(6):343-346.
7. Aaron JN, Carlisle CC, Carskadon MA, Meyer TJ, Hill
NS, Millman RP. Environmental noise as a cause of
sleep disruption in an intermediate Respiratory Care
unit. Sleep. 1996;19(9):707-710.
8. Kam PCA, Kam AC, Thompson JF. Noise pollution
in the anaesthetic and intensive care environment.
Anaesthesia. 1994;49(11):982-986.
9. Sabar R, Zmora E. Nurses response to alarms from
monitoring systems in NICU. Pediatric Research.
1997;41(4 Part 2):174.
10. Ahlborn V. False alarms in very low birthweight infants:
Comparison between three intensive care monitoring
systems. Acta Paediatr. 2000;89(5):571-576.
11. McLaughlin A, McLaughlin B, Elliott J, Campalani G.
Noise levels in a cardiac surgical intensive care unit:
A preliminary study conducted in secret. Intensive and
Critical Care Nursing. 1996;12(4):226-230.
12. Tsien CL, Fackler JC. Poor prognosis for existing
monitors in the intensive care unit. Critical Care
Medicine. 1997;25(4):614-619.
13. Oberli C, Urzua J, Saez C, Guarini M, Cipriano A,
Garayar B, et al. An expert system for monitor alarm
integration. Journal of Clinical Monitoring and Computing. 1999;15(1):29-35.
14. Salyer JW. Neonatal and pediatric pulse oximetry.
Respiratory Care. 2003;48(4):386-396.
15. Bitan Y, Meyer J, Shinar D, Zmora E. Nurses reactions
to alarms in a neonatal intensive care unit. Cognition,
Technology & Work. 2004;6(4):239-246.
16. Kestin IG, Miller BR, Lockhart CH. Auditory alarms

during anesthesia monitoring. Anesthesiology.
1988;69(1):106-109.
17. Edworthy J, Hellier E. Fewer but better auditory alarms
will improve patient safety. British Medical Journal.
2005;14(3):212-215.
18. Chambrin M, Ravaux P, Calvelo-Aros D, Jaborska
A, Chopin C, Boniface B. Multicentric study of
monitoring alarms in the adult intensive care unit
(ICU): A descriptive analysis. Intensive Care Medicine.
1999;25(12):1360-1366.
19. Castillo A, Sola A, Baquero H, Neira F, Alvis R,
Deulofeut R, et al. Pulse oxygen saturation levels
and arterial oxygen tension values in newborns
receiving oxygentherapy in the neonatal intensive
care unit: is 85% to 93% an acceptable range?
Pediatrics. 2008;121(5):882-889.
20. Askie L. Appropriate levels of oxygen saturation for
preterm infants. Acta Paediatr. 2004;93(s444):26-28.
21. Chen M, Guo L, Smith L, Dammann C, Dammann O.
High or low oxygen saturation and severe
retinopathy of prematurity: A meta-analysis.
Pediatrics. 2010;125(6):e1483-1492.
22. Deulofeut R, Critz A, Adams-Chapman I, Sola A.
Avoiding hyperoxia in infants less than or equal to 1250
g is associated with improved short-and long-term
outcomes. J Perinatol 2006;26:700-705.
24. Network SSGotEKSNNR. Target ranges of oxygen
saturation in extremely preterm infants. The New
25. Noori S, Patel D, Friedlich P, Siassi B, Seri I, Ramanathan R. Effects of low oxygen saturation limits on
the ductus arteriosus in extremely low birth weight
infants. J Perinatol. 2009;29:553-557.
British Medical Journal. 2001;84(2):F106-110.
27. VanderVeen DK, Mansfield TA, Eichenwald EC. Lower
oxygen saturation alarm limits decrease the severity
of retinopathy of prematurity. Journal of AAPOS.
2006;10(5):445-448.
28. Wright KW, Farzavandi S. A physiologic reduced
oxygen protocol decreases the incidence of threshold
retinopathy of prematurity. Trans Am Ophthalmol Soc.
2006;104:78-84.
02-08-12 11:37
106
CHAPTER 5
29. Wallace DK, Veness-Meehan KA, Miller WC. Incidence

of severe retinopathy of prematurity before and after
a modest reduction in target oxygen saturation levels.
Journal of AAPOS. 2007;11(2):170-174.
of prematurity in very low birth weight infants? Pediatrics. 2003;111(2):339-345.
32. Wallace DK. Oxygen Saturation Levels and
Retinopathy of Prematurity - Are We on Target?
Journal of AAPOS. 2006;10(5):382-383.
33. Sears JE, Pietz J, Sonnie C, Dolcini D, Hoppe G. A
change in oxygen supplementation can decrease the
incidence of retinopathy of prematurity. Ophthalmology. 2009;116(3):513-518.
34. Tokuhiro Y, Yoshida T, Nakabayashi Y, Nakauchi S,
Nakagawa Y, Kihara M, et al. Reduced oxygen protocol
decreases the incidence of threshold retinopathy of
prematurity in infants of< 33 weeks gestation.
Pediatrics International. 2009;51(6):804-806.
35. Higgins RD, Bancalari E, Willinger M, Raju TNK.
Executive summary of the workshop on oxygen in
neonatal therapies: controversies and opportunities
for research. Pediatrics. 2007;119(4):790-796.
Pediatrics. 2002;110(3):615-616.
37. Tin W, Wariyar U. Giving small babies oxygen:
50 years of uncertainty. Seminars in Neonatology.
2002;7(5):361-367.
saturation be maintained in the desired range? J Perinatol. 2006;26:337-341.
Pediatrics. 2008;121(5):e1039-1046.
41. Mills BA, Davis PG, Donath SM, Clucas LM, Doyle LW.
Improving compliance with pulse oximetry alarm limits
for very preterm infants? Journal of Paediatrics and
Child Health. 2010;46(5):255-258.
42. van der Eijk AC, Dankelman J, Schutte S, Simonsz

HJ, Smit BJ. An observational study to quantify
manual adjustments of the inspired oxygen fraction
in extremely low birth weight infants. Acta Paediatr.
2012;101(3):e97104.
2004;114(5):1341-1347.
44. McDonald CJ. Protocol-based computer reminders, the quality of care and the non-perfectibility
of man. The New England Journal of Medicine.
1976;295(24):1351-1355.
45. Grimshaw J, Hutchinson A. Clinical practice guidelines
- do they enhance value for money in health care?
British Medical Bulletin. 1995;51(4):927.
46. Leonard M, Graham S, Bonacum D. The human factor:
The critical importance of effective teamwork and
communication in providing safe care. British Medical
Journal. 2004;13(Supplement 1):i85-90.
47. McIntyre J. Ergonomics: Anaesthetists use of auditory
alarms in the operating room. International Journal of
02-08-12 11:37
107
02-08-12 11:37
108
02-08-12 11:37
CHAPTER 6
Manual control of
oxygenation in
extremely low birth
weight infants:
what is the nurses
point of view?
A.C. van der Eijk, K. Henken, J. Dankelman, B.J. Smit
02-08-12 11:37
110
CHAPTER 6
The results obtained in the observational study (Chapter 4 and 5) provided insight in the
manual control behaviour of oxygenation by nurses and neonatologists. However, these
results cannot uncover the decision making processes of these healthcare professionals.
Therefore, to obtain more insight in the knowledge, the opinions and the attitude of the
nursing staff towards supplemental oxygen therapy, a survey was performed. In this survey, manual control of oxygenation, pulse oximetry, and alarm limits for pulse oximetry
were addressed. The results of the survey are discussed in this chapter.
OBJECTIVE To explore the decision making processes and obtain insight in the knowledge, opinions, and attitude of the nursing staff towards supplemental oxygen therapy
in extremely low birth weight infants.
02-08-12 11:37
A survey on the control of oxygenation in ELBW infants
111
6.1 INTRODUCTION
In preterm infants supplemental oxygen therapy is associated with, amongst others, retinopathy of prematurity (ROP) and bronchopulmonary dysplasia (BPD). To prevent these
negative effects, oxygenation needs to be monitored closely. When oxygenation of preterm
infants is monitored and controlled manually by the nursing staff, the oxygen saturation
measured by pulse oximetry (SpO2) is frequently outside of the target range.1-3
To reach and maintain adequate oxygenation of preterm infants, several groups have been
working on the development of devices for (semi-)automatic control. These devices adjust
the fraction of inspired oxygen (FiO2) supplied to the patient automatically when the SpO2
deviates from the target.4-9 However, the step size of FiO2 adjustments made by nurses
does not depend on the SpO2 levels just before the adjustment (Chapter 4). Hence, for the
nursing staff, SpO2 is probably not the (only) parameter to determine oxygen requirements
of preterm infants. To predict the consequences of SpO2 as the only input parameter in
automatic control devices, insight in the actual decision making process of the neonatal
intensive care unit (NICU) personnel is needed.
Therefore, the aim of this study was to explore the decision making processes and obtain
insight in the knowledge, opinions, and attitude of the nursing staff towards supplemental
oxygen therapy in extremely low birth weight (1000 g.; ELBW) infants. To obtain this information a survey was performed. This survey focussed on manual control of oxygenation,
pulse oximetry, and alarm limits for SpO2.
To predict the consequences of SpO2 as the only

input parameter in automatic control devices,
insight in the actual decision making process
of the NICU personnel is needed.
02-08-12 11:37
112
CHAPTER 6
6.2 Methods
A survey was performed among the nursing staff of the level-III-c10 NICU of the Erasmus
Medical Centre - Sophia Childrens Hospital in Rotterdam, the Netherlands. There were no
selection criteria for nurses to participate in the survey. The researchers selected eight patients which were born with a gestational age (GA) <30 weeks and with a birth weight (BW)
1000 g. At the moment of the selection the patients had a postnatal age (PNA) <14 days
and received supplemental oxygen therapy (FiO2 >21%). For three working shifts in row (i.e.
three working shifts of eight hours = 24 hours) nurses were asked to fill out a questionnaire
immediately after their working shift specifically for the selected patient under their care in
the preceding working hours.
The protocol for preterm infants with a postmenstrual age (PMA) <32 weeks prescribed
alarm limits for SpO2 of 88 and 94% when supplemental oxygen was supplied (FiO2 >21%)
or 88 and 100% for patients in room air (FiO2 = 21%). An alarm sounded when SpO2 was
outside these limits for more than 20 seconds. An extra loud low saturation alarm was set
for SpO2 <80%. The nursing staffs task was to keep SpO2 within the target range. There
was, besides their normal education, neither a protocol nor specific training for nurses describing when and with what step size FiO2 should be adjusted.
6.2.1 The questionNaire
The questionnaire used for the survey consisted of 34 questions. These questions were
created by the researchers, and were partly based on questions asked in earlier surveys
(for neonatal intensive care).11-28 The first eight questions focussed on general information
about the nurse (e.g., working experience). The following 15 questions addressed knowledge
and opinions about (prescribed) alarm limits for pulse oximetry, and the actions taken in
response to pulse oximetry alarms. The next four questions focussed on manual adjustments in FiO2, and were followed by six questions about pulse oximetry. The final question
encouraged the subjects to come with suggestions for improvement. Questions were either
multiple choice or open to answer. The answers to the open questions were allocated to
categories by the researchers.
6.3 Results
A total of 24 questionnaires were filled out for eight different infants at the end of day (11),
evening (7), and night (6) shifts. All nurses who were invited to fill out the questionnaire
responded. The nurses who filled out the questionnaires were between 25 and 53 years old
((mean SD) 36 9.5 years), and had between 0 to 20 years working experience as NICU
nurse (7.4 5.6 years). They worked on the floor between 20 and 36 hours each week (32
4.4 hours) in the function of nurse administrator (1 nurse), senior NICU nurse (16 nurses),
NICU nurse (5 nurses), and student NICU nurses (2 nurses).
02-08-12 11:37
113
6.3.1 Manual control of oxygenation

The nurses answered that, when an alarm sounds for a Lowsat (SpO2 <88%), they first
observe the infant, focusing especially on the respiration pattern and skin colour (Table 6.1).
Based on their findings, they performed one or more of the nine actions listed in Table 6.2.
Seventeen nurses mentioned to act differently when the extra loud low saturation alarm
goes off (SpO2 <80%). These nurses stated that the actions performed to recover the SpO2
value were the same as for a normal alarm, but in these cases the nurses say that they tend
to start sooner with suctioning and increasing the FiO2.
For a Highsat alarm (SpO2 >94%), nurses mentioned four different aspects they take into
account to assess the oxygen requirement of the ELBW infant, and four different kinds
of actions to bring back SpO2 within the target range. The nurses indicated that they
focus on the infant in general, the respiration, the skin colour, and the patient monitor. The
actions performed were adjusting FiO2, waiting, suppressing the alarm, or adjusting the
upper alarm limit.
Table 6.1 Overview of the points of interest mentioned by the nurses (n=24) when SpO2 alarm sounds
Number of times mentioned
Points of interest
Lowsat alarm (normal or extra loud)
Highsat alarm
Activity of infant
Measurement of pulse oximeter
Obstruction of sputum present
Patient in general
Position of CPAP
Position of infant
Position of pulse oximeter
Position of endotracheal tube
Respiration pattern of patient
Skin colour of infant
Watch monitor
02-08-12 11:37
114
CHAPTER 6
Table 6.2 Overview of actions performed by nurses (n=24) when SpO2 alarm sounds
Number of times mentioned
Lowsat alarm
Extra loud Lowsat alarm
Adjust FiO2
Adjust position of infant
Adjust settings mech. ventilator
Adjust upper alarm limit
Increase comfort infant
Replace CPAP
Suctioning
14
Supplemental manual inspiration
Suppress alarm
Wait
10
Actions performed
According to the nurses, the FiO2 supplied to the infants was between 22 and 42% at the
start of that working shift ((mean SD) 30 6.5%). Each nurse changed FiO2 at least once
during the shift. The reason for making a change in the FiO2 was because of a change in the
SpO2, sometimes caused by an intervention like suctioning or changing the infants nappy.
The answers concerning the step size used for an increase in FiO2 reported values between 1
and 10%. For decreases in FiO2 the step size varied between 1 and 5%. In one case, FiO2 was
decreased with 20% at once. According to the responsible nurse, this was done for: a young
infant with high saturation and an initial FiO2 of 42%.
A multiple choice question regarding the reliability of pulse oximeter measurements showed
that most nurses feel that pulse oximetry is reliable (18 times) or very reliable (3 times). Two
nurses gave neutral as answer, one nurse did not answer. The nurses stated that the curve
of the plethysmogram clearly shows when the measurement of the pulse oximeter is not
reliable. According to the nurses, unreliable measurements can be caused by cold feet of the
patient or when the patient is moving.
In an open question concerning the quality of the pulse oximeter measurement, nurses
mentioned the plethysmogram (22 times), the latest SpO2 value (18 times), and the trend
of the SpO2 (16 times) as their main source for information. In addition, heart rate, clinical
status of the patient, location of the pulse oximeter, and blood perfusion values were each
mentioned once as a source of information.
02-08-12 11:37
115
6.3.3 Pulse oximetry alarm limits

All 24 nurses claimed that they learned about pulse oximetry alarm limits for preterm infants during their NICU education. They all learned to use 88% as the lower alarm limit.
Twenty-two nurses learned that 94% is the prescribed upper limit. One nurse (46 years old,
working 6 years as NICU nurse) learned that 95% is the correct upper boundary, although in
a refresher course he learned 94% as well. Another nurse (37 years old, working 12 years as
NICU nurse) learned an upper limit of 96%.
All nurses stated that the lower alarm limit for the oxygen saturation was indeed set to 88%
at the beginning of their shift. None of the nurses changed the lower alarm limit during
their shift. Nineteen nurses stated that the upper limit was 94% at the start of their shift;
four nurses stated that the upper limit was 96%, and one nurse stated that the upper limit
was 99% at the start of their working hours. Five nurses changed the upper limit during
their shift: in two cases, the upper limit was corrected to 94%; in the other three cases, the
upper limit was set >94%. A reason mentioned for these adjustments was to reduce the
number of alarms, because the oxygen saturation of the ELBW infant was very instable and
consequently FiO2 could not be decreased.
In three out of five cases in which the upper limit was changed, the new upper limit was not
corresponding to the upper limit of the same infant at the beginning of the next shift. This
could suggest that nurses may not be continuously aware about the upper limits or did not
answer honestly.
Twenty-three nurses answered that they find the alarms for the oxygen saturation (very)
useful. They stated that the use of the alarms is more important than the (possible) negative
noise effect on the patient.
Twenty-one nurses stated that they did not need more authorization regarding the pulse
oximetry alarm limits, because they were already allowed to adjust the limits when the situation demands this. One nurse stated that he would like to have more authority regarding
the alarm limits, so that the limits can be adjusted to the individual needs of the patient. The
nurse mentioned that this is already done frequently, but that it is not according to the protocol. Another nurse suggested defining standard profiles of patient groups with specified alarm
limits and add these profiles to the monitor to make sure that alarm limits are set correctly.
6.3.4 Suggestions for improvement
In the final question, where suggestions for improvement were asked, one nurse advised to
change nothing to the current situation. This nurse stated that it has been proven that too
many alarm leads to numbness of nurses, even in case of loud alarms. Apparently this nurse
felt that the amount of alarms is ok, and he was probably afraid that any change would lead
to more alarms. Another nurse complained that the loud alarm for desaturations is too loud.
Two nurses made a suggestion about the lower alarm limit. Both suggestions stated that the
lower limit should be decreased because in utero, the foetuses have an oxygen saturation of
75%. One nurse suggested to determine individual alarm limits for patients with BPD.
02-08-12 11:37
116
CHAPTER 6
Several nurses suggested to develop the possibility to connect SpO2, alarm limits, and FiO2
to each other. According to the nurses this improvement would make it possible for the
patient monitor to automatically set the right alarm limits when the FiO2 supplied to the
patient changes from 21 to >21% and vice versa. It was also suggested to change the colour of
the SpO2 value on the monitor when the patient receives supplemental oxygen therapy.
6.4 Discussion
This chapter discusses the results of a survey about the assessment of oxygen requirements in ELBW infants performed among 24 nurses of the NICU of the Erasmus Medical
Centre - Sophia Childrens Hospital in Rotterdam. The results show that the SpO2 level is
an important parameter for the assessment of oxygen requirement, but that nurses critically judge the reliability of the SpO2 value and take into account multiple sources of information before performing actions to recover SpO2 within the target range. Nevertheless,
the results also show that the sources used for decision making and the actions taken with
respect to control of oxygenation are not well defined, vary among the nurses, and are quite
subjective.
When performing a survey, it should be realised that answers of subjects may possibly be
biased. The reasons for the bias in our questionaire could be related to e.g., the fact that
subjects (subconsiously) feel the need to give publically acceptable answers.
While a survey among 24 nurses is relatively small, even with this number of subjects a
variability in approaches is already shown. Though some sources of information and actions were mentioned multiple times by various nurses, the answers show that there is
no systematic approach in the (sequence of) actions performed by the nursing staff with
respect to the control of oxygenation in ELBW infants. Furthermore, our results are in
agreement with the results of earlier performed surveys aimed at obtaining information
about the knowledge and opinions of (alarm limits for) pulse oximetry, and bloodtransfusion practices.11-13, 17, 19, 21, 23, 25, 26, 28
One of the most recent survey studies, performed by Solberg et al.,23 was held among five
regional hospitals in Norway and aimed to document nurses opinion about their assessment
for oxygen requirements in preterm infants. Their results showed that 95.5% of the nurses
included the skin colour, 81% of the nurses used pulmonary auscultation, 82% checked for
chest rise, and 79.3% assessed the childs synchrony with the ventilator. They concluded that
assessment of oxygen requirement is based on insufficient information and education is
needed. Unfortunately, none of the other surveys, either executed in a single centre,17, 19, 28
or in multiple centres,11-13, 21, 23, 25, 26 provide (motivations for) actions performed to control of
oxygenation.
To overcome variability in the sources used for decision making and in the (sequence of) actions performed by the nursing staff to restore or maintain adequate oxygenation, there are
two studies which focused on the development of protocols.29, 30 The protocol described by
02-08-12 11:37
117
one of these studies30 was implemented in several NICUs worldwide, and actually showed
a reduction in the incidence of ROP. In the studys centre there is no protocol prescribing
which (sequence of) actions need to be performed to control oxygenation. We advise to
implement a protocol for the control of oxygenation in neonatal intensive care to increase
standardization. This protocol could be similar to already existing protocols,29, 30 but we suggest that a more detailed protocol may improve manual control of oxygenation even more.
The beneficial effect of and the need for education and protocols was also emphasised by
Attin et al.,17 Rodriguez et al.,13 and Nghiem et al.21 However, the development of a protocol
is not that simple. A paradox in the control of oxygenation is that medical and nursing care
aiming to prevent adverse events like desaturation of the preterm infant, actually can lead
to adverse events. For instance, tracheal suctioning helps to maintain a patent airway, but
because it is a stressful procedure, it is frequently associated with desaturation and other
adverse events.31 This kind of controversies makes it harder to provide guidelines and protocols for healthcare professionals telling them how to control oxygenation of the patient.
Today, optimal target levels for SpO2 in ELBW infants are under debate.32, 33 Due to this
lack of consensus, the protocols for target levels for SpO2 and the matching pulse oximetry
alarm limits differ between centres18, 34 and even within one department.2, 21, 35, 36 Deviations
from the protocol for pulse oximetry alarm limits were seen in the NICU of the Erasmus
MC - Sophia Childrens hospital as well (Chapter 5).
However, based on the answers in the questionnaire, it is questionable whether NICU personnel sees the deviations from the protocol for alarm limits as something undesirable.
The answers suggest that, although there is a protocol for alarm limits, the nursing staff is
allowed to adjust the alarm limits (after agreement of the medical staff). We want to emphasise that, when the decisions is made to make use of non-protocolled alarm limits, it is
important to register both these alarm limits and the motivations for using them. Otherwise there will be a lack of overview of the used alarm limits, which increases the risk for
insufficient supplemental oxygen therapy.
To conclude, a change in SpO2 is for nurses the most important trigger to adjust FiO2.
However, in their decision for the adjustment and step size of the FiO2 adjustment, nurses
take into account multiple sources of information. Besides the FiO2 adjustments, nurses
perform other actions to recover and maintain adequate oxygenation. Both the sources and
the (sequence of) actions are not well defined, are quite subjective, and vary among NICU
personnel. The necessity of this variation in clinical practice is questionable and asks for the
development of clear protocols to define what sources of information and what (sequence
of) actions should be performed with respect to supplemental oxygen therapy. Furthermore, current developed devices for automatic control of oxygenation are based on a single
input parameter (SpO2) and can only change FiO2 to recover SpO2 levels. These devices do
not include other monitored parameters, e.g., heart rate and respiration rate. Hence the
impact of these devices should be known before they are implemented in clinical practice.
02-08-12 11:37
118
CHAPTER 6
6.5 References
than 28 weeks gestation: The AVIOx study. Pediatrics.
2006;118(4):1574-1582.
J Perinatol. 2006;26:337-341.
3. Sink DW, Hope SAE, Hagadorn JI. Nurse:patient
ratio and achievement of oxygen saturation goals in
premature infants. Arch Dis Child Fetal Neonatal Ed.
2011;96(2):F93-98.
Bhutani VK, Taube JC, Antunes MJ, Delivoria4.
Papadopoulos M. Adaptive control of inspired oxygen delivery to the neonate. Pediatric Pulmonology.
1992;14(2):110-117.
5. Sun Y, Kohane IS, Stark AR. Computer-assisted adjustment of inspired oxygen concentration improves
control of oxygen saturation in newborn infants
requiring mechanical ventilation. Journal of Pediatrics.
1997;131(5):754-756.
Herberts T, Miksch S, et al. Automatic control of the
inspired oxygen fraction in preterm infants: A randomized crossover trial. American Journal of Respiratory
and Critical Care Medicine. 2004;170(10):1095-1100.
7. Claure N, DUgard C, Bancalari E. Automated adjustment of inspired oxygen in preterm infants with
trial. Journal of Pediatrics. 2009;155(5):640-645
Conference of the IEEE, EMBS. Minneapolis,
Minnesota, USA; 2009. p. 3079-3082.
9. Tehrani FT, Abbasi S. Evaluation of a computerized
system for mechanical ventilation of infants. Journal of
2004;114(5):1341-1347.
11. Avery ME, Tooley WH, Keller JB, Hurd SS, Bryan MH,
Cotton RB, et al. Is chronic lung disease in low birth
weight infants preventable? A survey of eight centers.
Pediatrics. 1987;79(1):26-30.
12. Levy GJ, Strauss RG, Hume H, Schloz L, Albanese MA,
Blazina J, et al. National survey of neonatal transfusion practices: I. Red blood cell therapy. Pediatrics.
1993;91(3):523-529.

14. Vijayakumar E, Ward GJ, Bullock CE, Patterson ML.
Pulse oximetry in infants of <1500 g birth weight on
supplemental oxygen: a national survey. J Perinatol.
1997;17(5):341-345.
15. Alberdi E, Gilhooly K, Hunter J, Logie R, Lyon A,
McIntosh N, et al. Computerisation and decision making in neonatal intensive care: A cognitive engineering investigation. Journal of Clinical Monitoring and
Computing. 2000;16(2):85-94.
16. Alberdi E, Becher JC, Gilhooly K, Hunter J, Logie R,
Lyon A, et al. Expertise and the interpretation of
computerized physiological data: implications for the
design of computerized monitoring in neonatal intensive care. International Journal of Human Computer
Studies. 2001;55(3):191-216.
17. Attin M, Cardin S, Dee V, Doering L, Dunn D, Ellstrom
K, et al. An educational project to improve knowledge
related to pulse oximetry. American Journal of Critical
Care. 2002;11(6):529-534.
19. Popovich DM, Richiuso N, Danek G. Pediatric health
care providers knowledge of pulse oximetry. Pediatric
Nursing. 2004;30(1):14-20.
20. Rdiger M, Tpfer K, Hammer H, Schmalisch G, Wauer
RR. A survey of transcutaneous blood gas monitoring
among European neonatal intensive care units. BMC
pediatrics. 2005;5(1):30-36.
Pediatrics. 2008;121(5):e1039-1046.
22. Morriss Jr FH, Abramowitz PW, Carmen L, Wallis AB.
Nurses dont hate change-survey of nurses in a neonatal intensive care unit regarding the implementation,
use and effectiveness of a bar code medication administration system. Healthcare quarterly. 2009;12:135-140.
23. Solberg MT, Hansen TWR, Bjrk IT. Nursing assessment during oxygen administration in ventilated
preterm infants. Acta Paediatrica. 2010;100(2):193-197.
24. Stoneham MD, Saville GM, Wilson IH. Knowledge
about pulse oximetry among medical and nursing staff.
The Lancet. 1994;344(8933):1339-1342.
02-08-12 11:37
119
25. Ellsbury DL, Acarregui MJ, McGuinness GA, Klein

JM. Variability in the use of supplemental oxygen for
bronchopulmonary dysplasia. Journal of Pediatrics.
2002;140(2):247-249.
26. Fouzas S, Politis P, Skylogianni E, Syriopoulou T,
Priftis KN, Chatzimichael A, et al. Knowledge on pulse
oximetry among pediatric health care professionals:
A multicenter survey. Pediatrics. 2010;126(3):e657-662.
27. Rao H, May C, Hannam S, Rafferty GF, Greenough
A. Survey of sleeping position recommendations
for prematurely born infants on neonatal intensive
care unit discharge. European Journal of Pediatrics.
2007;166(8):809-811.
28. Kruger PS, Longden PJ. A study of a hospital staffs
knowledge of pulse oximetry. Anaesthesia and Intensive care. 1997;25(1):38-41.
2008;95(4):306-310.
Pediatrics. 2003;111(2):339-345.
31. Pritchard MA, Flenady V, Woodgate PG. Preoxygenation for tracheal suctioning in intubated, ventilated
newborn infants. Cochrane Database Syst Rev.
2001;3(1).
32. Higgins RD, Bancalari E, Willinger M, Raju TNK.
Executive summary of the workshop on oxygen in
neonatal therapies: controversies and opportunities
for research. Pediatrics. 2007;119(4):790-796.
Pediatrics. 2002;110(3):615-616.
34. Tin W, Wariyar U. Giving small babies oxygen:
50 years of uncertainty. Seminars in Neonatology.
2002;7(5):361-367.
36. Mills BA, Davis PG, Donath SM, Clucas LM, Doyle LW.
Improving compliance with pulse oximetry alarm limits
for very preterm infants? Journal of Paediatrics and
Child Health. 2010;46(5):255-258.
02-08-12 11:37
120
02-08-12 11:37
CHAPTER 7
Defining hazards of
supplemental oxygen
therapy in neonatology
using the Failure Mode
and Effects Analysis
(FMEA) tool
A.C. van der Eijk, D. Rook, J. Dankelman, B.J. Smit
02-08-12 11:37
122
CHAPTER 7
Human beings, and especially preterm infants, are extremely sensitive for either too
little as well as too much oxygen in their tissues. It is therefore of great importance
that oxygenation of preterm infants is monitored closely, and kept within a (narrow)
safe range. This chapter is based on the article titled Defining hazards of supplemental
oxygen therapy in neonatology using the Failure Mode and Effects Analysis (FMEA) tool
(accepted for publication by the American Journal of Maternal Child Nursing), and evaluates the process of supplemental oxygen therapy in very preterm infants hospitalised on
the neonatal intensive care unit.
OBJECTIVE To prospectively evaluate hazards in the process of supplemental oxygen
therapy in very preterm infants hospitalised in a NICU.
02-08-12 11:37
An FMEA on supplemental oxygen therapy
123
7.1 INTRODUCTION
As a result of on-going developments in neonatal care, long term outcome of preterm infants is still improving.1 However, innovations in neonatal intensive care units (NICUs) are
accompanied with an increased complexity of daily practice. In complex environments like
NICUs, where people from different disciplines work in shifts and with high-tech equipment, the occurrence of errors is, unfortunately, undeniable.
In the year 2000, the Institute of Medicines published To Err Is Human: Building a Safer
Health System.2 Thanks to this report, patient safety and the reduction of medical errors
obtained major priority in health care. In the NICU, where care focuses on a specific and
very vulnerable patient group, errors may have even more impact than in other disciplines of
health care. Therefore, although much work has already been performed, each opportunity
to increase (patient)safety in the NICU should be seized.3
In 2005, a NEOnatology System for Analysis and Feedback on medical Events (NEOSAFE) was
introduced in the Netherlands to establish specialty-based learning from (near-)accidents.4
The (near-)accidents are reported by NICU personnel on a voluntary basis. In the NICU of
the Erasmus Medical Centre - Sophia Childrens Hospital in Rotterdam, 909 (near-)
accidents were reported in 2010. While only 58 (6%) of the reports were related to supplemental
oxygen therapy, these 58 reports were responsible for 90% of the (near-)accidents that
were categorised as most risky.
In supplemental oxygen therapy, a gas mixture with >21% of oxygen is supplied to the patient
via (mechanical) ventilation. Due to the immaturity of the preterm infants lungs, supplemental oxygen therapy is often needed immediately after birth to reach and maintain adequate
oxygenation. Unfortunately, supplemental oxygen therapy has a very narrow therapeutic
range. Both too high and too low blood oxygen levels may have severe consequences for
the outcome of these infants.5
The aim of this study was to evaluate hazards in the process of supplemental oxygen
therapy in very preterm infants hospitalised on the NICU prospectively, and to provide recommendations for improvement. To evaluate the process, a Failure Mode and Effects Analysis (FMEA) tool was used.
In complex environments like

NICUs, the occurrence of errors is,
unfortunately, undeniable.
02-08-12 11:37
124
CHAPTER 7
7.1.1 FMEA
To prevent failure modes in high risk industries, like aviation, and nuclear power plants,
engineers analyse products and processes for potential hazards. Since the mid-1960s, FMEAs
have been performed to standardise the approach in failure mode detection. Although
performing an FMEA is often considered as time-consuming and needs organizational
commitment, the method is a useful tool for structural analysis, identification of (unnoticed) errors, and has been demonstrated to increase safety.6 Unfortunately, FMEAs
could not be applied easily to healthcare situations. Therefore, several adaptations on the
methodology of FMEAs have been considered, developed and used.7-9
7.2 Methods
On the NICU of the Erasmus Medical Centre - Sophia Childrens Hospital in Rotterdam,
the Netherlands, an FMEA on supplemental oxygen therapy in very preterm infants was
performed between May and August 2011. Approval from the hospital research ethics board
was not necessary because no patients were involved in the study.
The Erasmus MC has a level-III-c10 NICU with 30 beds divided over three subunits. The
nursing staff works in shifts of 8 hours. After each shift, half an hour of handover takes
place. During the handover, nurses inform each other verbally about the (clinical) situation
of the infants. Moreover, relevant patient data is collected and validated at least once per
hour in the Patient Data Management System (PDMS) by the nursing and/or medical staff.
The FMEA consisted of five steps (Table 7.1).9, 11 With the FMEA-team, several plenary
sessions were held to complete each of the steps. Besides the actual FMEA-team, other
employees were asked to join in a separate panel of FMEA advisors: the FMEA support-team
(as suggested by Ashley et al.12). This support-team was included in e-mail conversations,
but was not present at the plenary meetings.
The FMEA-team analysed the process of supplemental oxygen therapy and subdivided it
in (sub)sub-steps. For each (sub)sub-step potential failure modes, and possible causes and
effects of these failure modes were identified. The team was encouraged to identify as
many failure modes, causes and effects as possible by brainstorming.
After defining the hazards, i.e. the failure mode-cause-effects combinations, each team
member individually accredited three scores for each of these hazards by using the Hazard
Scoring Matrix (Table 7.2). The scores were provided for severity (S), likelihood of occurrence (O), and detectability (D). Based on these scorings, for each hazard a risk priority
number (RPN = S x O x D) was calculated. For each hazard the overall RPN was calculated
by taking the mean of all RPNs that were given to this specific hazard by each of the individual team members. The ten hazards with the highest overall RPN were defined as the
top ten hazards. Finally, the team made recommendations to minimise the hazards.
02-08-12 11:37
125
Table 7.1 The 5 steps forming the FMEA

Step
Object
Define a topic
Assemble a multidisciplinary team
Analyse the process by dividing the process in (sub)steps
Perform Hazard analysis

Identify Hazards, i.e. potential failure modes,
and possible causes and effects of these failure modes
Accredit risk priority numbers (RPN) for each of the identified Hazards
Determine the top ten hazards
Develop recommendations to resolve Hazards
Table 7.2 Hazard Scoring Matrix
Rating
Severity of hazard (S)
Occurrence of hazard (O)
Detectability of hazard (D)
1-3
No consequences
Rare
Always detected
4-6
Minor consequences
Occasional
Probably detected
7-8
Temporary consequences
Frequent
Low probability of detection
Permanent consequences
Often
Probably undetected
10
Fatal consequences
Always
Undetected
The Risk Priority Number (RPN) is calculated by multiplying the Severity (S), Occurrence (O) and
Detectability (D) of the hazard. For example, a hazard that has minor consequences (rate 4), that
occurs often (rate 9) and that has a flow probability of detection (rate 7) has a RPN of 4 x 9 x 7 = 252.
02-08-12 11:37
126
CHAPTER 7
7.3 Results
7.3.1 Step 1. Defining the topic
In April 2011 the topic for the FMEA was defined in detail. It was decided to focus on
supplemental oxygen therapy for very preterm infants with a postmenstrual age (PMA) <30
weeks, who are admitted to the NICU of the Erasmus MC. In the NICU, every bed space has
a separate outlet in the wall for oxygen. Devices for respiratory support are connected to
these outlets with a tube. It was decided to include the devices for respiratory support in the
FMEA, but to exclude the supply of oxygen from the central oxygen storage to the device for
respiratory support.
7.3.2 Step 2. Team assembly
A multidisciplinary team of six members and two team leaders was composed in May 2011.
The members of the team (1 neonatologist, 1 nurse practitioner, 2 NICU-nurses, 1 leading NICU-nurse, and 1 PhD student) were instructed about the FMEA procedure by the
team leaders, and by a consultant for quality improvement and patient safety from the
directorate patient care of the studys centre. The team was chosen such that all parties
relevant in the process of supplemental oxygen therapy in the NICU were represented. In a
four month period, the team had seven meetings of approximately two hours. In between
meetings, communication via e-mail was used to spread relevant documents. The FMEA
support team was included in the email conversation, and existed of seven members.
7.3.3 Step 3. Process analysis
During the process analysis, all process-steps involved in supplemental oxygen therapy
were defined (Figure 7.1). To prevent unnecessary complexity of the FMEA, only situations
that were directly related to supplemental oxygen therapy in preterm infants with a GA <30
weeks admitted to the NICU were included in the analysis. Where applicable, the processsteps were divided in (sub)sub-steps. For example, for process-step 3, eleven sub-steps were
defined (Figure 7.2).
Step 1
Nurse prepares admission of neonate to NICU
Step 2
Neonate arrives at NICU
Step 3
Neonate is hospitalised on NICU
Step 4
Neonate is discharged from NICU
Figure 7.1
The four main steps in the process of supplemental oxygen therapy as analysed during the FMEA.
02-08-12 11:37

or
Caregiver
performs
tasks to
recover SpO2
Neonate is
judged by
caregiver
Normal alarm
for low SpO2
(<85 %)
Normal
alarm for high
SpO2 (>93 %)
FiO2 from
roomair to
>21 %
FiO2 from
>21 % to
room air
Upper alarm
limit from 1 00
to 93 % for
PMA<3 7 wks,
or from 1 00
to 96% for
PMA 3 7
or
or
Upper alarm
limit to 1 00%
FiO2
increased
FiO2
decreased
FiO2 from
>21 % to
>21 %
Loud alarm for

very low SpO2
(<70%)
or
Respiratory
settings
adjusted
To step 4
or
Alarms 3
minutes
suppressed by
snooze button
Respiratory
device is
changed
Neonate
requires other
type of
respiratory
device
Respiratory
support is
stopped
FiO2 adjustment
while SpO2 is
within target
range
Transport
monitor
settings
adjusted
Neonate is
prepared for
transport
Nurse
records
adjustments
in PDMS
Nurse
adjusts
alarm limits
in monitor
MD records
alarm limits
in patient
chart and
PDMS
MD orders
deviating alarm
limits for SpO2
figure 7.2 A simplified flowchart of the sub(sub)steps of process step 3: Neonate is hospitalised on NICU.
Subsubstep
Substep
Process step
Situation
Legend
Normal
alarm for
interference of
pulse oximeter
Neonate in
incubator
From step 2
Nurse
checks
respiratory
device and
monitor
settings
Neonate is
handed over to
other nurse
Nurse
records
alarm limits
in PDMS
Nurse
adjusts limits
according to
protocol
(92-96 for
>21 % or 921 00 for
room air)
Neonate
reaches PMA of
3 7 wks
127
02-08-12 11:37
128
CHAPTER 7
7.3.4 Step 4. Hazard analysis

In step four of the FMEA, the actual hazard analysis was performed. In total, 134 hazards
were defined. The overall RPN ranged from a minimum score of 45 to the highest RPN of
507. The 10 hazards with the highest overall RPN are shown in Table 7.3. These hazards were
all related to incorrect adjustment of the fraction of inspired oxygen (FiO2), incorrect alarm
limit settings for oxygen saturation measured by pulse oximetry (SpO2), or incorrect pulse
oximetry alarm limit settings on patient monitors for temporary use.
7.3.4.1 Incorrect adjustments of FiO2
Incorrect adjustments of FiO2 can result in periods of hyperoxia and hypoxia. The severity
of this effect is difficult to quantify, but probably depends on the length and frequency of
the periods. Incorrect adjustments of FiO2 are defined by the combination of frequency and
step size in FiO2 adjustments. For example, multiple adjustments with a small step size, or
a single adjustment with a large step size may have a similar effect on the patient, thus both
situations were defined as too quick.
Too slow decreasing of FiO2 was thought to result from the fact that the risk for hyperoxia
might not be sufficiently realised by the nursing staff and the lack of a protocol describing
the step size and timing of FiO2 adjustments. The lack of a protocol was also thought to be
the cause for the too quick adjustments in FiO2. Another cause for too quick increases was
the high workload of the nursing staff. Because the nurse has to perform other tasks (simultaneously), he might adjust FiO2 without waiting for the result of previous adjustment.
7.3.4.2 Incorrect pulse oximetry alarm limit settings
Just as incorrect adjustments in FiO2, no or incorrect adjustment of alarm limits for SpO2
may also lead to periods of hyperoxia or hypoxia. In the studys centre, the protocol for alarm
limits for patients with a PMA <37 weeks, prescribes alarm limits of 85 - 100% when patients
are on room air, and 85 - 93% for FiO2 >21%. Thus, when the need for oxygen increases
from 21% to >21% or vice versa, the alarm limit need to be adjusted. The causes for incorrect
alarm limits were thought to lie in the fact that nurses forget to adjust alarm limits, and that
nurses deliberately do not adjust alarm limits because, for instance, he supposes the need for
supplemental oxygen will not last long.
Another situation in which incorrect alarm limits can occur is related to conscious deviations
from the protocol. In consultation with the physician, nurses can deviate from the protocol
for alarm limits for medical reasons. However, when the non-protocolled alarm limits are not
recorded, other NICU personnel may not be aware of the deviating alarm limits, and furthermore, cannot detect who decided to adjust the alarm limits for what reason.
When a preterm infant reaches a PMA of 37 weeks, the protocol for alarm limits change. The
protocol for this patient group prescribes 92 - 100% for patients on room air, and 92 - 96%
for FiO2 >21%. It often happens that the alarm limits are not adjusted at 37 weeks of PMA,
because the responsible nurse is not aware of the actual age of the patient.
02-08-12 11:37
129
7.3.4.3 Incorrect pulse oximetry alarm limit settings on patient

monitors for temporary use
Another hazard was the incorrect alarm limit setting for SpO2 on patient monitors for
temporary use. Next to the standard monitor belonging to each incubator, there are
situations in which the patient is monitored by separate monitors. For instance, during
transport, during a surgical procedure, or when an MRI is performed, the patient is disconnected from its own monitor, and connected to another monitor specifically used for
transport, surgery, or MRI. These monitors are used for multiple patients. A consequence
of the variety in patients is that monitor settings are not always correctly set for the specific
patient. The effects of these incorrect settings depend on the actual setting, but can, for
instance, cause unnoticed hypoxia or hyperoxia. The reasons for incorrect monitor setting
are probably the variety in default settings of the monitors, and the fact that the nurse
and/or physician forget to adjust monitor settings when connecting the patient.
7.3.5 Step 5. Develop Risk Reduction Methods
For each hazard in the top ten, it was determined whether the hazard should be eliminated, controlled or accepted. Afterwards recommendations were defined. Several of these
recommendations were applicable on multiple hazards. For instance, the fact that caregivers do not recognise the potential harm of hyperoxia could be improved by a structured
educational program. In the studys centre, this encompassed the inclusion of supplemental
oxygen therapy as a subject in the education of new NICU nurses, and in the obligatory
monthly training for all NICU nurses. The implementation of a protocol for FiO2 adjustments, like described by Chow et al.13 and Wilkinson & Andersen,14 could be the solution
to minimise incorrect FiO2 adjustment, and increase standardisation in patient care. The
oxygen blender (PM5200, Precision Medical Inc, Northampton, PA, USA) used for respiratory support with a nasal prong, or during manual ventilation, is associated with (near-)
accidents because the pre-set FiO2 is difficult to read. The possibility of a more user-friendly
oxygen blender will be investigated.
The frequency of incorrect alarm limit settings may decrease by introducing technical
solutions. In the most extreme solution, the protocol for alarm limits is implemented in
the patient monitor to make sure alarm limits are adjusted automatically. However, to be
able to achieve this, the monitor needs to be informed about the actual FiO2 supplied to the
infant. In the studys centre, the monitors cannot receive information from ventilators yet.
Other solutions could be related to automatic reminders in the PDMS to prompt caregivers
to verify pre-set alarm limits.
Incorrect setting of temporarily used patient monitors will probably be solved by
changing the default alarm limits of all monitors to the same, most conservative, settings.
Also, checking of the alarm limits is nowadays included in the checklist for transportation
of neonates. For each of the recommendations made in the FMEA, individuals responsible
for implementing were identified.
02-08-12 11:37
130
CHAPTER 7
Table 7.3 Top 10 hazards

#
Hazard
Category
Too slow reduction

of FiO2, resulting in
hyperoxia
Incorrect FiO2
adjustments
No adjustment of
SpO2 alarm limits
when reaching GA of
37 weeks, resulting in
hypoxia
Overall RPN
Action
Recommendation
507
Controlled
Protocol on manual
FiO2 adjustments
Incorrect alarm
limit settings
502
Eliminated
Technical
improvements in
patient monitors
No reduction of FiO2,
resulting in hyperoxia
Incorrect FiO2
adjustments
501
Controlled
Protocol on manual
FiO2 adjustments
Education on hazard
of supplemental
oxygen therapy
Incorrect alarm limits

of monitor during
transport, resulting in
hyperoxia or hypoxia
Incorrect alarm
limit settings
in temporarily
monitor
488
Eliminated
Check alarm limits

is added to
the checklist used
before transport
No adjustment
of SpO2 alarm
limits when going
from room air to
supplemental oxygen,
resulting in hyperoxia
Incorrect alarm
limit settings
486
Eliminated
Technical improvements in respiratory

devices and patient
monitors
No adjustment of
SpO2 alarm limits
when reaching GA of
37 weeks, resulting in
unnecessary alarms
for hyperoxia
Incorrect alarm
limit settings
434
Eliminated
Technical improvements in patient

monitors
Too fast increase

FiO2, resulting in
hyperoxia
Incorrect FiO2
adjustments
408
Controlled
Technical
improvements in
patient monitors
Physician does not

record deviating
alarm limits
Incorrect alarm
limit settings
402
Temporarily
accepted
Create awareness
for the need to fill in
patient records completely and correctly
Too fast reduction

of FiO2, resulting in
hypoxia
Incorrect FiO2
adjustments
402
Controlled
Protocol on manual
FiO2 adjustments
Education on
potential hazard of
supplemental oxygen
therapy
10
Nurse does not

record adjustments of
alarm limits in PDMS
Incorrect alarm
limit settings
376
Controlled
Technical
improvements in
patient monitors
The Hazard Scoring is represented by the overall Risk Priority Number (RPN), i.e. the mean RPN
calculated over all the RPNs given to the hazard by each of the team members.
02-08-12 11:37
131
7.4 Discussion
This article discusses the FMEA performed on supplemental oxygen therapy in very preterm
infants admitted to a NICU. The FMEA resulted in a top ten hazards for which recommendations were suggested. These hazards could be categorised in three topics, namely incorrect adjustment of the fraction of inspired oxygen (FiO2), incorrect alarm limit settings for
oxygen saturation measured by pulse oximetry (SpO2), and incorrect pulse oximetry alarm
limit settings on patient monitors for temporarily use.
With respect to FiO2 adjustments, protocols describing the step size and timing of FiO2
adjustments are lacking in the studys centre, and FiO2 adjustments are now highly dependent on the workload, personal training, and opinion of the caregiver. Literature on this topic
is scarce and no randomised studies on how and when to adjust the FiO2 exist. To improve
manual FiO2 adjustment, the FMEA-team will develop a protocol for FiO2 adjustments based
on current literature (e.g., Chow et al.,13 and Wilkinson & Andersen14), and expertise.
Another recommendation is to educate caregivers on the hazards of both hypoxia and
hyperoxia in preterm infants. However, the effect of education may be transient, and therefore, a structural repetitive program will be implemented.
Incorrect alarm limits for SpO2 may be resolved with technical solutions. First, awareness
by using reminders in the PDMS or the patient monitors could be introduced. These reminders will prompt caregivers to verify alarm limits. Ultimately, automated adjustment
of alarm limits for SpO2, based on the actual FiO2, should be constructed. Unfortunately,
currently used respiratory devices and patient monitors are not able to send and receive
information to/from each other. Solutions will be discussed with the industry so recommendations can be taken into account for future developments. Finally, the hazard of incorrect settings in patient monitors was tackled by changing the default alarm limits of all
monitors to equal, most conservative settings. Also, in the checklist for transportation of
neonates, the task check the alarm limits was included.
Implementation of the recommendations has proven to be difficult. Some of the recommendations are not possible to implement immediately due to technical limitations, other
recommendations are part of larger changes in the NICU (e.g., implementation of checklists
and protocols), and will be implemented in the near future. Certainly, the knowledge and
understanding obtained during performing the FMEA will be shared with other NICUs.
The study was performed in one centre, and the question remains whether the hazards
are generalizable to other NICUs. However, Nghiem et al.15 showed that 32% of the centres
in the US do not have a protocol for SpO2 alarm limits, which may also implicate absence
of protocols for FiO2 adjustment. Furthermore, in centres that have a protocol for alarm
limits, SpO2 is frequently outside of the alarm limits.16-18
02-08-12 11:37
132
CHAPTER 7
Compared to other processes that were prospectively analysed for hazards, like the administration of medication,19-24 in the process of supplemental oxygen therapy no clear endpoint can be defined. The process of supplemental oxygen therapy is continued until the
patient is discharged from the NICU. This complexity made it difficult to quantify the RPN
for hazards. For instance, in the hazard analysis, the severity of potential consequences was
scored ranging from no consequences to the patient to fatal consequences for the patient.
However, most failure modes in this FMEA could result in either hypoxia or hyperoxia, with
unknown (long term) consequences. It is known that both hypoxia or hyperoxia can cause
permanent damage,25 but it is impossible to determine whether this damage will actually
occur and whether this damage will be either temporarily, permanent or fatal.
Besides the difficulty of determining the occurrence and severity of hazards, Ashley and
Armitage26 illustrated that determination of RPNs of hazards is also subjective. To minimise
subjectivity, it was decided to first determine RPNs individually, and to discuss the results
afterwards in a plenary session. Although the RPNs of the hazards differed slightly between
individuals, there was overall consensus about the hazards in the final top ten.
To quantify the effects of implementation of recommendations, the FMEA-team is dependent of the voluntary reports of (near-)accidents. The number of reports is not only
dependent on the actual (near-)accidents, but also on the willingness, and focus of caregivers
to detect, and report the (near-)accidents.3 Since these reports of (near-)accidents thus only
provide an indication for the actual number of (near-)accidents, it is impossible to compare
failure rates before and after this FMEA. However, despite the lack of quantification, performing an FMEA on supplemental oxygen therapy in very preterm infants hospitalised on
the NICU provided interesting insights, and is expected to enhance patient safety
7.4.1 Lessons learnt by the FMEA-team
Performing the FMEA was a new experience for most of the team members. Though, despite
the inexperienced team, the FMEA was considered as an easy to learn, and intuitive method.
The discussions in the plenary meetings were very open minded, and, thanks to the multidisciplinary character, diverse ideas and visions were mentioned.
The team members all agreed that performing an FMEA is time consuming. One of the
largest challenges for the team leaders was to plan appointments for plenary sessions which
all of the team members were able to join. However, despite the time issue, it is expected
that the FMEA tool will be used more frequently in the near future for other processes in
the studys centre.
To conclude, performing an FMEA on supplemental oxygen therapy in very preterm infants
hospitalised on the NICU provided interesting insights. The main hazards were all related
to incorrect adjustment of the FiO2, incorrect alarm limit settings for SpO2, and incorrect
pulse oximetry alarm limit settings on patient monitors for temporarily use. Thanks to the
structured approach of the FMEA, and the multidisciplinary team, several recommendations for improvement were made and implemented.
02-08-12 11:37
133
7.5 References
1. Saigal S, Doyle LW. Preterm birth 3: An overview of
mortality and sequelae of preterm birth from infancy
to adulthood. The Lancet. 2008;371(9608):261269.
2. Kohn LT, Corrigan JM, Donaldson MS. To err is human:
building a safer health system. Washington DC:
The National Academy Press; 2000.
3. Edwards W. Patient safety in the neonatal intensive
care unit. Clinics in perinatology. 2005;32(1):97-106.
4. Snijders C, Van der Schaaf T, Klip H, van Lingen W,
Molendijk A. Feasibility and reliability of PRISMAMedical for specialty-based incident analysis. Quality
and Safety in Health Care. 2009;18(6):486-491.
6. McDermott RE, Mikulak RJ, Beauregard MR. The
basics of FMEA. Portland: Productivity Press; 1996.
7. DeRosier J, Stalhandske E, Bagian JP, Nudell T. Using
health care failure mode and effect analysis: the VA
National Center for Patient Safetys prospective risk
analysis system. Joint Commission Journal on Quality
and Patient Safety. 2002;28(5):248-267.
8. Habraken M, Van der Schaaf T, Leistikow I, ReijndersThijssen P. Prospective risk analysis of health care processes: a systematic evaluation of the use of HFMEA
in Dutch health care. Ergonomics. 2009;52(7):809-819.
9. Van der Hoeff NWS. Theory and practice of in-hospital
patient risk management Delft: Thesis, Delft University of Technology; 2003.
2004;114(5):1341-1347.
11. Van der Hoeff NWS. www.patientveiligheid.org/
handleidingen. Published 2011.
12. Ashley L, Armitage G, Neary M, Hollingsworth G. A
practical guide to Failure Mode and Effects Analysis
in health care: Making the most of the team and its
meetings. Joint Commission Journal on Quality and
Patient Safety. 2010;36(8):351-358.
of prematurity in very low birth weight infants? Pediatrics. 2003;111(2):339-345.
14. Wilkinson D, Andersen C. Bedside Algorithms
for Managing Desaturation in Ventilated Preterm
Infants: A Randomised Crossover Trial. Neonatology.
2008;95(4):306-310
Pediatrics. 2008;121(5):e1039-1046.

J Perinatol. 2006;26:337-341.
than 28 weeks gestation: The AVIOx Study. Pediatrics.
2006;118(4):1574-1582.
18. Armbruster J, Schmidt B, Poets CF, Bassler D. Nurses
compliance with alarm limits for pulse oximetry:
qualitative study. J Perinatol. 2009;30:531-534.
19. Adachi W, Lodolce AE. Use of failure mode and effects
analysis in improving the safety of iv drug administration. American Journal of Health-System Pharmacy.
2005;62(9):917-920.
20. Apkon M, Leonard J, Probst L, DeLizio L, Vitale R.
Design of a safer approach to intravenous drug infusions: failure mode effects analysis. Quality and Safety
in Health Care. 2004;13(4):265-271.
21. Cheng CH, Chou CJ, Wang PC, Lin HY, Kao CL, Su CT.
Applying HFMEA to Prevent Chemotherapy Errors.
Journal of Medical Systems. 2010:1-9.
22. Esmail R, Cummings C, Dersch D, Duchscherer G,
Glowa J, Liggett G. Using Healthcare Failure Mode and
Effect Analysis tool to review the process of ordering
and administrating potassium chloride and potassium
phosphate. Patient Safety. 2005;8(Sp):73-80.
23. Moss J. Reducing errors during patient-controlled
analgesia therapy through Failure Mode and Effects
Analysis. Joint Commission Journal on Quality and
Patient Safety. 2010;36(8):359-364.
24. Kunac DL, Reith DM. Identification of priorities
for medication safety in neonatal intensive care.
Drug Safety. 2005;28(3):251-261.
25. Askie L, Henderson-Smart DJ, Ko H. Cochrane review:
Restricted versus liberal oxygen exposure for preventing morbidity and mortality in preterm or low birth
weight infants. Evidence-Based Child Health:
26. Ashley L, Armitage G. Failure Mode and Effects
Analysis: An empirical comparison of failure mode
scoring procedures. Journal of Patient Safety.
2010;6(4):210-215.
02-08-12 11:37
02-08-12 11:37
Part 2
intravenous
therapy
02-08-12 11:37
136
02-08-12 11:37
CHAPTER 8
Flow-rate variability
in neonatal IV therapy:
what do we know
about the flow?
A.C. van der Eijk, M.F.P.T. van Rens, J. Dankelman, B.J. Smit
02-08-12 11:37
138
CHAPTER 8
Virtually all patients in a neonatal intensive care unit receive intravenous (IV) therapy.1-3
In IV therapy, parenteral nutrition, various types of drugs and/or fluids are administered
directly into the veins of the patient via an intravascular catheter or cannula. Unfortunately, IV therapy is hampered by a number of limitations. One of these limitations is
the unpredictability of the actual amount of substances delivered to the patient. This
chapter is based on the article titled: Flow-rate variability in neonatal IV therapy: what
do we know about the flow? (submitted). In this chapter the most important factors
influencing volume delivery in IV therapy are discussed.
OBJECTIVE To study which factors are responsible for flow-rate variability in IV therapy
with syringe pumps.
02-08-12 11:37
139
8.1 INTRODUCTION
Virtually all patients in a neonatal intensive care unit (NICU) receive intravenous (IV)
therapy.1-3 Because medications are administered simultaneously and intravenous access in
neonates is limited, multi-infusion is often applied.4 Due to the interaction between flows
from multiple syringe pumps, the volume delivered to the patient may vary.5, 6 This effect is
increased in IV therapy for neonates in which the required infusion flow-rate is relatively
low (0.1 ml/h to ~5 ml/h) compared to adults.7
The need for low flow-rates in neonatal IV therapy hampers the accuracy and predictability
of the actual volumes delivered to the patient.8 Several studies have described changes in
the haemodynamics and oxygenation of neonates due to the flow-rate variability in supplied drugs (e.g., dopamine or epinephrine).9-13 A study by Sherwin et al.14 demonstrated that
in preterm infants weighing 0.5 kg, only 70% of the intended drug dose had been delivered
75 minutes after the infusion began.
In addition to the required pre-programmed low flow-rates other factors influence the
flow-rate variability in IV therapy. In this review, the factors contributing to flow-rate
variability in IV therapy for newborn infants are discussed and suggestions for improvement are provided.
Because medications are administered

simultaneously and intravenous access
in neonates is limited, multi-infusion
is often applied.
02-08-12 11:37
140
CHAPTER 8
8.2 Methods
To find relevant literature on the topic, the following search strategy was used. In September 2011, the online databases PubMed and Web of Knowledge were searched using the
keywords Syringe pump, Syringe, Infusion, Infusion pump, Intravenous, Infusion line,
Stop cock, Vascular access separately and in an AND combination with Neonatal intensive care OR Neonate OR Preterm.
From the collected articles, two researchers independently selected articles (reviews, quantitative research, letters, and case studies) for relevancy based on the title and abstract. Articles focusing on the (design of) parts of the IV-administration set, flow-rate variability
causes, low flow IV therapy, and possible complications due to flow-rate variability were
considered relevant. Articles focusing on complications due to extravasation, the management of extravasation, medical errors in drug administration, drug compatibility, the
manual flushing of drugs, not written in English, and those originating before 1980 were
excluded from this overview.
8.3 Results
Forty-one articles were selected for this paper. These articles provided an overview
of factors known to influence the flow-rate variability in IV therapy. These factors
(Figure 8.1) are listed below and discussed in this overview:
1. Vertical syringe pump displacement relative to the patient
2. Syringe size and design
3. Infusion tubing size and design
4. Check valves
5. Inline filters
6. Add-on devices
7. Vascular access devices
8.3.1 Factor 1: Vertical syringe or patient displacement
The delivery of IV substances to the patient is influenced by vertical displacement of the
syringe pump relative to the patient. Four studies were found that assessed the effect of
vertical displacement for different brands and models of syringe pumps using an in vitro
set-up.8, 15-17 These studies all used 50-ml syringes, but other study characteristics differed.
The methods, materials, results, and conclusions of these studies are elaborated on in the
following paragraphs and in Table 8.1.
Lnnqvist and Lfqvist15 studied the effects on the flow-rate variability after downward and
upward displacement for five types of syringe pumps. The tip of the infusion needle was
positioned in a small container beneath the fluid surface. Neff et al.16 evaluated the effect
of upward and downward displacement for three types of syringe pumps. For each type,
two identical syringe pumps were tested. The tip of the infusion line was submerged in
sterile water to create 5.9 mmHg (0.8 kPa) pressure against the infusion to simulate venous
02-08-12 11:37
2
3
Syringe pumps with syringes
Infusion tubing
5
1
141
Height
between catheter
tip and syringe
pump
Check valves
Inline filters
Add-on device
To patient
Vascular
acces device
figure 8.1 Schematic illustration of a multi-infusion IV-administration set (not to scale). The numbers
refer to the factors discussed in this overview.
pressure. Both studies measured the delivered volume with an electronic balance to determine the bolus volume (i.e., the sudden delivery of a relatively large amount of fluid), the
aspiration volume (i.e., the sudden retraction of fluid into the IV-administration set), and
the time without flow (i.e., the period during which no fluid is delivered to the patient) after
vertical displacement.
Kern et al.17 measured the effects of a vertical downward displacement for three types of
syringe pumps. In contrast to the other two studies, the delivered volume in this study was
measured in a closed system using glass capillaries (diameter: 0.5 mm; length: 1 m). Both
Neff et al.16 and Kern et al.17 also studied the effect of the administration set compliance on
the study results. They found a negative correlation between infusion rate and time without flow. Kern et al.17 stated that the compliance of the complete IV-administration set was
linearly correlated with the effective time without flow.
The most recent study, performed by Donald et al.,8 tested the effect of an upward and
downward displacement for two types of syringe pumps. A methylene blue dye solution was
pumped into a standard tubing set and a central line. Approximately 1 meter of the tubing
was positioned next to a tape measure and the dye front was advanced onto the scale. The
study showed that for an upward displacement of 30 cm, the relative increase of the delivery
rate was sevenfold for the lowest tested pre-programmed flow-rate (2 ml/h) and twofold for
the highest pre-programmed flow-rate (20 mL/h). This increase in delivery is equivalent
to 4.3 to 9.7 g of inotrope for a solution with a concentration of 3 mg/50 ml. After the
downward displacement of the syringe pump, it took as long as 180 seconds before steady-
02-08-12 11:37
142
CHAPTER 8
Kern
et al. [17]
Lnnqvist and
Lfqvist [15]
Kern
et al. [17]
2001
2001
1997
2001
50
80
100
130
Not
specified
Fresenius
Injectomat
Fresenius
Injectomat
B. Braun
Perfusor
Fresenius
Injectomat
50
50
50
50
50
Not
specified
Not
specified
Not
specified
20
Not
specified
Infusion tubing type
Not
specified
Standard
triple-lumen
central line
5.9
Fresenius Injectomat-S, device 1

Fresenius Injectomat cp-IS, device 1
IVAC P 4000 anaesthesia, device 1
PE
infusion line
Fresenius Injectomat
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
9000911
n872296/0
g30402
9000911
n872269/0
G30402
9000911
N872296/0
G30402
9000911
n872296/0
g30402
9000911
n872296/0
g30402
Not
specified
Bd Modular
Bd Pilot C
B. Braun Perfusor Securra FT
B. Braun Perfusor FM
IVAC P 4000
B. Braun
Perfusor
tubing type N
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
9000911
n872296/0
g30402
9000911
n872269/0
G30402
9000911
N872296/0
G30402
9000911
n872296/0
g30402
9000911
n872296/0
g30402
Syringe pump
brand & type
3.7
Pressure working
against infusion
[mmHg]
Syringe filling volume

[ml]
Syringe size [ml]
Syringe brand
Vertical
displacement [cm]
30
Needle/catheter
brand & size [mm]
Neff
et al. [16]
2007
Infusion tubing
length [m]
Donald
et al. [8]
Year
Authors
[reference number]
Table 8.1 Characteristics of the studies examining the effect of the vertical displacement of syringe
pumps relative to the catheter tip
02-08-12 11:37
1.5
2
1.5
2
1.5
2
1.5
2
1.5
2
1.5
2.0 x
1.5
2
1.5
2
1.5
2
1.5
2
1.5
2
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
9000911
n872296/0
g30402
9000911
n872269/0
G30402
9000911
N872296/0
G30402
9000911
n872296/0
g30402
9000911
n872296/0
g30402
Bd Modular
Bd Pilot C
B. Braun Perfusor Securra FT
B. Braun Perfusor FM
IVAC P 4000
B. Braun
Perfusor
tubing type N
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
B. Braun Perfusor
IVAC 770
9000911
n872296/0
g30402
9000911
n872269/0
G30402
9000911
N872296/0
G30402
9000911
n872296/0
g30402
9000911
n872296/0
g30402
45.8 5.4
44.1 3.2
67.6 2.1
61.6 2.1
77.1 5.1
59.0 5.1
20.9 1.3
20.4 0.8
22.6 1.6
22.0 0.8
26.8 1.6
24.2 0.8
2.98 0.53
2.78 0.29
4.65 0.56
4.26 0.77
5.99 1.09
4.32 0.56
2
5
10
20
1.5
Bolus volume after

upward displacement
[L]
~ 50 s
~ 10 s
< 10 s
< 10 s
Number of tests
Pre-programmed
flow rate [ml/h]
Time without
flow (mean SD)
[minutes]
PE
infusion line
Aspiration volume
after downward
displacement [L]

71.18
121.16
141.08
161.01
Needle/catheter
brand & size [mm]
Standard
triple-lumen
central line
Infusion tubing
length [m]
Not
specified
Syringe pump
brand & type
10
3.90 0.35
4.72 0.30
7.17 0.45
2.15 0.233
1.83 0.15
5.17 0.40
1.35 0.12
1.12 0.08
4.05 0.17
0.72 0.05
0.65 0.02
2.17 0.12
0.38 0.03
0.48 0.02
1.05 0.03
1360
2280
520
190
1210
340
280
60
60
240
61
105
24
8
44
6.95 0.98
6.82 0.23
10.55 1.08
4.33 0.38
3.15 0.17
6.10 0.30
2.60 0.23
2.08 0.12
6.17 0.22
1.72 0.10
1.82 0.08
3.53 0.15
0.75 0.03
0.80 0.02
1.67 0.03
2
1.5
2
1.5
2
1.5
2
1.5
10
1.5
2.0 x 80
1.5
1
2
1.5
2
1.5
2
1.5
2
1.5
2
10
10
143
02-08-12 11:37
144
CHAPTER 8
state delivery was resumed for a pre-programmed flow-rate of 2 ml/h. For the raising and
lowering of the pump, the delivery rate was increased and decreased in inverse proportion
to the pre-programmed flow-rate.
The results of the studies mentioned above indicate that syringe pump design needs
attention to minimise syringe plunger movement and increase patient safety. Furthermore,
vertical displacement of the pump relative to the patient should be minimised, especially
when potent drugs are infused. Kern et al.17 advised that, when vertical displacement is
necessary, the flow-rate should be at least 5 ml/h. In addition, because the studies showed
performance differences between syringe pumps of the same model, syringe pumps should
be tested for vertical displacement effects before they are used in clinical practice.
The actual clinical effect of vertical displacement was shown by Igarashi et al.18 in a rabbit
model. The researchers assessed the influence of a 50-cm downward displacement of the
syringe pump on the internal pressure (IP) and volume delivered by the IV-administration
set. Furthermore, they examined the changes in systolic blood pressure (SBP) after vertical
displacement during norepinephrine infusion in rabbits with and without haemorrhagic
shock. The downward movement of the syringe pump with a pre-programmed flow-rate of
1 ml/h resulted in an internal pressure increase of 37.4 2.6 mmHg (5.0 0.3 kPa). The SBP
decreased significantly by 17 6.9%. The elevation of the syringe pump caused a significant
decrease in the IP of 37 2.8 mmHg (4.9 0.4 kPa) and an SBP increase of 45.7 21.5% from
the starting point values.
In addition to the studies mentioned above, four others have performed tests with vertical
displacements of syringe pumps.19-22 In these in vitro studies, vertical displacement is used
to evaluate the effect of the size and design of various parts of the IV-administration set.
These results will be discussed later.
8.3.2 Factor 2: Syringes
The effect of both the size and design of syringes on the accuracy of the delivered
volume has been discussed in various publications (Tables 8.2, 8.3, and 8.4). These studies
were performed in vitro with different brands of syringe pumps, different brands and size of
syringes, and different pre-programmed flow-rates.
Kim and Steward23 compared three syringe sizes in an in vitro set-up. The outcome
parameters were the effect of syringe size on the time to an occlusion alarm and the bolus
delivered after the relief of the occlusion. The authors also measured the volume of water
discharged from the syringe over precisely 1 hour after starting the infusion (n=10). They
found that the volume of water delivered over a one-hour period was not significantly dif-
02-08-12 11:37
145
ferent for the three different syringes sizes. Dnmez et al.24 tested the time to an occlusion
alarm for two types of syringe pumps with two sizes of syringes from two manufacturers.
The syringes were occluded with a stopcock. The most recent study was performed by Neal
and Lin.25 They evaluated the time to occlusion alarm and the time from the start of the
infusion to steady state flow for three syringe sizes at two pre-programmed flow-rates. To
determine the time to occlusion alarm, the IV tubing was folded and clamped. The results
of these three studies are shown in Table 8.2.
Two years before the Neal and Lin study25, Neff et al.3 also measured the time required
to reach steady state. They used four different syringe sizes from two manufacturers for
different pre-programmed flow-rates. The actual delivered volumes were obtained gravimetrically. Capes et al.11 studied the effect of the syringes design by measuring the force
needed to initiate and maintain syringe plunger motion and the time to the start of the
flow. Two brands of syringe drivers and four brands of syringes from various countries were
included, with pre-programmed flow-rates between 0.1 and 20 ml/h. The measurement of
the syringe plunger force revealed regular fluctuations indicative of plunger sticking. The
researchers concluded that syringe plunger sticking resulting in intermittent boluses can
occur at low flow-rates and with certain syringe brands. In Table 8.3 the results of these
three studies are shown. For Neff et al.3 and Capes et al.,11 the results are shown for 0.1 and 1
ml/h pre-programmed flow-rates only.
Weiss et al.22 studied the effect of four different syringe designs on volume delivery after
downwards and upwards displacement. The no-flow time, aspiration and bolus volume
were measured with an electronic balance. To determine syringe compliance, the bolus
volume released after occlusion at 100 mmHg (13.3 kPa) was measured. In another study by
Weiss et al.19 the effect of syringe size was evaluated by comparing the volume delivery for
different pre-programmed flow-rates after the vertical downward and upward displacement
of the syringe pump. The delivered volume was measured every second using an electronic
balance. The results of these two studies are shown in Table 8.4.
The results of the studies examining syringe size and design show that the time to occlusion alarm and the time without flow, as well as the aspiration and bolus volume caused by
vertical displacement, increase with syringe size. The time to occlusion alarm is also longer
for lower pre-programmed infusion rates. Syringe compliance can be minimised by a small
diameter, a rigid syringe wall, and a close fit of the plastic plunger to the silicon plunger
head. Based on these results, it was concluded that low occlusion alarm pressures should be
selected and that syringe pumps should be improved for easy adaptation of the occlusion
alarm level.
02-08-12 11:37
146
CHAPTER 8
Occlussion pressure
[mmHg]
Pre-programmed
Syringe size [ml]
Syringe brand
Number of tests
Needle/catheter
brand & size [mm]
Infusion tubing
length [m]
Syringe pump
brand & type
Pressure working
against infusion
[mmHg]
Year
Authors
[reference number]
Table 8.2 Characteristics of the studies examining the effects of syringe size and design.
Part 1: Studies of the time to occlusion for various syringe sizes.
Not
specified
590-890
0.5
1
2
5
370-660
0.5
1
2
5
590-890
0.5
1
2
5
370-660
0.5
1
2
5
590-890
0.5
1
2
5
590-890
0.5
1
2
5
10
Kim and
Steward [23]
1999
Not
specified
Medfusion 2001
Not
specified
0.87
Not
specified
20
60
JMS SP-100
JMS
JMS SP-500
Not
specified
20
20
JMS SP-100
Hayat
JMS SP-1500
Dnmez
et al. [24]
2005
JMS SP-100
Not
specified
JMS SP-100
JMS
JMS SP-500
Neal and Lin

[25]
2009
Not
specified
Baxter
AS50
50
20
Hayat
Medex
extension
tubing
1.52
BD Insyte
Autogard,
18-ga, 1.16 inch,
intravenous
catheter,
no needle
4
3
4
3
6
3
370-660
3
Covidien
Kendall
Monoject
02-08-12 11:37
12
60
Not
specified
0.5
1
2
5
0.2
5
0.2
5
0.2
5
7.4
0.09
20
Not
specified
~ 29
Not
specified
~ 0.35
84.4
1.03
Not
specified
60
590-890
0.5
1
2
5
94.1 28.2
33.4 21.2
13.2 8.2
8.1 4.3
370-660
0.5
1
2
5
39.3 17.0
17.6 8.0
8.6 3.0
3.7 1.8
590-890
0.5
1
2
5
86.4 32.2
49.8 39.4
20.4 14.4
11.6 8.4
370-660
0.5
1
2
5
48.2 25.3
14.3 4.4
7.2 2.2
4.5 7.2
590-890
0.5
1
2
5
112.7 16.5
95.6 32.2
42.4 17.6
17.5 6.9
590-890
0.5
1
2
5
117.3 9.4
85.5 33.1
38.6 17.1
15.0 7.1
0.5
1
2
5
83.0 24.8
34.4 15.2
15.9 8.5
6.4 2.2
0.2
5
0.2
5
0.2
5
> 30
1.08 0.42
> 30
3.35 0.58
> 30
27.06 3.26
JMS
20
20
Hayat
JMS
50
20
Hayat
4
3
4
3
6
3
370-660
3
Covidien
Kendall
Monoject
12
Not
specified
60
147
Bolus volume after

relief of occlusion
[L]
Occlussion pressure
[mmHg]
10
Syringe size [ml]
Time to
occlusion alarm
(mean SD) [
minutes]
Not
specified
Pre-programmed
flow rate [ml/h]
Syringe brand
Number of tests
brand & size [mm]
D Insyte
togard,
ga, 1.16 inch,
ravenous
theter,
needle
02-08-12 11:37
148
CHAPTER 8
Medex
extension
tubing
4
3
4
3
6
Covidien
Kendall
Monoject
12
60
Neff et al. [3]
Clinico PEinfusion line
B.Braun,
single lumen
22-G CVC
30
50
Becton
Dickinson
Omnifix
30
Terumo
(Japan)
Terumo
(US)
Becton
Dickinson
50
IVAC
Omnifix
Terumo
(Australia)
Terumo
(Japan)
0.1
1
1
20
Not
specified
Syringe pump
brand & type
0.1
0.1
Terumo
(Japan)
Not
specified
Becton
Dickinson
Not
specified
Alaris
AsenaTM
GH
0.1
0.1
Codan
1995
0.1
Codan
Capes et al.
[11]
Baxter
AS50
0.2
BD Plastipack
Codan
BD Plastipack
0.2
0.1
20
10
1
10
BD Plastipack
2007
0.2
0.1
BD Plastipack
Codan
Pre-programmed
flow rate [ml/h]
Syringe size [ml]
Syringe brand
152
BD Insyte
Autogard,
18-ga,
1.16 inch,
intravenous
catheter, no
Number of tests
Infusion tubing
type
Pressure working
against infusion
[mmHg]
Not
specified
Needle/catheter
brand & size
[mm]
2009
Infusion tubing
length [m]
Neal and Lin

[25]
Year
Authors
[reference
number]
Part 2: studies of the time to reach (steady-state) flow for various syringe sizes.
02-08-12 11:37
Covidien
Kendall
Monoject
12
60
BD Plastipack
Codan
10
BD Plastipack
20
Codan
BD Plastipack
Codan
0.2
5
50
Time to
steady state
(mean SD) [s]
Force required to
maintain syringe
plunger motion
at 0.053 mm/min
(mean, max) [N]
Time to start
flow [minutes]
Baxter
AS50
51 38
-
0.2
79 40
0.1
20.5 10.8
3.6 0.9
0.1
12.3 3.8
3.9 1.7
0.1
25.2 15.6
6.2 2.9
0.1
22.6 7.9
Alaris
AsenaTM
GH
4.1 0.8
27.3 10.1
4.8 1.5
0.1
15.4 8.3
3.9 0.5
0.1
74.5 26.6
13.9 7.4
0.1
20.4 10.4
5.3 2.1
0.1
1
0.1
1
0.1
1
0.1
1
1
0.1
1
0.1
1
0.1
1
0.1
1
0.1
1
0.1
1
0.1
1
149
24 10
0.1
30
Syringe pump
brand & type
0.2
Codan
BD Plastipack
Pre-programmed
flow rate [ml/h]
Syringe size [ml]
Syringe brand
Atom 235
ivac 770
Atom 235
ivac 770
0.2
0
5
0.1
7.7, 9.4
Atom 235
ivac 770
Atom 235
ivac 770
28.4
0
0.2
0.1
13.0, 15.4
Atom 235
ivac 770
Atom 235
ivac 770
Atom 235
ivac 770
Atom 235
ivac 770
99.1
0
2.7
0.2
15.5, 20.4
1.7
10.8
0
0.3
1.4, 1.9
Atom 235
ivac 770
Atom 235
ivac 770
Atom 235
ivac 770
Atom 235
ivac 770
115
357.8
0
1,6
19.8, 30.4
139.2
21.8
0.8
2.8
Atom 235
ivac 770
Atom 235
ivac 770
63
0
0.2
0.1
13.5, 14.6
Atom 235
ivac 770
Atom 235
ivac 770
81
16.6
0.5
1.2
11.5, 13.5
Atom 235
ivac 770
Atom 235
ivac 770
4.4
0
1.5
0.1
5.6, 6.7
Atom 235
ivac 770
Atom 235
ivac 770
80.4
551.9
12.1
22.4, 28.4
Atom 235
ivac 770
Atom 235
ivac 770
91.5
45.2
19.7
39.5
29.0, 39.6
02-08-12 11:37
150
CHAPTER 8
Syringe size [ml]
Pre-programmed
Syringe brand
Vertical displacement
[cm]
Number of tests
Occlussion pressure
[mmHg]
Needle/catheter
brand & size [mm]
Infusion tubing
length [m]
Syringe pump
brand & type
Pressure working
against infusion
[mmHg]
Year
Authors
[reference number]
Part 3: studies of the vertical displacement of the syringe pumps for various syringe sizes.
50
Codan
Weiss et al. [22]
2000
Not
specified
IVAC P4000
Fresenius
InjectomatLine
Ivac
2
100
50
BD Plastipak
Fresenius
Injectomat
50
Weiss et al. [19]
2000
10
IVAC Alaris
50
Arrow,
CV-Catheterisation Set,
14 Ga./ 30 cm
300
130
02-08-12 11:37
14.7 1.4
3.3 0.4
43.6 2.2
19.6 2.2
4.8 0.2
44.3 3.1
20.7 0.6
4.9 0.4
57.9 2.2
26.5 2.0
6.4 0.6
41
31
0.4 0.05
40 3
21 3
5.1 0.16
0.5
10 3
83
2.1 0.4
20 3
10 3
1.1 0.08
20 3
10 3
0.61 0.05
0.5
42 3
28 3
8.7 1.1
42 3
26 5
3.8 0.9
42 3
22 3
1.6 0.05
0.5
56 5
40 3
21.1 0.6
50 5
40 3
5.6 0.16
158 6.4
30 3
2.2 0.2
0.5
166 4.8
78 6.4
29.7 2.1
160 4.8
68 3.2
12.8 0.2
160 4.8
6.2 3.2
5.5 0.15
BD Plastipak
50
Pre-programmed
flow rate [ml/h]
38.9 4.0
Ivac
Syringe size [ml]
Codan
Syringe brand
Time without flow

after vertical
displacement
(mean SD)
[minutes]
50
Aspiration volume
after vertical displacement
(mean SD) [L]
Vertical displacement
[cm]
Number of tests
Occlussion pressure
[mmHg]
100
Bolus volume after

vertical displacement
(mean SD) [L]
Set,
cm
Fresenius
Injectomat
BD Plastipak
10
50
10
300
BD Plastipak
20
30
50
151
02-08-12 11:37
152
CHAPTER 8
8.3.3 Factor 3: Infusion tubing

Although the volume of infusion tubing is relatively low (2 to 3 ml) compared to the syringe
volume, two studies by Weiss et al.21, 26 demonstrated that the infusion tubings compliance
actually affects the flow-rate variability. In one of the two studies,26 the impact of the three
different infusion tubings compliance on the time to an occlusion alarm and the size of the
subsequent occlusion release bolus were determined. Tests were performed with 10-ml and
50-ml syringes and a pre-programmed flow-rate of 1 ml/h. A measured occlusion pressure
of 230 mmHg was defined as the occlusion alarm level. The syringe compliance and infusion tubing compliance were determined by dividing the bolus volume after the pressure
release by the occlusion pressure. The mean occlusion time differed significantly between
the three types of infusion tubing. In combination with the 50-ml syringe, the times to
occlusion alarm were 20.5 0.46, 24.8 0.36 and 26.2 0.19 minutes for each of the three
tested infusion tubings. For the 10-ml syringe, the times to occlusion alarm were 2.6 0.08,
5.6 0.28 and 6.9 0.35 minutes. The calculated tubing compliances for each of the three
types of tubing were 0.038 0.025, 0.22 0.07 and 0.32 0.08 l/mmHg.
The other study evaluated the effect of infusion tubing compliance during vertical displacements of syringe pumps.21 Five different types of infusion tubing were tested at preprogrammed flow-rates of 0.5, 1.0 and 1.5 ml/h. The catheter tip was submerged in water
to create a pressure of 10 mmHg. After steady-state flow conditions were reached, a 70cm vertical infusion tubing loop below the syringe pump level was performed. The compliance of each infusion tubing type was calculated using a pressure transducer and the
bolus volume measurement after an occlusion pressure of 300 mmHg. Creating the loop in
the tubing at pre-programmed flow-rates of 0.5 ml/h resulted in a no-flow time between 5.1
1.5 seconds for low-compliance (lc) tubing and 44.0 6.8 seconds for high-compliance (hc)
tubing. For higher pre-programmed flow-rates, the no-flow time was significantly lower.
The researchers also studied the effect of infusion tubing compliance on drug delivery after
a 50-cm downward displacement. The no-flow times were 295.8 20.7 seconds (lc) and 463.3
24 seconds (hc). Returning the syringe pump to the original position delivered boluses of
34.2 4.9 l (lc) and 53.0 3.1 l (hc).
From these two studies, Weiss et al. concluded that infusion tubing compliance has an important impact on the time to occlusion alarm and on the flow-rate variability associated
with the vertical displacement of syringe pumps.
8.3.4 Factor 4: Check valves & anti-siphon valves
In IV-administration sets, special valves can be placed in the infusion tubing between the
syringe and the patient to prevent backflow and siphonage, two frequently occurring phenomena in IV therapy. Backflow can occur when multiple infusions are interconnected via
add-on devices.4 Due to the differences in resistance, the administered substances do not
flow into the patient, but into another line.27 In siphonage, there is uncontrolled emptying
or free flow of substances from a syringe into the patient or vice versa. Siphonage can occur
when the syringe is poorly clamped into the syringe pump, not clamped at all, or when there
are air leaks in the IV-administration set.15, 28-30
02-08-12 11:37
153
The valves available for the prevention of siphonage have a relatively high opening pressure. In combination with a low pre-programmed flow, this high resistance is responsible
for flow-rate variability. The effect of the resistance was shown by McCarroll et al.29 They
compared the start-up time, i.e., the time from the start of the infusion to first delivery of
substances from the catheter, for three commercially available anti-siphon valves with preprogrammed flow-rates of 2, 10, and 50 ml/h. Each test was performed 15 times. The results
showed longer start-up times with valves compared with the control set-up without valves.
This effect was the largest for the pre-programmed flow-rate of 2 ml/h. While the mean
SD start-up time for a flow of 2 ml/h without a valve was 3.5 2.09 minutes, the start-up
times with each of the three tested valves were significantly higher: 12.3 7.48, 12.3 5.06,
and 18.4 9.26 minutes.
Weiss et al.20 studied whether pressure regulation with anti-siphon valves reduces the flow
variation during the vertical displacement (50 cm) of syringe pumps. The set-up consisted
of a 50-ml syringe and a 2-meter lc infusion line. The set-up was tested with two different
brands of syringes without anti-siphon valves and with two valves of differing operating
pressures: 75 mmHg (10 kPa) and 155 mmHg (20.6 kPa). Without an anti-siphon valve, the
no-flow times after lowering the syringe pump were (mean SD) 2.4 0.2 minutes and 4.09
0.55 minutes for the two different syringes. The no-flow time increased significantly when
the valves were introduced and depended on the type of syringe used. For the valve with an
opening pressure of 75 mmHg, the no-flow time increased by 58% and 43% for each of the
two syringes, respectively. For the valve with an opening pressure of 155 mmHg, the no-flow
times increased by 88% and 81% for each of the syringes, respectively. The researchers concluded that anti-siphon valves have a negative effect on flow-rate variability.
8.3.5 Factor 5: Inline filters
Due to the wear of syringes used for IV therapy, plastic particles from the syringes can
enter the patients circulation. These infused particles have the potential to cause e.g.,
endothelial damage or act as a nidus for thrombosis.31, 32 Cant et al.33 described a fatal
bowel necrosis in a neonate due to irregular 50- to 200-m fragments of polypropylene.
The authors concluded that manufacturers must make greater efforts to minimise particulate contamination and that the use of filters in intravenous lines should be considered.
In 2006 a systematic review by Foster et al.34 assessed whether in-line filters on intravenous
lines prevent morbidity and mortality in neonates. They conclude that there is insufficient
evidence to recommend the use of in-line filters to prevent mortality in neonates.
One study evaluated the effect of inline filtration on IV fluid delivery. In this study, gentamicin was infused at a pre-programmed flow-rate of 8.26 ml/hr.35 The authors tested three
different filters (Travenol 0.22 micron, Travenol 0.5 micron, and Pall Ultipor 0.2 micron) in
four positions (horizontal and vertical position with ascending or descending fluid flow).
With the Travenol 0.22-micron filter, drug delivery was significantly faster with an ascending fluid flow than with a descending flow in both positions. No studies examining the
effect of the presence or absence of in-line filters on the flow variability or on start-up times
were found.
02-08-12 11:37
154
CHAPTER 8
8.3.6 Factor 6: Add-on devices

Add-on devices are used to connect multiple infusion lines to a single vascular access device. One study related to add-on devices and flow-rate variability was found. This study, by
Levi et al.,27 evaluated the efficiency of a stopcock system in an in vitro experimental set-up.
To test the efficiency two syringe pumps with pre-programmed flow-rates of 0.1 and 8 ml/h
were connected to an array of stopcocks. A diluted dye was used to quantify the volume
reaching the patient. Over a 90-minute period, samples were collected and analysed using
spectrophotometry. After the pre-programmed flow-rate was changed from 0.1 ml/h to 0.2
ml/h, the time required to double the delivery of the diluted red dye was 32 minutes. The
researchers concluded that significant delays were associated with the stopcock array and
that healthcare professionals in a neonatal care setting should minimise dead space in IVadministration set.
8.3.7 Factor 7: Vascular access devices
There is considerable variety in the vascular access devices that can be used in IV therapy
for newborn infants.36, 37 The usefulness of specific vascular access devices depends on the
intended length of therapy and the type of drug used (e.g., the osmolarity and pH of the
substance and the use of vesicants or irritants), among other factors.
Several studies were found that evaluated the influence of the vascular access device design on the flow variability. Angle et al.38 studied 16 different peripherally inserted central
catheters (PICCs) from 6 manufacturers with pre-programmed flow-rates between 25 and
270 ml/h. This study showed that the PICCs material influenced the flow variability. For
instance, polyurethane PICCs showed much better flow-rates than silicone PICCs with a
comparable outside diameter. The reason for this difference is the relatively thin wall of the
polyurethane catheters. The researchers concluded that there is significant variability in the
flow capabilities of available PICCs.
In 2007 Lovich et al.39 evaluated the delivery of IV substances for central venous catheters
in an in vitro multi-infusion experimental set-up. They showed that there are substantial
differences between central venous catheters in their delivery of IV substances. According
to the authors the delivery of the IV substances depends on the dead volume and the rate
of carrier flow.
Bartels et al.40 determined how the dead space of central venous catheters affects the
volumes delivered to the patient for (low) pre-programmed flow-rates. In this study the
researchers mimicked new infusions and the resumption of stopped infusions. Two syringe
pumps were connected to a Y-piece and compared at pre-programmed flow-rates of 2 and 12
ml/h. The Y-piece was either primed or not. Each condition was tested at least 3 times. The
delivered volume was measured once every minute. The times to achieve half of the steadystate delivery for the low and high flow-rates were 23.5 2.1 minutes and 15.7 2.9 minutes,
respectively. For the low flow, the expected steady-state delivery rate was not achieved for
more than 40 minutes. Priming the Y-piece shortened the time required to reach half of the
steady-state delivery: 12.7 0.6 minutes for low flow and 5.2 0.8 minutes for high flow.
02-08-12 11:37
155
8.4 Discussion
This literature review shows that several factors increase the flow-rate variability of IV
therapy for newborns, which have clinically relevant consequences. The factors affecting
the volumes delivered to the patient include the vertical displacement of the syringe pump
relative to the patient, the internal compliance of the complete IV-administration set, and
the absence or presence of valves and filters in the IV-administration set.
The factors actual effect on the delivered volume varied widely amongst the included studies. This variation finds its origin in the differences in the studies test methods, materials
and experimental set-ups and reflects the problems that healthcare professionals are confronted with in clinical practice. Each change in the IV-administration set will affect the
actual volume delivered to the patient,4 but the actual change in delivered volume cannot
easily be predicted.
Despite the differences in the included studies results, it can be concluded that the start-up
time, the time to steady state flow, and the no-flow time after lowering the syringe pump
decrease with decreasing internal compliance of the IV-administration set and with increasing pre-programmed flow-rate. Not surprisingly, the same applies to the time to an
occlusion alarm. It has been shown that the time to an occlusion alarm is lengthened at
low rates with larger syringes.23-25 Other studies have shown that the total volume of boluses
and aspirations is not related to the pre-programmed flow-rate, but depends solely on the
degree of vertical displacement of the syringe pump relative to the patient and the compliance of the IV-administration set.
The compliance of the IV-administration set depends on the total volume, design and
materials of the parts of the IV-administration set. Although small syringes are preferable
in terms of reducing compliance, it is necessary to find the optimum balance between the
smallest syringe size and the least frequent need to replace empty syringes. This optimal
balance is necessary because despite all precautions syringe changes are still associated with
increased risks of blood stream infections.
Concerning the other parts of the IV-administration set, it can be concluded that valves
in the IV-administration set are used to prevent backflow and/or siphonage; however, the
relatively large resistance of some valves enhances flow-rate variability. Therefore, healthcare professionals should be aware of the differences between valves and should choose the
type that meets their needs. One study showed that the position of inline filters affects the
flow-rate. Add-on devices can significantly delay the delivery of IV substances. Therefore,
low-volume add-on devices are preferable to those with larger volumes. The effect of the
material, length and position of vascular access devices on flow-rate variability has not been
studied extensively. Although these variables may influence the delivery of substances, the
actual effect is unclear.
02-08-12 11:37
156
CHAPTER 8
The clinical relevance of the changes in delivered volume depends on patient characteristics
and the substances being supplied to the preterm neonate. For example, when vasopressors
or inotropic agents are delivered to the patient in a non-continuous manner, the subsequent haemodynamic instability may lead to cerebral haemorrhage in preterm neonates.41
To conclude, flow-rate variability at low infusion rates is a persistent problem in IV therapy
for newborn infants. This flow-rate variability and its clinical relevance are due primarily
to the pre-programmed flow-rate, changes in hydrostatic pressure, the compliance of the
complete IV-administration set, and the type of substances being supplied to the patient.
To prevent serious negative effects on (preterm) neonates, healthcare professionals must
be educated about the dynamics of the IV-administration set and be made aware that they
should take steps to minimise flow-rate variability. These steps include minimising the
internal compliance of the complete IV-administration set, using the highest possible preprogrammed flow-rate, using valves with a low resistance and selecting low occlusion alarm
pressure. Manufacturers can help physicians by developing products that are optimised for
neonatal IV therapy.
02-08-12 11:37
157
8.5 References
1. Loisel D, Smith M, MacDonald M, Martin G. Intravenous access in newborn infants: impact of extended
umbilical venous catheter use on requirement for
peripheral venous lines. J Perinatol. 1996;16(6):461-466.
2. Hermansen M, Goetz Hermansen M. Intravascular
catheter complications in the neonatal intensive care
unit. Clinics in Perinatology. 2005;32(1):141-156.
3. Neff S, Neff T, Gerber S, Weiss M. Flow rate,
syringe size and architecture are critical to start-up
performance of syringe pumps. European Journal of
Anaesthesiology. 2007;24(7):602-608.
4. Timmerman A, Riphagen B, Klaessens J, Verdaasdonk
RM. Confirmation of uncontrolled flow dynamics
in clinical simulated multi-infusion setups using
absorption spectral photometry. In: Raghavachari
R, Liang R, editors. Design and Quality for Biomedical Technologies. Bellingham: Spie-Int Soc Optical
Engineering; 2010.
5. Lovich M, Doles J, Peterfreund R. The impact of
carrier flow rate and infusion set dead-volume on the
dynamics of intravenous drug delivery. Anesthesia &
Analgesia. 2005;100(4):1048-1055.
6. Lovich M, Kinnealley M, Sims N, Peterfreund R. The
delivery of drugs to patients by continuous intravenous
infusion: Modeling predicts potential dose fluctuations
depending on flow rates and infusion system dead
volume. Anesthesia & Analgesia. 2006;102(4):1147-1153.
7. Bell E, Acarregui M. Restricted versus liberal water
intake for preventing morbidity and mortality in
preterm infants. Cochrane Database of Systematic
Reviews 2001(3):CD000503.
8. Donald A, Chinthamuneedi M, Spearritt D. Effect
of changes in syringe driver height on flow: a small
quantitative study. Critical Care and Resuscitation.
2007;9(2):143-147.
9. Schulze K, Graff M, Schimmel M, Schenkman A, Rohan
P. Physiologic oscillations produced by an infusion
pump. Journal of Pediatrics. 1983;103(5):796-798.
10. Hurlbut J, Thompson S, Reed M, Blumer J, Erenberg A,
Leff R. Influence of infusion pumps on the pharmacologic response to nitroprusside. Critical Care
Medicine. 1991;19(1):98-101.
11. Capes D, Dunster K, Sunderland V, McMillan D,
Colditz P, McDonald C. Fluctuations in syringe-pump
infusions: association with blood pressure variations in
infants. American Journal of Health-System Pharmacy.
1995;52(15):1646-1653.
12. Stowe CD, Storgion SA, Lee KR, Phelps SJ. Hemodynamic response to intentionally altered flow continuity
of dobutamine and dopamine by an infusion pump in
infants. Pharmacotherapy. 1996;16(6):1018-1023.
13. Cunningham S, Deere S, McIntosh N. Cyclical variation
of blood pressure and heart rate in neonates. Arch Dis
Child. 1993;69(1 Spec No):64-67.
14. Sherwin CMT, McCaffrey F, Broadbent RS, Reith DM,
Medlicott NJ. Discrepancies between predicted and
observed rates of intravenous gentamicin delivery for
neonates. Journal of Pharmacy and Pharmacology.
2009;61(4):465-471.
15. Lnnqvist P, Lfqvist B. Design flaw can convert
commercially available continuous syringe pumps to
intermittent bolus injectors. Intensive Care Medicine.
1997;23(9):998-1001.
16. Neff T, Fischer J, Schulz G, Baenziger O, Weiss M.
Infusion pump performance with vertical displacement: effect of syringe pump and assembly type.
Intensive Care Medicine. 2001;27(1):287-291.
Spies C, et al. Downward movement of syringe pumps
reduces syringe output. British Journal of Anaesthesia.
2001;86(6):828-831.
18. Igarashi H, Obata Y, Nakajima Y, Katoh T, Morita K,
Sato S. Syringe pump displacement alters line internal
pressure and flow. Canadian Journal of Anesthesia.
2005;52(7):685-691.
19. Weiss M, Hug M, Neff T, Fischer J. Syringe size and
flow rate affect drug delivery from syringe pumps.
21. Weiss M, Bnziger O, Neff T, Fanconi S. Influence of
infusion line compliance on drug delivery rate during
acute line loop formation. Intensive Care Medicine.
2000;26(6):776-779.
22. Weiss M, Fischer J, Neff T, Baenziger O. The effects
of syringe plunger design on drug delivery during
vertical displacement of syringe pumps. Anaesthesia.
2000;55(11):1094-1098.
23. Kim D, Steward D. The effect of syringe size on
the performance of an infusion pump. Pediatric
Anesthesia. 1999;9(4):335-337.
02-08-12 11:37
158
CHAPTER 8
24. Dnmez A, Araz C, Kayhan Z. Syringe pumps take too

long to give occlusion alarm. Pediatric Anesthesia.
2005;15(4):293-296.
Care Medicine. 2009;10(5):592-596.
infusions in the newborn: Problems and possible
solutions. Pediatric Critical Care Medicine.
2010;11(2):275-281.
28. Southern D, Read M. Lesson of the Week: Overdosage
of opiate from patient controlled analgesia devices.
British Medical Journal. 1994;309(6960):1002.
29. McCarroll C, McAtamney D, Taylor R. Alteration
in flow delivery with antisyphon devices. Anaesthesia.
2000;55(4):355-357.
30. Rooke G, Bowdle A. Syringe Pumps for Infusion of
Vasoactive Drugs. Anesthesia & Analgesia.
1994;78(1):150-156.
31. Puntis J, Wilkins K, Ball P, Rushton D, Booth I. Hazards
of parenteral treatment: do particles count? Arch Dis
Child. 1992;67(12):1475-1477.
32. Lehr HA, Brunner J, Rangoonwala R, Kirkpatrick CJ.
Particulate matter contamination of intravenous antibiotics aggravates loss of functional capillary density in
postischemic striated muscle. American Journal of Respiratory and Critical Care Medicine. 2002;165(4):514520.
33. Cant A, Lenney W, Kirkham N. Plastic material from
a syringe causing fatal bowel necrosis in a neonate. British Medical Journal (Clinical research ed).
1988;296(6627):968-969.
34. Foster J, Richards R, Showell M. Intravenous in-line
filters for preventing morbidity and mortality in
neonates. Cochrane database of systematic reviews.
2006(2):CD005248.
35. Nazeravich DR, Otten NH. Effect of inline filtration
on delivery of gentamicin at a slow infusion rate.
Am J Hosp Pharm. 1983;40(11):1961-1964.
36. Stovroff M, Teague WG. Intravenous access in infants
and children. Pediatric Clinics of North America.
1998;45(6):1373-1393.
37. Ramasethu J. Complications of vascular catheters

in the neonatal intensive care unit. Clinics in
perinatology. 2008;35(1):199-222.
38. Angle JF, Matsumoto AH, Skalak TC, OBrien RF,
Hartwell GD, Tegtmeyer CJ. Flow characteristics of
peripherally inserted central catheters. Journal of vascular and interventional radiology. 1997;8(4):569-577.
39. Lovich MA, Peterfreund GL, Sims NM, Peterfreund RA.
Central venous catheter infusions: A laboratory model
shows large differences in drug delivery dynamics
related to catheter dead volume. Critical Care
Medicine. 2007;35(12):2792-2798.
40. Bartels K, Moss D, Peterfreund R. An analysis
of drug delivery dynamics via a pediatric central
venous infusion system: quantification of delays in
achieving intended doses. Anesthesia & Analgesia.
2009;109(4):1156-1161.
41. Funato M, Tamai H, Noma K, Kurita T, Kajimoto Y,
Yoshioka Y, et al. Clinical events in association with
timing of intraventricular hemorrhage in preterm
infants. Journal of Pediatrics. 1992;121(4):614-619.
02-08-12 11:37
159
02-08-12 11:37
160
02-08-12 11:37
CHAPTER 9
Flow-rate variability in
neonatal IV therapy
caused by the use of
check valves
A.C. van der Eijk, A.J. van der Plas, C.J.N.M. van der Palen, J.Dankelman, B.J. Smit
02-08-12 11:37
162
CHAPTER 9
Intravenous (IV) therapy is provided to virtually all patients in a neonatal intensive care
unit, but unfortunately, the accuracy and predictability of the substances delivered to
the patient is limited. One of the factors that can influence the flow delivery is the
presence or absence of valves in the IV-administration set. The use of valves in the IVadministration set is twofold. Although valves are needed to prevent siphonage, and
flow in opposite directions, the relative large resistance of current available valves
increases the flow rate variability. This chapter is based on the article titled: Flow-rate
variability in neonatal IV therapy caused by the use of check valves (submitted). In this
chapter three different types of check valves are compared in an in vitro set-up.
OBJECTIVE To evaluate the effect of three different types of check valves on the flow
characteristics in a low-flow multi-infusion set.
02-08-12 11:37
Check-valves in neonatal IV therapy
163
9.1 INTRODUCTION
Intravenous (IV) therapy is provided to virtually all neonates admitted to a neonatal intensive care unit (NICU).1-3 Because of the limited vascular access possibilities, the illness, and/
or the low body weight of (preterm) neonates, multi-infusion is used during IV therapy in
NICUs. In multi-infusion therapy, several infusions are supplied to the infant via a single
catheter.4-9 Although it is expected that the IV substances are supplied to the patient with a
pre-programmed flow-rate, it has been shown that the actual volume delivered to the patient
can vary over time.5, 7, 10-16
The uncontrollability of the delivered volume is caused by backflow and siphonage, in addition to other factors. In siphonage, there is uncontrolled emptying or free flow of substances
from a syringe into the patient. Siphonage can occur when the syringe is not clamped or is
poorly clamped in the syringe pump or when there are air leaks in the IV-administration
set.8, 16-18 Backflow can occur when multiple infusions are interconnected to each other (e.g.,
via a stopcock). Because of differences in resistance in the IV-administration set, it is possible
that fluids do not flow from the syringe into the patient but instead into another line.19
To prevent backflow, various types of check valves can be inserted in the IV-administration
set. Some of these check valves can prevent siphonage as well. Unfortunately, it has been
shown that the presence of check valves and/or anti-siphon valves in the IV-administration
set can introduce a delay in the delivery of IV substances.16, 20
The aim of this study was to evaluate the effect of three different types of check valves on the
flow characteristics of a low-flow multi-infusion set-up for preterm infants. To visualise the
flow characteristics, an in vitro experimental IV set-up with in-line flow meters was used to
simulate clinical situations.
Although it is expected that the IV substances

are supplied with a pre-programmed flow-rate,
the actual volume delivered to the patient can
vary over time.
02-08-12 11:37
164
CHAPTER 9
9.2 Methods
The experiments were performed in the Erasmus Medical Centre - Sophia Childrens
Hospital in Rotterdam, the Netherlands. Approval from the hospital research ethics board
was not necessary because no patients were involved in the study.
9.2.1 Check valves
Table 9.1 provides the characteristics of the three different types of check valves that were
used in the tests. These valves were all designed to prevent backflow of IV fluids. Two of the
three check valves, the HOP-S and the LOP-S, have anti-siphon functionality, as well. The
LOP valve has the lowest opening pressure but does not have anti-siphon functionality.
Table 9.1 Manufacturers specifications of the tested valves. The opening pressure of the valves is
provided in kPa (1 kPa equals 7.5 mmHg).
Abbreviation
Opening
pressure [kPa]
Prevent
backflow
Prevent
siphonage
Type
Manufacturer
Location,
country
<2
Infuvalve
BBraun
Melsungen,
Germany
LOP-S
3-7
SyphonSafe
(not commercially available)
Filtertek BV
Co.
Limerick,
Ireland
HOP-S
10 - 40
BC1000
BBraun
Melsungen,
Germany
LOP

To evaluate the effect of the check valves on the flow characteristics in a low-flow multiinfusion IV set for preterm infants, an in vitro experimental set-up was designed with two
syringe pumps connected to a catheter via infusion tubing and a stopcock (Figure 9.1).
The manufacturers specifications of the materials used in the experimental set-up are
shown in Table 9.2.
In daily practice, the pre-programmed flow-rates and concentrations of the IV substances
are chosen such that the supply of nutrition and drugs is maximal and the total volume of
fluids is maintained within the prescribed limits. To simulate daily practice, Pump 1 was preprogrammed with a flow-rate of 2.5 ml/h (to simulate, for example, total parenteral nutrition) and Pump 2 was pre-programmed with a flow-rate of 0.1 ml/h (to simulate for example,
vasoactive drugs). The experiments were performed with disposable 20-ml luer-lock syringes
filled with a 15% glucose solution. To test a check valve, the valve was positioned between the
syringe and the infusion line. The open end of the catheter was positioned in a column filled
02-08-12 11:37
165
with water to create a pressure of 10 cmH2O (7.4 mmHg), to simulate the central venous
pressure in a clinical setting. In-line bidirectional flow meters were positioned at two points
of the experimental set-up. The flow meters were connected to a PC, and the flow in the infusion lines was recorded continuously, with a sample frequency of approximately 100 Hz.
Valve
Syringe pump 2,
0.1 ml/h
Valve
Height
= 30 cm
Syringe pump 1,
2.5 ml/h
Stopcock with
3-way valves
Flowmeter 1
Height
= 20 cm
single lumen
catheter, 1 Fr
10
cmH2O
Flowmeter 2
Figure 9.1 Experimental set-up for Experiment II. Both pumps were positioned on top of each other,
20 cm above the catheter tip. The flow meters and the stopcock were positioned level with the catheter
tip on a horizontal surface. The distance from each of the flow meters to the stopcock was equal. At T2.1
(30 minutes after the start at T2.0), the pumps were raised 30 cm. The experimental set-up for Experiment I was similar, except that both pumps were positioned level with the flow meters, the stopcock,
and the catheter tip.
9.2.3 Study set-up

The experimental set-up was used to perform two different experiments.
9.2.3.1 Experiment I: Adding syringes
The aim of Experiment I was to obtain insight regarding the flow-rate variability that occurs
when a syringe pump with a low pre-programmed flow-rate is added to an already running
IV-administration set. Both pumps were placed on a horizontal plane to ensure that the
height of the syringe outlet was level with the catheter tip to prevent differences in hydrostatic pressure. According to standard practice, the pump with the lowest pre-programmed
flow-rate (Pump 2) was connected to the entry of the stopcock proximal to the patient;
Pump 1 was connected to the entry of the stopcock distal to the patient. Pump 1 was started
at time T1.0. At time T1.1, 30 minutes after T1.0, Pump 2 was started, and the corresponding three-way valve of the stopcock was opened. At time T1.2, 90 minutes after T1.0, both
pumps were stopped, and the experiment was finished.
02-08-12 11:37
166
CHAPTER 9
Table 9.2 Manufacturers specifications of materials used for the experimental set-up
Equipment
Type
Manufacturer
Location, country
Perfusor fm
BBraun
Melsungen, Germany
Syringes
Plastipak, disposable,
luer-lock, Vol. 20 ml
BD
Drogheda, Ireland
Infusion line
1.0-2.0 mm, L. 150 cm,

Vol. 1.5 ml
Vygon
Ecouen, France
Vygon
Aachen, Germany
Bronkhorst
Veenendaal,
The Netherlands
Pump 1,
(pre-set flow-rate 2.5 ml/h)
Pump 2,
(pre-set flow-rate 0.1 ml/h)
Four 3-way stopcocks with

bionector
Biovystar 4
Catheter
Premicath, 1 Fr,
length 20 cm, Vol. 0.09 ml,
Flow meter for Pump 1
mini cori-flow
Flow meter for Pump 2
-flow
02-08-12 11:37
167
9.2.3.2 Experiment II: Changing height

The aim of the second experiment was to measure the flow-rate variability that occurs when
syringe pumps are vertically displaced with respect to the catheter tip. The experimental setup was similar to that in Experiment I except that the pumps were positioned on top of each
other, 20 cm above the catheter tip. According to clinical practice, the pump with the lowest
flow-rate (Pump 2) was positioned on top. Both pumps were started at time T2.0. At time
T2.1, 30 minutes after T2.0, the height of both pumps was increased to more than 30 cm. At
time T2.2, 60 minutes after T2.0, the pumps were returned to their original position. At T2.3,
90 minutes after T2.0, both pumps were stopped, and the experiment was finished.
Both experiments were performed at room temperature twelve times (three repeated tests
without the use of a check valve and three tests using each type of check valve described in
Table 9.1: a LOP, a LOP-S, and a HOP-S valve). Before each test, care was taken to extrude the
air bubbles from the syringe, to prime the complete IV-administration set, and to calibrate
the flow meters. To prevent bias caused by re-use of the syringe or check valve, new syringes
and check valves were used for every test. The actual flow in the lines, measured by the two
flow meters, was analysed after the experiment to determine the following factors:
the start-up time;
the distribution of flow over time;
the total time spent within 75 to 125% of the pre-programmed flow-rate;
the area under the curve of the flow-rate.
The start-up time was defined as the time required to reach 75% of the pre-programmed
flow-rate (i.e., 1.875 ml/h for Pump 1 and 0.075 ml/h for Pump 2) for the first time after the
pumps were started. The area under the curve of the flow-rate was calculated to determine
the actual delivery of fluids to the patient.
02-08-12 11:37
168
CHAPTER 9
9.3 Results
Measured Flow [ml/h]
Twenty-four tests were performed, providing 36 hours of flow characteristics. Figure 9.2
shows typical examples of the measured flow characteristics in the set-up without a check
valve in Experiments I and II.
Pre-programmed flow rate = 2.5 ml/h

Expected flow rate for 0.1 ml/h
51
T1.0: Pump 1
is started
T1.1: Pump 2
is added to the IV
administration set
T1.2: Experiment
is ended
Time [minutes]
Measured Flow [ml/h]
-0.

Expected flow rate for 0.1 ml/h
T2.0: Both pumps T2.1: Both pumps

are started
have a 30 cm
increase
in height
Time [minutes]
T2.2: Both pumps

have a 30 cm
decrease
in height
T2.3: Experiment
is ended
Figure 9.2A&B Two typical examples of flow characteristics during the tests. Figure 9.2A shows the
flow-rate over 90 minutes for Experiment I, without check valve. Figure 9.2B shows the flow-rate over 90
minutes for Experiment II, without check valve.
02-08-12 11:37
169
9.3.1 Experiment I: Adding syringes

9.3.1.1 Start-up time
In Experiment I, the start-up time of Flow Meter 1 was 1.73 0.04 minutes (mean SD)
when a check valve was not used. The longest measured start-up time was 3.76 minutes,
measured when the HOP-S check valve was used. Starting Pump 2 and opening one of the
three-way valves on the stopcock at time T1.1 caused a sudden change in flow-rate in both
flow meters in all of the tests. After this initial change in the flow, the start-up time for Flow
Meter 2 was 27.59 3.80 minutes when a check valve was not used. The use of a check valve
caused the start-up time to increase, and the HOP-S check valve prevented the attainment
of the 0.075 ml/h set point within the measured time (Table 9.3).
Table 9.3 Start-up times (time to reach 75% of the pre-programmed flow-rate) in Experiment I for
Flow Meter 2. The pre-programmed flow-rate was 0.1 ml/h.
Flow Meter 2 (0.1 ml/h)
Start-up time (at T1.1)
Flow-rate distribution
(time period: T1.1 - T1.2)
mean SD [minutes]
(number of tests with a start-up time
<60 minutes)
median (25th - 75th percentile)

[ml/h]
No check valve
27.6 3.8 (n = 3 out of 3)
0.056 (0.025 - 0.081)
LOP
36.1 6.6 (n = 3 out of 3)
0.044 (-0.006 - 0.073)
LOP-S
43.7 2.7 (n = 3 out of 3)
0.020 (-0.002 - 0.056)
HOP-S
- (n = 0 out of 3)
0.018 (-0.001 - 0.035)
Experiment I
9.3.1.2 Flow-rate variability

The total time for which the flow measured in Flow Meter 1 was within 75 to 125% of the
pre-programmed flow-rate of Pump 1 (i.e., between 1.875 and 3.125 ml/h) ranged from 96
to 98% of the total duration of the test in all of the tests. For Flow Meter 2, the total time
spent within 75 to 125% of the pre-programmed flow-rate of Pump 2 (i.e., between 0.075
and 0.125 ml/h) was 15%, 22%, 10% and 0% for tests performed without a check valve, with a
LOP valve, with a LOP-S valve and with a HOP-S valve, respectively. The median (25th 75th
percentile) of the flow-rate measured by Flow Meter 2 in Experiment I is shown in Table 9.3
for each check valve. The corresponding frequency histogram of the flow is shown in Figure
9.3. The minimum and maximum flow-rates measured by Flow Meter 2 were -3.5 and 2.3
ml/h, respectively.
02-08-12 11:37
170
CHAPTER 9
The minimum and maximum measured total flow-rates through the catheter,
(i.e., the sum of the measured flow in both flow meters) were -5.2 and 5.2 ml/h,
respectively. The corresponding frequency histogram is illustrated in Figure 9.4.
Occurrence [%]
LOP
LOP-S
HOP-S
No valve
Flow rate [ml/h]
Figure 9.3 Frequency histogram of the measured flow from Pump 2 in Experiment I. Only flow-rates
ranging from -0.3 to 0.3 ml/h are shown. The frequency histogram includes the flow from the moment Pump 2 was started until the end of the experiment (time period: 60 minutes, from T1.1 to T1.2).
The pre-programmed flow-rate was 0.1 ml/h.
12
LOP
LOP-S
HOP-S
No valve
Occurrence [%]
10
8
6
4
2
0
2.2
2.3
2.4
2.5
2.6
2.7
2.8
Flow rate [ml/h]
Figure 9.4 Frequency histogram of the measured flow through the catheter (sum of the flow
measured in both Flow meters 1 and 2) in Experiment I. Only flow-rates ranging from 2.2 to 2.8 ml/h are
shown. The frequency histogram includes the flow from the moment Pump 2 was started until the end
of the experiment (time period: 60 minutes, from T1.1 to T1.2). The total pre-programmed flow-rate was
2.6 ml/h.
02-08-12 11:37
171
9.3.1.3 Delivered volumes

The actual amount of fluid delivered from Pump 1 ranged from 103 to 110% of the expected
delivery (90 min * 2.5 ml/h = 3.75 ml) in all of the tests. The percentages of fluid delivered
from Pump 2 relative to the expected delivery (60 min * 0.1 ml/h = 0.1 ml) were 52 21%,
32 40%, 12 23% and 12 24% (mean SD) in the tests without a check valve, with a LOP
valve, with a LOP-S valve and with a HOP-S valve, respectively.
9.3.2 Experiment II: Changing height
9.3.2.1 Start-up time
The start-up time of Flow Meter 1 in Experiment II was 2.00 0.25 minutes (mean SD)
in the tests without a check valve. The longest measured start-up time was 4.62 minutes,
measured when the LOP-S valve was used. For Flow Meter 2, in eight of the twelve tests,
the 0.075 ml/h set point was not reached within the first 30 minutes of the experiment. In
the remaining four tests (two tests without a check valve and once for each of the LOP and
LOP-S valves), the start-up time was >20 minutes.
After the height of the pumps was increased to 30 cm at T2.1, an increase in flow in both
flow meters was measured. As a result of this increase, the flow in Flow Meter 2 reached the
pre-programmed flow (0.1 ml/h) in all of the tests, except for the tests performed with the
HOP-S check valve. At T2.2, when the pumps were returned to their original positions, a decrease in flow was measured in both flow meters. In Flow Meter 1, the flow did not decrease
below 2.1 ml/h. In Flow Meter 2, the flow decreased below 75% of the pre-programmed
flow-rate in all of the tests, with a minimum measured flow of -0.3 ml/h. Table 9.4 shows
the times required to reach a flow-rate of 0.075 ml/h in Flow Meter 2 after the height of the
pumps was changed.
Table 9.4 Start-up times (time to reach 75% of the pre-programmed flow-rate) in Experiment II for Flow
Meter 2 after decreasing the height of the pumps (at T2.2). The pre-programmed flow-rate was 0.1 ml/h.
Flow Meter 2 (0.1 ml/h)
Start-up time after decreasing
the height of the pumps (at T2.2)
Flow-rate distribution
(time period: T2.0 - T2.3)
mean SD (number of tests with

a start-up time <30 minutes) [minutes]
median (25th - 75th percentile)

[ml/h]
No check valve
6.0 1.7 (n = 3 out of 3)
0.072 (0.026 - 0.090)
LOP
6.5 1.8 (n = 3 out of 3)
0.068 (-0.003 - 0.083)
LOP-S
9.3 2.6 (n = 3 out of 3)
0.049 (-0.003 - 0.074)
HOP-S
25.6 (n = 1 out of 3)
0.016 (-0.002 - 0.034)
Experiment II
02-08-12 11:37
172
CHAPTER 9
9.3.2.2 Flow-rate variability

The total time that the flow measured in Flow meter 1 was within 75 to 125% of the preprogrammed flow-rate of Pump 1 (i.e., between 1.875 and 3.125 ml/h) ranged from 95% to
98% of the total duration of the test in all of the tests. In Flow Meter 2, the total time spent
within 75 to 125% of the pre-programmed flow-rate (i.e., between 0.075 and 0.125 ml/h)
was 43%, 36%, 22% and 0% for tests performed without a check valve, with the LOP valve,
with the LOP-S valve and with the HOP-S valve, respectively. The median (25th - 75th percentile) of the flow in Pump 2 is shown in Table 9.4. Figure 9.5 illustrates the corresponding
frequency histogram. The minimum and maximum measured flow-rates in Flow Meter 2
were -0.8 and 0.3 ml/h, respectively.
The minimum and maximum measured flow-rates of the total flow through the catheter
(i.e., the sum of the measured flow in both flow meters) in Experiment II were -1.2 and 9.9
ml/h, respectively. In Figure 9.6, the corresponding frequency histogram is shown.
9.3.2.3 Delivered volumes
The actual amount of fluid delivered from Pump 1 ranged from 100 to 103% of the expected
delivery (90 minutes*2.5 ml/h = 3.75 ml) for all of the tests. For Pump 2, the percentages
of delivered fluid were 56 8%, 52 13%, 18 9% and 18 7% (mean SD) of the expected
delivery (90 minutes*0.1 ml/h = 0.15 ml) for tests without a check valve, with a LOP valve,
with a LOP-S valve and with a HOP-S valve, respectively.
26
LOP
LOP-S
HOP-S
No valve
24
22
20
Occurrence [%]
18
16
14
12
10
8
6
4
2
0
-0.3
-0.2
-0.1
0.1
0.2
0.3
Flow rate [ml/h]
Figure 9.5 Frequency histogram of the measured flow from Pump 2 in Experiment II. Only flow-rates
ranging from -0.3 to 0.3 ml/h are shown. The frequency histogram includes the flow from the moment
Pumps 1 and 2 were started until the end of the experiment (time period: 90 minutes, from: T2.0 to
T2.3). The pre-programmed flow-rate was 0.1 ml/h.
02-08-12 11:37
173
12
LOP
LOP-S
HOP-S
No valve
Occurrence [%]
10
2.2
2.3
2.4
2.5
2.6
2.7
2.8
Flow rate [ml/h]
Figure 9.6; Frequency histogram of the measured flow through the catheter (sum of the flow
measured in both Flow meters 1 and 2) in Experiment II. Only flow-rate ranging from 2.2 to 2.8 ml/h are
shown. The frequency histogram includes the flow from the moment Pumps 1 and 2 were started until
the end of the experiment (time period: 90 minutes, from: T2.0 to T2.3).The total pre-programmed flowrate was 2.6 ml/h.
9.4 Discussion
The aim of this study was to evaluate the effect of three different types of check valves on
the flow characteristics in a low-flow multi-infusion set for preterm infants. To visualise the
flow characteristics, an in vitro experimental set-up with two syringe pumps and two inline flow meters was used. The results show that fluids at a low pre-programmed flow-rate
(0.1 ml/h) in combination with a higher pre-programmed flow-rate (2.5 ml/h) flow much
slower than expected, both with and without the use of check valves. The actual amount
of the low-flow fluid delivered through the catheter at the end of the experiments varied
within the tests, but was, at best, approximately half of the expected amount. The percentage of time the actual flow-rate was between 75 to 125% of the low pre-programmed
flow-rate was 0% for the check valve with the highest opening pressure and was 43%, at
most, for tests without a check valve. Overall, the distribution in flow and the actual delivery of fluids using a pre-programmed flow-rate of 2.5 ml/h were much more accurate
than those obtained using the low-flow setting.
02-08-12 11:37
174
CHAPTER 9
In clinical practice, the results for low pre-programmed flow-rates will be worse than those
observed in this study for several reasons. First, in clinical practice, the number of syringes
connected to one catheter is often greater than two. Increasing the number of lines connected to one catheter will lead to a more complex interaction between fluids and will consequently cause more flow differences. Secondly, in clinical practice, the total internal volume
of the administration system will probably be larger than what was used in this study because of the use of, for example, larger syringes and longer infusion lines. This larger volume
may result in a larger compliance of the system and, consequently, more flow differences. In
this study, the syringes had a volume of 20 ml. Larger syringes (i.e., 50 ml) are associated with
a higher compliance and will therefore increase start-up times and flow-rate variability.3, 21
Third, in clinical practice, there will be more frequent and possibly larger changes in height
between the catheter tip and the syringe pumps. In this study, the increase and decrease in
the height of the pumps was chosen to be 30 cm. This change in height was based on the
maximal change possible in the incubator used in the studys centre (Caleo, Drger, Lbeck,
Germany). The changes in flow caused by changes in hydrostatic pressure will also be greater
when the differences in height are increased or occur more frequently.
The actual effect of check valves on the IV-administration set-up depends on factors including the position of the check valve, the opening pressure of the check valve, and the preprogrammed flow-rate. In this experimental set-up, the check valves were placed directly
after the syringe. Thus, backflow of IV fluids could occur in the lines, but fluids could not
return into the syringe. A check valve opens when the pressure before the valve (i.e., in the
syringe) minus the pressure after the valve (i.e., in the infusion line) is higher than the opening pressure of the valve. When a check valve is positioned distal to the catheter tip, i.e., next
to the stopcock, the advantage is that fluids cannot flow back into the infusion lines. However, the pressure to open the check valve has to be increased both in the syringe and in the
infusion lines. Therefore, it takes longer for the check valve to open. The position chosen in
this study prevents backflow into the syringe while minimising the time required to reach
the opening pressure of the check valve. The higher the pre-programmed flow-rate of the
syringe pump and the lower the compliance of the syringe, the faster the opening pressure
of the check valve will be reached.
The results also showed that turning the three-way valves on the stopcock to add an extra
pump to the IV-administration set has a short, but very large, influence on the flow in the
infusion lines. At that moment, the pressure levels in the system suddenly change. Fluids
will follow the path of least resistance and will thus flow into the line where the pressure is
lowest. The time to reach a stable situation after introducing an extra syringe depends on
factors including the pre-programmed flow-rates, the compliance of the IV-administration
set and the type and position of the check valves.
As far as we know, our study is the first to measure the flow-rate in the infusion lines. The
advantage of this method is that accurate and frequent data of the flow-rate characteristics
can be obtained. The flow meters used in this study were able to yield bi-directional measurements with a frequency of 100 Hz. The knowledge of flow-rate characteristics is espe-
02-08-12 11:37
175
cially useful for very low pre-programmed flow-rates that are used in NICUs. The design of
this study allowed for the actual amount of fluid that flows backward or forward into the
lines to be shown in detail.
The results of our study are consistent with other studies focussing on flow-rate
variability.8, 11, 20-27 McCarroll et al.16 tested anti-siphon valves at flow-rates greater than 2 ml/h
and showed longer start-up times when check valves were used compared to the control setup without valves. Weiss et al.20 studied whether pressure regulation by check valves reduces
the flow-rate variation during vertical displacement (50 cm) of the pumps. The time without
fluid delivery increased significantly when check valves were introduced and was dependent
on the type of syringe used. One of the most recent studies, performed by Donald et al.,23
tested the effect of a 30 cm upward and downward displacement for two different syringe
pumps with a 50 ml syringe, and a pre-programmed flow-rate of 2 ml/h. Each test was repeated three times. They showed that when the pumps were either elevated or lowered, the
delivery rate was inversely related to the pre-programmed flow-rate. One study used a preprogrammed flow-rate of 0.1 ml/h.3 In this study by Neff et al., the time from the start of the
infusion to reach steady state flow was measured for four different syringe sizes (10, 20, 30,
and 50 ml). The time varied from 3.6 0.9 minutes for a 10 ml syringe at a flow-rate of 2 ml/h
to 74.5 26.6 minutes for a 50 ml syringe at a flow-rate of 0.1 ml/h. These studies did not
perform tests with a multi-infusion set. According to our knowledge, our study is the first to
show the effects on the flow within the infusion lines when a second infusion was added to
the system, when check valves were added and when the height of the pumps was changed.
The necessity for check valves with or without anti-siphon function in IV systems remains
unresolved. While the use of the valves prevents backflow and/or siphoning, the flow-rate
can be worsened because of the introduction of these check valves in the IV-administration
set. Literature on this topic is scarce, and the actual clinical effect of the flow-rate variability
depends on the type of fluid that is administered and the patient characteristics, among
other factors. For example, in small infants, variation in flow-rate in highly concentrated
inotropic or vasoactive drugs has been associated with serious haemodynamic effects.5-7, 10
Nevertheless, although even check valves with low opening pressure can cause a delay in
the delivery of fluids, we advise their use to ensure that IV fluids flow in only one direction.
It is the task of manufacturers to develop check valves (with anti-siphon function) with the
lowest possible opening pressure to minimise the negative effects of check valves on flowrate variability. Until it is possible to prevent both backflow and siphonage with a (very)
low opening pressure, we advise medical staff to evaluate the need for check valves (with or
without anti-siphon function) for each situation separately.
To conclude, this study showed that actual flows in a multi-infusion administration set
with a low pre-programmed flow-rate are much lower than the pre-set value, and the total
delivered volume is far less than expected both with and without the use of check valves. To
optimise the delivery of IV fluids, we advise the use of the highest possible pre-programmed
flow-rate and to evaluate the need for check valves (with or without anti-siphon function)
for each situation separately.
02-08-12 11:37
176
CHAPTER 9
9.5 References
1. Loisel D, Smith M, MacDonald M, Martin G. Intravenous access in newborn infants: impact of extended
umbilical venous catheter use on requirement for
peripheral venous lines. J Perinatol. 1996;16(6):461-466.
2. Hermansen M, Goetz Hermansen M. Intravascular
catheter complications in the neonatal intensive
care unit. Clinics in Perinatology. 2005;32(1):141-156.
3. Neff S, Neff T, Gerber S, Weiss M. Flow rate,
syringe size and architecture are critical to start-up
performance of syringe pumps. European Journal
of Anaesthesiology. 2007;24(7):602-608.
4. Bartels K, Moss D, Peterfreund R. An analysis
of drug delivery dynamics via a pediatric central
venous infusion system: quantification of delays in
achieving intended doses. Anesthesia & Analgesia.
2009;109(4):1156-1161.
5. Schulze K, Graff M, Schimmel M, Schenkman A, Rohan
P. Physiologic oscillations produced by an infusion
pump. Journal of Pediatrics. 1983;103(5):796-798.
6. Leff R, Roberts R. Problems in drug therapy for
pediatric patients. American Journal of Health-System
Pharmacy. 1987;44(4):865-870.
7. Capes D, Dunster K, Sunderland V, McMillan D,
Colditz P, McDonald C. Fluctuations in syringe-pump
infusions: association with blood pressure variations in
infants. American Journal of Health-System Pharmacy.
1995;52(15):1646-1653.
8. Lnnqvist P, Lfqvist B. Design flaw can convert
commercially available continuous syringe pumps to
intermittent bolus injectors. Intensive Care Medicine.
1997;23(9):998-1001.
9. Cunningham S, Deere S, McIntosh N. Cyclical variation
of blood pressure and heart rate in neonates. Arch Dis
Child. 1993;69(1 Spec No):64-67.
10. Klem S, Farrington J, Leff R. Influence of infusion pump
operation and flow rate on hemodynamic stability
during epinephrine infusion. Critical Care Medicine.
1993;21(8):1213-1217.
Spies C, et al. Downward movement of syringe pumps
reduces syringe output. British Journal of Anaesthesia.
2001;86(6):828-831.
12. Evans A, Winslow E. Oxygen saturation and hemodynamic response in critically ill, mechanically ventilated
adults during intrahospital transport. American Journal
of Critical Care. 1995;4(2):106-111.
13. Hurlbut J, Thompson S, Reed M, Blumer J, Erenberg A,
Leff R. Influence of infusion pumps on the pharmacologic response to nitroprusside. Critical Care Medicine.
1991;19(1):98-101.

Care Medicine. 2009;10(5):592-596.
16. McCarroll C, McAtamney D, Taylor R. Alteration in
flow delivery with antisyphon devices. Anaesthesia.
2000;55(4):355-357.
17. Rooke G, Bowdle A. Syringe Pumps for Infusion
of Vasoactive Drugs. Anesthesia & Analgesia.
1994;78(1):150-156.
18. Southern D, Read M. Lesson of the Week: Overdosage
of opiate from patient controlled analgesia devices.
infusions in the newborn: Problems and possible
solutions. Pediatric Critical Care Medicine.
2010;11(2):275-281.
21. Kim D, Steward D. The effect of syringe size on the
performance of an infusion pump. Pediatric Anesthesia. 1999;9(4):335-337.
22. Neff T, Fischer J, Schulz G, Baenziger O, Weiss M.
Infusion pump performance with vertical displacement: effect of syringe pump and assembly type.
Intensive Care Medicine. 2001;27(1):287-291.
23. Donald A, Chinthamuneedi M, Spearritt D. Effect
of changes in syringe driver height on flow: a small
quantitative study. Critical Care and Resuscitation.
2007;9(2):143-147.
24. Weiss M, Hug M, Neff T, Fischer J. Syringe size and
flow rate affect drug delivery from syringe pumps.
25. Weiss M, Bnziger O, Neff T, Fanconi S. Influence of
infusion line compliance on drug delivery rate during
acute line loop formation. Intensive Care Medicine.
2000;26(6):776-779.
26. Weiss M, Fischer J, Neff T, Baenziger O. The effects
of syringe plunger design on drug delivery during
vertical displacement of syringe pumps. Anaesthesia.
2000;55(11):1094-1098.
27. Dnmez A, Araz C, Kayhan Z. Syringe pumps take too
long to give occlusion alarm. Pediatric Anesthesia.
2005;15(4):293-296.
02-08-12 11:37
177
02-08-12 11:37
178
02-08-12 11:37
CHAPTER 10
General discussion
02-08-12 11:37
180
CHAPTER 10
This chapter discusses the main conclusions and the research approach of the work performed in this thesis. Furthermore, recommendations for future work are provided. The
primary objective of this thesis was to determine the limitations of supplemental oxygen therapy and intravenous therapy in current neonatal intensive care and to identify
areas for improvements (see 1.8). To meet this objective, research was performed in
cooperation between the Department of Biomechanical Engineering of the Delft University of Technology, and the Department of Neonatology of the Erasmus Medical Centre Sophia Childrens Hospital in Rotterdam.
02-08-12 11:37
General discussion
181
10.1 Main conclusions

The research in this thesis covers two distinct subjects. Part I (Chapters 2 to 7) of the thesis
concerns the evaluation of the process of supplemental oxygen therapy in preterm infants.
In Part II (Chapters 8 and 9) the flow-rate variability in intravenous (IV) therapy for newborn infants is studied. The main conclusions of the two parts are presented below:
Monitoring oxygenation (Chapter 2 and 3):
Simultaneously obtained pulse oximeter measurements from the feet of extremely low
birth weight (1000 g.; ELBW) infants differ from each other, both for equal and different brands (median (range) SpO2 was 2 (0 to 26) %). The variances of the fluctuations
of the three tested pulse oximeter brands were not significantly different. Additionally,
differences in the SpO2 were observed after switching pulse oximeter probes between
feet. These differences suggest that the handling of the pulse oximeter in neonatal intensive care, like the place and positioning of the probe, influences the performance of the
pulse oximeter the most.
Current available techniques for monitoring oxygenation in preterm infants are suboptimal with respect to accuracy and user-friendliness. Furthermore, the techniques can
not provide a complete overview of the patients oxygenation (oxygen delivery, oxygen
demand, and oxygen reserve) .
Manual control of the oxygenation (Chapter 4, 5, and 6):
To reach and maintain adequate oxygenation of the preterm infant, healthcare professionals manually adjust the fraction of inspired oxygen (FiO2) in the gas mixture that is
supplied to the patient. These manual adjustments in FiO2 vary widely in frequency and
step size for equal SpO2 values.
For nurses, a change in SpO2 is the most important trigger to adjust FiO2. Neither the
sources of information that should be used to determine oxygen requirement, nor the (sequence of) actions that should be performed to recover oxygenation are well defined and
both vary among NICU personnel.
While the alarm limits for SpO2 deviate from the protocol frequently, the motivations for
these non-protocolled alarm limits are rarely registered in patient charts.
The (unregistered) variation in FiO2 adjustments and in alarm limits for pulse oximetry
may influence outcome results without providing a clear insight in the causes for the outcome differences, and should therefore be minimised.
Hazards in supplemental oxygen therapy (Chapter 7):
A Failure Mode and Effects Analysis (FMEA) showed that the top ten hazards in the process of supplemental oxygen therapy in very preterm infants can be categorised in three
main topics: incorrect adjustment of the FiO2, incorrect alarm limits for SpO2, and incorrect alarm limits for SpO2 on patient monitors for temporary use. The recommendations
to mitigate the hazards were aimed at technical developments in patient monitoring, and
at the improvement of checklists and structural repetitive education programs.
02-08-12 11:37
182
CHAPTER 10
10.1.2 IV therapy
Flow-rate variability in intravenous (IV) therapy (Chapter 8 and 9):
Flow-rate variability in IV therapy is affected by changes in hydrostatic pressure, the
specifications of the IV-administration set (volume, stiffness, etc.), and the preprogrammed flow-rate. The clinical relevance of the flow-rate variability depends on the
patient characteristics and the type of substances supplied to the patient.
The actual flows in a multi-infusion set-up with a low pre-programmed flow-rate are
much lower than the pre-set value on the pump. The total delivered volume is far less
than expected both with and without the use of check valves.
10.2 On the research approach

To come to the conclusions described above, two literature reviews and five different
studies were performed. In the next paragraphs the motivations for, and the limitations of
the five studies are discussed briefly.
A review of literature on the monitoring of oxygenation in (preterm) newborn infants
showed that pulse oximetry is easily influenced by artefacts (Chapter 2). To study the actual
performance of pulse oximetry in clinical practice, three different new-generation pulse
oximeters were tested in ELBW infants (Chapter 3). The choice was made to record SpO2
measurements from two feet simultaneously instead of comparing single SpO2 measurements with the oxygen saturation in the blood (SaO2). The decision to make use of dual
SpO2 measurement was made because the comparing of SpO2 values with intermittently
obtained blood samples is hampered by several limitations. Firstly, due to the small amount
of blood in preterm infants (80 ml/kg), the frequency and the amount of blood samples
taken from these patients should be minimised. Secondly, these samples are taken preferably when the infant is in a relatively stable condition. At those stable moments, pulse
oximetry will perform rather well; it is more interesting to know how pulse oximeters perform during desaturations or periods of low blood perfusion. By making use of dual SpO2
measurement we were able to uncover situations that would not have been noticed when
the SpO2 measurements would have been compared intermittently with SaO2 values obtained from blood samples.
An observational study was performed to obtain insight in the behaviour of NICU personnel with respect to the manual control of oxygenation in ELBW infants (Chapter 4 and
5). Although an observational study is useful to obtain insight in the behaviour of subjects,
the observations cannot be used to uncover the actual decision making processes of the
subjects. Therefore, to explore the decision making processes, the observational study
was followed by a survey amongst the nursing staff (Chapter 6). It was decided to make
use of questionnaires to be sure that the nurses could answer at their own convenience
02-08-12 11:37
General discussion
183
without the risk of bias due to an interviewer or other colleagues. A limitation of survey
studies is that the answers of the respondents are not always completely reliable. Reasons
for bias in the results are related to the fact that subjects may be not aware of the actual
reasons for their behaviour, or because subjects do not answer honestly. Reasons for the dishonest answers are, amongst others, a lack of motivation, or the need to give answers that
are publically acceptable.
Based on the results of the first three studies, it was decided to make an overview of the
hazards in supplemental oxygen therapy for preterm infants. To make this overview, an
analysis was performed using the Failure Mode and Effects Analysis (FMEA) tool (Chapter
7). One of the unique features of an FMEA is the fact that the analysis is performed by a
multidisciplinary team. Although performing an FMEA is time-consuming and needs
organizational commitment, the method is a useful tool for structural analysis and identification of (unnoticed) errors, and has been demonstrated to increase safety.1
10.2.2 IV therapy
The fifth study evaluated the influence of three different types of check valves on the flowrate variability in a low flow in-vitro multi-infusion set-up (Chapter 9). This study was performed because literature showed that, amongst others, the introduction of check valves
in an IV-administration set can delay the delivery of IV substances, especially for low preprogrammed flow-rates (Chapter 8). To test the effect of the introduction of check valves,
two bi-directional in-line flow meters were positioned in the infusion tubing. The advantage of this set-up was that actual flow rates in the infusion tubing could be measured. A
limitation was that the set-up can not be used in daily practice because of the in-line flow
meters. Consequently, NICU personnel does not receive feedback about the actual realtime flow-rates in IV therapy in daily practice.
10.3 The need for standardization

The overall conclusion that can be drawn from Part I of this thesis is that there is variation
in the process of the manual control of oxygenation in preterm infants. Part 2 of this thesis
showed, amongst others, that in IV therapy there are endless possibilities to vary the composition of the parts in IV-administration sets. As a result, the type and order of the parts in
the IV-administration set vary between hospitals, departments, and even between patients.
Also, although it was not shown in this thesis, it is to be expected that the (sequence of)
acts in administering IV therapy varies amongst healthcare professionals. Thus, in current
clinical practice, there is no standard approach for either supplemental oxygen therapy or
IV therapy.
The lack of standardization in daily practice for both supplemental oxygen therapy
and IV therapy is undesirable for various reasons. One of the reasons is that, due to the
02-08-12 11:37
184
CHAPTER 10
variation in care, unnoticed differences in care between patients can occur. An example
of non-standardised care is the non-compliance to the protocol for pulse oximetry alarm
limits (Chapter 5). However, since it is not known whether the non-compliance to the protocol influences the outcome of patients in either a negative or a positive way, it should
be realised that the actual deviations from the protocol are not per definition undesirable.
Moreover, as the current protocol in the NICU of the Erasmus MC - Sophia Childrens
Hospital provides only two different alarm limit settings, it could be argued easily that this
protocol is invalid or incomplete for the patient group it is used for. Thus, the actual weak
link in this example is not the deviation from the protocol for alarm limits but the fact that
arguments for the deviations are rarely registered in patient charts. This poor registration
could lead to unnoticed and untraceable (motivations for) deviations.
The lack of standardization in clinical practice may also lead to uncertainty of healthcare
professionals. It has been shown that nurses in operating theatres prefer clear rules. When
protocols are not present, incomplete or insufficient, nurses may become insecure about
the best practice, optimal care etc. Ironically, this uncertainty about the best practice is
probably also the main reason for the lack of standardization and the absence of sufficient
protocols. Both in supplemental oxygen therapy and in IV therapy there are uncertainties
about the best methods and solutions for daily recurring situations, problems and questions. If the solutions to these uncertainties were straightforward, the development of and
compliance to protocols would not lead to any difficulties.
Unfortunately this is not the case. Although the knowledge about neonatal intensive care
has improved enormously in the last few decades there are still (and always will be) issues
to solve. Working with these unanswered questions and situations is the reality with which
healthcare professionals are confronted every day. Besides dealing with these uncertainties
about optimal care, healthcare professionals have to anticipate the fact that each patient is
unique, and could, therefore, need specific treatment. Being aware of this, it is not surprising that the persons responsible for the development of protocols, i.e. the physicians, think
that guidelines are unnecessary and potentially harmful.2
Paradoxically, as long as protocols are lacking or insufficient, and NICU personnel does
not register motivations for deviations in an adequate way, it will virtually be impossible
to define which strategies or therapies are superior. Thus, there is a vicious circle: To solve
the uncertainties in daily practice, research needs to be performed. To obtain reliable results in either retrospective or prospective research, daily care should be standardised. To
standardise care, uncertainties in neonatal intensive care need to be solved.
Again, it should be emphasised that standardization of care does not implicate that all
patients receive the same treatment: it only guarantees the presence of a standard. The
presence of this standard makes the healthcare professionals more aware of the fact the he
02-08-12 11:37
General discussion
185
decides (not) to deviate from the standard and to register his decision in the patient records.
Furthermore, it is very conceivable that one standard prescribes different care for different
patient groups (for instance in the protocol for pulse oximetry alarm limits).
10.4 Future work

To improve both supplemental oxygen therapy and IV therapy in neonatal intensive care,
the focus on the work in the (near) future should lie on the breaking of the vicious circle in
the standardisation of care. To break the vicious circle, both changes in culture3-5 and technical improvements may help.
10.4.1 Changes in culture
Changes in culture are, for example, changes in education and authorisation or autonomy
of NICU personnel. With respect to supplemental oxygen therapy, several studies showed
that healthcare professionals have inadequate theoretical understanding of pulse oximetry
and/or other basic concepts in neonatal oxygenation.6-14 Other studies showed that implementation of a structured program, education, and/or clear handover processes leads to
fewer errors.19, 20 In IV therapy, education could increase awareness among healthcare professionals about, for instance, the laws of physics (e.g., principles of compliance, siphonage,
or interconnected tanks).
Changes in authorisation or autonomy could be implemented in the policy for pulse oximetry alarm limits. Nurses permission to adjust alarm limits could be restricted, and it
could be demanded that motivations for alarm limit adjustments are registered in the patient records. For sure, the implementation of structured education programs and/or the
requirement for written motivations will cost extra time. However, it is to be expected that
the invested time pays back in safer and more effective care.15-19
Simultaneously with the implementation of structured education programs and more
adequate administration, a start could be made to develop (or improve) protocols. For supplemental oxygen therapy, protocols could help to describe which input parameters and
which (sequences of) tasks should be performed to control oxygenation.20, 21 In IV therapy,
protocols could help to prescribe the optimal assembly for an IV-administration set. For
sure, not only the guideline itself, but also the discussion about the topics with healthcare
professionals will lead to a greater attention for the actual needs in both supplemental oxygen therapy and IV therapy.
To support healthcare professionals with the development of protocols, with the implementation of the protocols in clinical practice, and with the reduction of time required for
administration, technical improvements may help.
02-08-12 11:37
186
CHAPTER 10
10.4.2 Technical improvements

The most straightforward solution to e.g., reduce the time required for administration, is
the realisation of the possibility for medical devices and digital patient charts to communicate with each other. Unfortunately, this is also probably the most challenging solution.
For instance, in the NICU of the Erasmus MC Sophia Childrens Hospital the mechanical ventilator, the incubator, and other relevant medical devices such as IV-pumps are not
connected to each other or to the PC with the patient data management system (PDMS).
Currently, the only communication that is possible between medical devices is the information exchange from the patient monitor to the PC with PDMS. Data exchange between
medical devices could be useful because it creates opportunities to build in automatic
checks for differences in the prescribed and the actual supplied therapies.
Concerning supplemental oxygen therapy, the information flow between medical devices
could help to improve alarm management. Currently, when nurses adjust alarm limits in
the patient monitor, the change in alarm limits has to be registered in the PDMS system
manually. This manual registration is time-consuming and easy to forget. Communication between the patient monitor and PDMS could create the possibility to automatically
register alarm limits in the PDMS. This guarantees correct registration. An extra implementation could be that alarm limits can only be adjusted when authorization for the adjustment is provided in PDMS first. Another possibility is that the protocol is implemented in
the patient monitor, to make sure that limits are adjusted automatically according to the
protocol (e.g., when FiO2 is set from >21% to 21% and vice versa). For safety reasons, an
additional pop-up could be used to give notice of the automatic adjustment.
Communication between devices may improve safety in IV therapy as well. For instance,
the patient characteristics, the administered substances, the prescribed medication, and the
required flow-rate could be checked by the syringe pump in the PDMS system. When these
data do not match, or when a potentially dangerous combination of drugs is to be supplied,
the pump can give an alarm to warn the healthcare providers of the potential error. Also, the
exact moment of the start and end of the therapy can be registered automatically.
Besides communication between medical devices, technical improvements could also help
to increase the meaning of (pulse oximetry) alarms by a reduction of the number of false
or irrelevant alarms. Currently, pulse oximetry alarms often ring without the presence of
a life-threatening situation. The alarm is more a notice, rather than a demand for action.
These notices may lead to a reduced response of the staff, i.e. the crying wolf effect.22 For
effective alarming, a more intelligent system is preferred, probably based on multiple monitored parameters. To reduce response time and to improve performance, these systems may
be developed such that they do not only inform about, but also advise which (subsequent)
activities are needed.23-27
02-08-12 11:37
General discussion
187
10.5 Conclusion
The research in this thesis was performed to determine the limitations of supplemental
oxygen therapy and IV therapy in current neonatal intensive care and to identify areas for
improvements. The main conclusion is that both supplemental oxygen therapy and IV
therapy could benefit from a more standardised approach. In both therapies variations in
clinical practice are observed. The actual effects of the variations in therapy depend on multiple factors, and are therefore not easy to predict. However, it is certain that when both
supplemental oxygen therapy and IV therapy are not standardised, or when deviations from
the protocols are not registered in patient records, unnoticed differences in care between
patients can occur. These unnoticed (and untraceable) differences may lead to differences in
outcome of patients and hamper the research to find optimal strategies.
To reach standardisation, changes in culture, including education, and technical improvements may help. Although the actual effect of each (small) change in culture and technology
on the outcome of NICU patients will be very difficult to quantify, it is likely that all small
changes together will enhance the quality of care in daily practice in neonatal intensive
care. Hopefully, the insights obtained during the studies described in this thesis will not
only lead to an improvement in the quality of supplemental oxygen therapy and IV therapy
in preterm infants, but will also lead to better care for NICU patients in general.
02-08-12 11:37
188
CHAPTER 10
10.6 References
1. McDermott RE, Mikulak RJ, Beauregard MR. The
basics of FMEA. Portland: Productivity Press; 1996.
2. McDonald R, Waring J, Harrison S, Walshe K,
Boaden R. Rules and guidelines in clinical practice:
a qualitative study in operating theatres of doctors
and nurses views. Quality and Safety in Health Care.
2005;14(4):290.
3. Edwards W. Patient safety in the neonatal intensive
care unit. Clinics in Perinatology. 2005;32(1):97-106.
4. Leonard M, Graham S, Bonacum D. The human factor:
The critical importance of effective teamwork and
communication in providing safe care. British Medical
Journal. 2004;13(Supplement 1):i85-90.
5. Berwick D. Errors today and errors tomorrow. New
6. Elliott M, Tate R, Page K. Do clinicians know how to
use pulse oximetry? A literature review and clinical implications. Australian Critical Care. 2006;19(4):139-144.
7. Popovich DM, Richiuso N, Danek G. Pediatric health
care providers knowledge of pulse oximetry. Pediatric
Nursing. 2004;30(1):14-20.
9. Attin M, Cardin S, Dee V, Doering L, Dunn D, Ellstrom
K, et al. An educational project to improve knowledge
related to pulse oximetry. American Journal of Critical
Care. 2002;11(6):529-534.
10. Fouzas S, Politis P, Skylogianni E, Syriopoulou T, Priftis
KN, Chatzimichael A, et al. Knowledge on pulse
oximetry among pediatric health care professionals: A
multicenter survey. Pediatrics. 2010;126(3):e657-662.
11. Stoneham MD, Saville GM, Wilson IH. Knowledge
about pulse oximetry among medical and nursing staff.
The Lancet. 1994;344(8933):1339-1342.
12. Kruger PS, Longden PJ. A study of a hospital staffs
knowledge of pulse oximetry. Anaesthesia and
Intensive Care. 1997;25(1):38-41.
13. Solberg MT, Hansen TWR, Bjrk IT. Nursing assessment during oxygen administration in ventilated
preterm infants. Acta Paediatr. 2010;100(2):193-197.
14. Sola A, Lee B. Education in neonatal oxygenation has
been insufficient: a need for darning. In: Society for
Pediatric Research; 2007; Toronto, Canada; 2007.
15. Grimshaw J, Thomas R, MacLennan G, Fraser C,
Ramsay C, Vale L, et al. Effectiveness and efficiency of
guideline dissemination and implementation strategies. International Journal of Technology Assessment
in Health Care. 2005;21(01):149-149.
16. Cosby J. Improving patient care: the implementation

of change in clinical practice. Quality and Safety in
Health Care. 2006;15(6):447.
17. McColl E. I just want the protocol, doctor! Quality and
Safety in Health Care. 2005;14(3):155.
18. West J, Wright J, Tuffnell D, Jankowicz D, West R. Do
clinical trials improve quality of care? A comparison of
clinical processes and outcomes in patients in a clinical
trial and similar patients outside a trial where both
groups are managed according to a strict protocol.
19. Grimshaw J, Russell I. Effect of clinical guidelines on
medical practice: a systematic review of rigorous
evaluations. The Lancet. 1993;342(8883):1317-1322.
Pediatrics. 2003;111(2):339-345.
2008;95(4):306-310.
22. Lawless ST. Crying wolf: False alarms in a pediatric intensive care unit. Critical Care Medicine.
1994;22(6):981-985.
23. Bitan Y, Meyer J, Shinar D, Zmora E. Nurses reactions
to alarms in a neonatal intensive care unit. Cognition,
Technology & Work. 2004;6(4):239-246.
24. Imhoff M, Kuhls S. Alarm Algorithms in Critical
Care Monitoring. Anesthesia & Analgesia.
2006;102(5):1525-1537.
25. Edworthy J, Hellier E. Fewer but better auditory alarms
will improve patient safety. British Medical Journal.
2005;14(3):212-215.
26. Westenskow D, Orr J, Simon F, Bender H, Frankenberger H. Intelligent alarms reduce anesthesiologists
response time to critical faults. Anesthesiology.
1992;77(6):1074.
27. Schoenberg R, Sands DZ, Safran C. Making ICU
alarms meaningful: a comparison of traditional vs.
trend-based algorithms. In; 1999: American Medical
Informatics Association; 1999. p. 379-383.
02-08-12 11:37
General discussion
189
02-08-12 11:37
190
02-08-12 11:37
Appendices
02-08-12 11:37
192
In these appendices additional information about the patients (n = 12) included in the
observational study is presented. The main results of the observational study are discussed in Chapter 4 & 5. In Appendix A two tables are shown. The first table (Table A1)
focusses on the patient characteristics. The second table (Table A2) presents characteristics of the recorded data for each patient separately. Appendix B belongs to Chapter
5 and elaborates on the information presented in Table 5.2. It provides detailed descriptions about the circumstances in which the adjustments in pulse oximetry alarm limits
were performed.
02-08-12 11:37
Appendices
193
Appendix A: Detailed patient characteristics
Mortality
BPD4
HFV
NIV
ROP
II
62
108
normal
26 4/7
800
2nd of
twins
HFO
41
84
normal
28
935
1st of
twins
12
SIMV
NIV
CPAP
ROP
II-iii
73
99
normal
27
545
SIMV
46
98
normal
27 3/7
885
2nd of
twins
+0
HFV
SIMV
CPAP
+
D28
NA
NA
NA
NA
deceased
26 6/7
640
+0
SIMV
ROP
I
46
99
normal
26 1/7
845
2nd of
twins
+0
SIMV
46
89
normal
26 1/7
805
1st of
twins
NIV
46
91
normal
26 1/7
690
+0
SIMV
HFO
ROP
III
113
174
normal
10
25 3/7
585
+5
+0
HFO
+
D10
NA
NA
NA
NA
deceased
11
24 2/7
720
only
liveborn of
triplet
+0
55
117
normal
12
25 2/7
875
+0
45
107
normal
SIMV
HFO
Follow up at 1 year
PDA3
Length of stay in hospital [days]
IRDS 2
Length of stay NICU [days]
Intubated at birth
ROP
Asphyxia 1
Multiple births
Caesarean section
Sexe
615
26 2/7
Gestational age [weeks]
Patient
PNA at start recording [days]
Ventilation modes during recording
Table A1 Characteristics of patients (n=12) included in the observational study.
0 = Indomethacine treatment during recording

1 = Asphyxia defined as Apgar score at 5 <7
2 = IRDS defined as surfactant treatment within 24 hours postpartum
3 = PDA diagnosed by echocardiography
4 = BPD, diagnosed when supplemental oxygen therapy was needed at a postmenstrual age of 36 weeks
5 = Surfactant treatment during recording
02-08-12 11:37
194
71.6
55.8
26 4/7
800
51.9
44.0
28
935
12
72.9
60.5
27
545
72.5
53.5
27 3/7
885
72.2
72.2
26 6/7
640
72.8
60.7
26 1/7
845
72.2
71.3
26 1/7
805
72
65.2
26 1/7
690
79.9
73.1
10
25 3/7
585
72.7
62.7
11
24 2/7
720
48.8
43.0
12
25 2/7
875
73.5
64.3
832.93
726
Total
22
(21-25)
27
(24-29)
35
(22-51)
77
(66-85)
25
(24-26)
32
(28-35)
27
(25-28)
21
(21-21)
35
(30-40)
42
(35-50)
24
(22-25)
30
(29-35)
100
2.9
41
64
2.4
80
121
3.3
38
62
1.9
136
204
4.7
77
125
3.3
86
148
3.3
25
26
0.8
83
121
2.8
67
102
2.7
70
100
4.0
87
136
3.5
851
1309
2.97
-2
(1, 34)
(-1, -34)
-2.5
(1, 16)
(-1, 17)
-4
(1, 40)
(-1 -40)
5.5
-3
(1, 37)
(-1 -50)
-2
(1, 28)
(-1, 30)
-3
(1, 40)
(-1, 57)
-2
(1-35)
(-1, 41)
-4.5
(1, 29)
(-1, 29)
-3
(1, 65)
(-1, 65)
-3
(1, 25)
(-1, 23)
-2
(1, 27)
(-1, 28)
-4
(1, 22)
(-1, 19)
10.7
28.7
6.7
23.6
17.0
7.6
37.8
1.2
59.9
38.5
32.8
22.3
2.3
52.7
7.0
21.7
33.8
4.0
47
39.4
38.5
13.8
1.3
60.5
0.6
28.7
43
8.3
52.4
43.3
41.3
14.7
2.4
64.6
0.0
22.2
5.6
0.0
70.9
68.6
29.4
13.3
0.5
62.4
62.4
16.6
12.3
0.8
42.4
37.3
39.4
13.6
0.6
62.4
36.4
34.7
13.0
1.0
654.5
361.2
0.3
Median (min-max) SpO2
55.7
Number of Lowsats with

minimum SpO2 <80 per hour
Step size FiO2 adjustments down

median (min, max)
Step size FiO2 adjustments up median

(min, max)
Number of FiO2 adjustments per hour
Number of FiO2 adjustments down
Number of FiO2 adjustments up

61
Time spent in Lowsat

(SpO2 <88%) [%]
Time spent in Highsat

(SpO2>94%) [%]
615
Time period that FiO2 >21% and alarm

limits 88-94 [hours]
26 2/7
Time period that FiO2, SpO2, and

alarm limits were available for analysis
[hours]
FiO2 median (interquartile range) [%]
FiO2 recorded [hours]
Duration of observation [hours]
PNA at start recording [days]
Patient
Gestational age [weeks]
Table A2 Characteristics of recorded data for each patient separately.
93
(61-100)
89
(65-100)
92
(29-100)
90
(68-100)
93
(48-100)
89
(74-100)
94
(46-100)
93
(86-99)
93
(47-100)
92
(50-100)
93
(42-100)
93
(7-100)
02-08-12 11:37
Appendices
195
Appendix B Detailed descriptions of alarm limit adjustments

Patient 1
Patient 1 was born at 26 2/7 weeks with a birth weight (BW) of 615 grams. Data recording
started at postnatal day 4. In total 56 hours of observation were suitable for analysis. Alarm
limits were adjusted twelve times by six different nurses. These adjustments were made
during day (3 times), evening (4 times) and night shifts (5 times). Three adjustments were
not according to the protocol.
The first of the non-protocolled adjustments was when the patient had a thorax drain, was
on room air and on high frequency ventilation (HFV). In the five minutes before and after
the adjustment the incubator was opened shortly but the patient was not disturbed. The
alarm limits were adjusted from 90 - 96% to 90 - 98%, thus both the upper and lower limit
were not according to the protocol (88 - 100%). This adjustment was not written in the
(digital) patient records.
In the second adjustment, alarm limits changed from 90 - 98% to 88 - 96%. The incubator
was closed. The adjustment was made 2.5 hours later than the first adjustment, in the same
shift but by a different nurse. At the time of adjustment, the FiO2 was changed from 23 to
25%, thus, although the limits were set in the correct direction, the upper limit was still too
high (should be 88 - 94%).
The third adjustment was made two days later, in the early morning. In the night before
the adjustment, the patient was extubated. Since the extubation the patient needed supplemental oxygen therapy again. According to the notes of the nurse, she was able to decrease
the FiO2 slowly. She adjusted the alarm limits from 88 - 94% to 88 - 100%, while FiO2 was
still between 24 - 22%. The incubator was opened shortly to check the patient.
Patient 2
Patient 2 was born at 26 4/7 weeks, with a birth weight of 800 grams. Data recording started
on day 9. This patient was on high frequency oscillation (HFO) during the complete 24
hours of observation. Alarm limits were adjusted three times, by a nurse and by a physician,
all during day shifts. In one situation, the adjustment was not according to protocol. The incubator was closed, and the nurse adjusted FiO2 from 26 to 21%. Alarm limits were adjusted
similarly from 88 - 96% to 88 - 99%. Thus, although the patient was on room air, there was
still an upper alarm limit. The adjustment was not noted in the (digital) patient record. In
the PDMS was written that alarm limits were 88 - 94%.
At the moment of adjustment the patient suffered from a patent ductus arteriosus (PDA),
and received Indocid. According to notes of the nurse, the ventilation settings were reduced
and the patient only needed supplemental oxygen therapy when the patient was disturbed
by medical or nursing staff.
02-08-12 11:37
196
Patient 3
Patient 3 was born at 28 weeks weighing 935 g. The data recording started on day 12. Alarm
limits were adjusted five times by three different nurses in 60 hours of recorded data. Three
times according to the protocol in the day shift; in two times the alarm limits deviated from
the protocol, both times in the evening shift. Alarm limits were changed from 88 - 94% to
88 - 100%, while the patient received supplemental oxygen therapy. In the hours before the
alarm adjustment the FiO2 was decreased in multiple steps from ~50% to room air. In the 10
minutes around the adjustment, FiO2 was adjusted from 30 to 21 to 27 and finally to 34%.
Approximately 21 hours after the first non-protocolled adjustment the alarm limits were
changed from 88 - 94% to 88 - 100% by another nurse. At that moment FiO2 was 22%. In the
hours before the adjustment FiO2 was decreased. During both adjustments the incubator
was closed. Neither of these two adjustments were noted in the patient records.
Patient 4
Patient 4 was born at 27 weeks with a birth weight of 545 g. Data recording started on day
9. Alarm limits were adjusted twice in the 53 hours of recording by two nurses. The first
adjustment was performed during an evening shift, and according to the protocol. In the
second adjustment, during a day shift, the alarm limits were adjusted from 88 - 94% to
88 - 98%. Corresponding FiO2 was 85%. At the moment of adjustment, the mother of the
patient was present, the incubator was open and the nurse was performing nursing care.
Two hours later alarm limits were written in the PDMS: per order 88 - 98%. Hence, this deviating alarm limit setting was discussed with a physician. The motivation could come from
the fact that the patient suffered from persistent pulmonary hypertension of the neonate
(PPHN). Probably, because PPHN can be increased by hypoxia, the SaO2 was kept higher
than normal.
Patient 5
In patient 5, data recording started on day 3. The patient was born at 27 3/7 weeks weighing
885 g, and suffered from a PDA. Alarm limits were adjusted eleven times (6 times during
day, 3 times during the evening, and 2 times during the night shift) during 47 hours of recording. These adjustments were made by three different nurses and a physician. Four adjustments were not according to the protocol. The first time was in the early morning. During
the night before, the need for supplemental oxygen varied. In the ten minutes around the
adjustment, the diaper of the patient was changed, and the FiO2 was adjusted from 23 to
21% and back. Alarm limits were adjusted from 88 - 94% to 88 - 100% .
Three hours later, during the day shift, alarm limits were adjusted three times in less than 15
minutes. A physician adjusted FiO2 from 25 to 22 to 21% and adjusted the alarm limits from
88 - 100% to 88 - 96%. Thus, an upper limit was created while the patient was on room air. A
couple of minutes later FiO2 was adjusted to 22%. Another couple of minutes later, a nurse
02-08-12 11:37
Appendices
197
adjusted FiO2 from 22 to 21%. Also, she adjusted the alarm limits from 88 - 96% to 88 - 94%.
Two minutes later the same nurse adjusted FiO2 from 21 to 22% and changed the limits
from 88 - 94% to 88 - 100%. During these adjustments, the incubator was closed.
Patient 6
Patient 6 was born at a gestational age (GA) of 26 6/7 weeks weighing 640 grams. Data recording was started on day 4. The alarm limits were adjusted six times in 60 hours by three
different nurses: Three times during day shifts, and four times during evening shifts. In
five times, the alarm limits deviated from the protocol. At 5 pm an ultrasonography of the
heart was made, showing a PDA. At 11 pm a nurse adjusted the alarm limits from 88 - 100%
to 88 - 99% while the patient received 28% FiO2. This adjustment was not written in the
patient records.
Two days later, in an evening shift another nurse adjusted the limits from 88 - 99% to 88 96%. At that moment FiO2 was adjusted from 38 to 29%. Thus, although the upper limit was
decreased, they were still not according to the protocol. Half an hour later, the same nurse
adjusted to alarm limits to 88 - 94%, according to the protocol. Another half an hour later,
the same nurse adjusted the limits to 92 - 98%. In the time around the adjustment FiO2 was
adjusted from 29 to 30, to 38, and to 37%, and the incubator was closed. In PDMS was written
that the patient suffers from a starting PPHN, and that SpO2 should be kept >95%.
The next morning the third nurse adjusted the limits from 92 - 98% to 88 - 98%, FiO2 was
adjusted from 34 to 40%. Simultaneously the nurse was taking care of the infant. In PDMS
the alarm limits were said to be 92 - 98%. Half an hour later, the same nurse adjusted limit
back to 92 - 98%.
Patient 7
Patient 7 was born after 26 1/7 weeks with a BW of 845 grams. Data recording started on day
3. In the 52 hours of recorded data, alarm limits were adjusted twice in the evening shift by
two nurses. One of these adjustment was not according to the protocol. A nurse adjusted
the limits from 88 - 94% to 88 - 96%. FiO2 was adjusted from 27 to 26%, the incubator was
closed. In the patient records, this adjustment in alarm limits was not noted.
Patient 8
Patient 8 was born after 26 1/7 weeks with a BW of 805 grams, and was twins with patient
#7. Data recording started on day 8. In the 65 hours of recorded data alarms were adjusted
twice by one nurse. Both adjustments were performed in the evening shift, and both were
according to the protocol.
02-08-12 11:37
198
Patient 9
For patient 9 (GA 26 1/7, BW 690 grams) 71 hours of data was available starting from day 2.
Although the need for supplemental oxygen therapy varied, the alarm limits were 88 - 94%
for the complete recording.
Patient 10
The same situation as in patient 9 was present in patient 10 (GA = 25 3/7, BW = 585 grams).
Data recording started on day 2. In the 62 hours of recorded data no alarm limit adjustments were made.
Patient 11
In patient 11 (GA = 24 2/7, BW = 720 grams) one alarm limit adjustment was made in 42
hours of recording. These recordings were started on day 7. This adjustment was according
to the protocol.
Patient 12
In patient 12 (GA = 25 2/7, BW = 875 grams), one adjustment in the alarm limits was performed in 62 hours of recording. These recordings were started on day 4. The adjustment
was made in a day shift, while the incubator was closed. The alarm limits were adjusted
from 88 - 94% to 88 - 95%, the FiO2 was 29%. This adjustment was not noted in the patient
records.
02-08-12 11:37
Appendices
199
02-08-12 11:37
200
02-08-12 11:37
SUMMARY
in neonatal
intensive care:
02-08-12 11:37
202
02-08-12 11:37
Summary
203
In neonatal intensive care units (NICUs), both supplemental oxygen therapy and intravenous (IV) therapy are essential, but potentially dangerous therapies. The objective of this
thesis was to determine the limitations of supplemental oxygen therapy and IV therapy in
current neonatal intensive care and to identify areas for improvements. The research was
performed in cooperation between the Delft University of Technology and the Erasmus
Medical Centre - Sophia Childrens Hospital, Rotterdam. To distinguish between supplemental oxygen therapy and IV therapy, the thesis was divided in two parts.
Part I Supplemental Oxygen therapy

In supplemental oxygen therapy, a gas mixture with >21% of oxygen is supplied to the patient via (mechanical) ventilation. Although supplemental oxygen therapy is often needed
in preterm infants to reach and maintain adequate oxygenation, this therapy is not without risks. To prevent negative outcome, an auditory alarm sounds when the blood oxygen
saturation of the patient (SpO2, measured by pulse oximetry) is outside of the target range
(88 - 94%). One of the actions that can be performed by healthcare professionals to recover
the SpO2 level is the adjustment of the fraction of inspired oxygen (FiO2) in the gas mixture
supplied to the preterm infant. This manual control is difficult and time consuming.
To determine the limitations of supplemental oxygen therapy and to identify areas for improvements, a literature review (Chapter 2) and four studies were executed. One study compared the performance of three different pulse oximeters in clinical practice in extremely
low birth weight (1000 g.; ELBW) infants (Chapter 3). The manual adjustments in FiO2
performed by NICU personnel in ELBW infants were quantified by an observational study
(Chapter 4). Simultaneously, the compliance to the protocol for pulse oximetry alarm limits
was studied (Chapter 5). Subsequently, a survey was held to obtain insight in the decision
making processes of NICU nurses (Chapter 6). Finally, a Failure Mode and Effects Analysis
(FMEA) was performed to identify hazards in the process of supplemental oxygen therapy
(Chapter 7).
The main conclusions of the literature review were that monitoring techniques for preterm infants are suboptimal, and the number of false and less relevant alarms is high.
Additionally, there is no consensus about the optimal target range for blood oxygen levels
in preterm infants. The other studies showed that in the manual control of oxygenation
in preterm infants both the sources of information and the (sequence of) actions to recover oxygenation are not well defined, are quite subjective, and vary among NICU personnel. Although, according to the nurses, a change in SpO2 is the most important trigger to
adjust FiO2, the frequency and step size of the adjustments varied for equal levels of SpO2.
Furthermore, the alarm limits for pulse oximetry deviated from the protocol in one third
of the time. Mostly, the motivations for these deviations were not registered in the patient
charts.
02-08-12 11:37
204
The (unregistered) variation in FiO2 adjustments and in alarm limits may influence outcome results without providing a clear insight in the causes for the outcome differences, and
should therefore be minimised. Recommendations for improvement include suggestions for
(changes in) protocols and education of NICU staff. Technical developments should focus
on the improvement of the accuracy and user-friendliness of sensors for preterm infants,
and on the possibilities for exchange of information between medical devices.
Part II Intravenous therapy

In IV therapy, various types of nutrition, drugs, and/or fluids are administered directly into
the veins of the patient. Because of limited vascular access possibilities in preterm infants,
several infusions are connected to a single vascular access device (i.e. multi-infusion). Due
to the low body weight of preterm infants, the flow-rate that is pre-programmed on the
syringe pumps is relatively low (0.1 - 3 ml/h per pump). Although it would be expected that
IV substances are supplied to the patient with the pre-programmed flow-rate, it has been
shown that the actual volume delivered to the patient can vary over time. Especially in preterm infants, these sudden changes in delivered volume can have severe consequences.
To study which factors are responsible for flow-rate variability in IV therapy, a literature
review was performed (Chapter 8). Subsequently an in-vitro study to evaluate the effect of
three different types of check-valves on the flow-rate characteristics in a low flow multiinfusion set was executed (Chapter 9).
The main conclusions were that the flow-rate variability in (multi-infusion) IV therapy is
influenced by the pre-programmed flow-rate, by changes in the hydrostatic pressure, by the
compliance of the complete IV-administration set, and by the type of substances supplied
to the patient. The clinical relevance of the flow-rate variability depends on the type of IV
substances and the patient characteristics.
To optimise the delivery of IV substances, we advise healthcare professionals to use the
highest allowable pre-programmed flow rate and to use check valves with a low opening
pressure in the IV-administration set. Furthermore, the vertical displacement of syringe
pumps and the internal compliance and the volume of the complete IV-administration set
should be minimised.
02-08-12 11:37
Summary
205
Conclusion
In both supplemental oxygen therapy and IV therapy, (unregistered) variations in clinical
practice are observed. The actual effects of the variations in therapy depend on multiple factors, and are therefore, not easy to predict. However, it is to be expected that the unnoticed
(and untraceable) differences in care lead to differences in outcome of patients and hamper
the research to find optimal strategies. Therefore, both therapies could benefit from a more
standardised approach. To reach standardisation, changes in culture, including education,
and technical improvements may help. Although the actual effect of each (small) change in
culture and technology on the outcome of NICU patients will be very difficult to quantify,
it is likely that all small changes together will enhance the quality of care in daily practice in
neonatal intensive care.
02-08-12 11:37
206
02-08-12 11:37
SAMENVATTING
Observaties op
de neonatale
intensive care unit:
Op weg naar veiligere toediening van zuurstof & IV therapie
02-08-12 11:37
208
02-08-12 11:37
Samenvatting
209
Op neonatale intensive care afdelingen (ICNs) zijn zuurstof therapie en intraveneuze therapie essentile, maar risicovolle therapien. Het doel van dit proefschrift was het bepalen van
beperkingen in het proces van zowel zuurstof therapie als van intraveneuze therapie in de
huidige neonatale intensive care, en om punten voor verbetering te benoemen. Het onderzoek werd uitgevoerd in een samenwerking tussen de Technische Universiteit Delft en het
Erasmus Medisch Centrum - Sophia Kinderziekenhuis, te Rotterdam. Het proefschrift is
verdeeld in twee delen om een onderscheid te maken tussen de beide therapien.
Deel I Zuurstof therapie

Zuurstof therapie houdt in dat de patint een gasmengsel met >21% zuurstof toegediend
krijgt met behulp van een machine voor beademing of ademhalingsondersteuning. Hoewel
zuurstof therapie vaak noodzakelijk is om de oxygenatie van te vroeg geboren patinten op
een juist niveau te brengen en te houden, is deze therapie niet zonder risicos. Om negatieve
gevolgen voor te vroeg geborenen te voorkomen gaat er een auditief alarm wanneer de
zuurstofsaturatie in het bloed van de patint (SpO2, gemeten d.m.v. puls oximetrie) buiten
de ingestelde grenzen komt (88 - 94%). En van de acties die door de verpleegkundigen of
artsen ondernomen kan worden om de SpO2 te herstellen, is het aanpassen van de hoeveelheid zuurstof (FiO2) in het gasmengsel toegediend aan de patint. Dit handmatig regelen
van de SpO2 is moeilijk en tijdrovend.
Om de beperkingen in het proces van zuurstof therapie in kaart te brengen en suggesties
voor verbetering te kunnen doen, werden er een literatuur onderzoek (Hoofdstuk 2) en
vier studies uitgevoerd. En studie vergeleek drie verschillende merken puls oximeters
tijdens het dagelijks gebruik bij patinten met een extreem laag geboorte gewicht (1000
g.; ELBW) (Hoofdstuk 3). Met behulp van een observatie studie werden de door het ICN
personeel uitgevoerde handmatige aanpassingen in de FiO2 bij ELBW patinten gekwantificeerd (Hoofdstuk 4). Ook werd er onderzocht in hoeverre het bestaande protocol voor
alarm limieten voor puls oximetrie werd gevolgd (Hoofdstuk 5). Vervolgens zijn er schriftelijke enqutes uitgedeeld onder de ICN verpleegkundigen om inzicht te krijgen in de beslissingen die zij nemen en de kennis die zij hebben m.b.t. zuurstof therapie (Hoofdstuk 6). Als
laatste werd er een zogenaamde Failure Mode and Effects Analysis (FMEA) uitgevoerd om
de potentiele faalwijzen en bijbehorende gevolgen in het proces van zuurstof therapie te
identificeren (Hoofdstuk 7).
Het literatuuronderzoek bracht naar voren dat de beschikbare technieken voor het monitoren van te vroeg geboren patinten suboptimaal zijn, en dat het aantal valse en/of niet kritieke alarmen groot is. Tevens ontbreekt er consensus over de optimale zuurstofniveaus in
het bloed van te vroeg geborenen. De belangrijkste conclusie die getrokken kon worden uit
de vier vervolg studies was dat tijdens het handmatig regelen van de oxygenatie van te vroeg
geborenen, noch de gebruikte informatiebronnen noch de (volgorde van de) acties om de
02-08-12 11:37
210
oxygenatie te herstellen duidelijk gedefinieerd waren. Ook waren de gebruikte informatie

bronnen en de verrichte acties subjectief en varieerden ze onder het ICN personeel. Hoewel
volgens de verpleegkundigen een verandering in de SpO2 de belangrijkste reden is voor een
aanpassing in de FiO2, varieerden de frequentie en de stapgrootte van de gemaakt aanpassingen voor gelijke SpO2 waardes. Daarnaast weken de alarmlimieten voor puls oximetrie
af van het protocol in een derde van de tijd. De motivaties om af te wijken van het protocol
werden slechts enkele keren vermeld in de status van de patint.
De (ongeregistreerde) variatie in handelingen (zoals de aanpassingen in FiO2 en in alarm
limieten) zou geminimaliseerd moeten worden omdat de lange termijn uitkomst van
patinten hierdoor benvloed zou kunnen worden zonder dat daar een duidelijke reden
voor te vinden is. Aanbevelingen voor verbetering richten zich o.a. op suggesties voor het
ontwikkelen en aanpassen van protocollen en scholing van het ICN personeel. Technische
innovaties zouden gefocust moeten zijn op de verbetering van de nauwkeurigheid en gebruiksvriendelijkheid van sensoren voor te vroeg geborenen, en op de mogelijkheden voor
informatieuitwisseling tussen de gebruikte medische apparatuur.
Deel II Intraveneuze therapie

Met intraveneuze (IV) therapie wordt het rechtstreeks toedienen van voeding, medicatie
en/of vloeistoffen in een bloedvat van een patint bedoeld. Omdat de mogelijkheden voor
het aanprikken van een bloedvat bij te vroeg geborenen beperkt is, wordt er vaak gebruik
gemaakt van multi-infusie. Bij multi-infusie worden meerdere infusen aangesloten op
n katheter. Hoewel het vaak verwacht wordt dat de IV vloeistoffen met een constante,
voorgeprogrammeerde, snelheid aan de patint toegediend worden, is gebleken dat het
daadwerkelijk toegediende volume varieert. Vooral bij te vroeg geborenen kunnen deze
plotselinge veranderingen in het toegediende volume ernstige gevolgen hebben.
Om te bepalen welke factoren invloed hebben op de variatie in de stroomsnelheid van IV
vloeistoffen werd er een literatuur onderzoek gedaan (Hoofdstuk 8). Omdat de introductie
van kleppen in het IV-toedieningssysteem geassocieerd werd met vertraagde toediening van
medicatie werd er vervolgens een in-vitro studie uitgevoerd. In deze studie werd het effect
van drie verschillende type kleppen op de variatie in stroomsnelheid van multi-infusie setup met een lage voorgeprogrammeerde stroomsnelheid bepaald (Hoofdstuk 9).
De belangrijkste conclusies waren dat de variatie in stroomsnelheid van IV therapie afhangt
van de voorgeprogrammeerde stroomsnelheid, de veranderingen in de hydrostatische druk,
de compliantie van het toedieningssysteem, en het type substantie dat is toegediend aan
de patint. De klinische relevantie van de variatie in stroomsnelheid is afhankelijk van de
toegediende substantie en de eigenschappen van de patint.
02-08-12 11:37
211
Om de toediening van IV vloeistoffen te optimaliseren wordt het medisch personeel geadviseerd om de hoogst mogelijke voorgeprogrammeerde stroomsnelheid te gebruiken,
en om alleen gebruik te maken van kleppen met een (zeer) lage openingsdruk. Daarnaast
dienen de verticale verplaatsing van infuuspompen en de interne compliantie van het gehele toedieningssysteem geminimaliseerd te worden.
Conclusie
In zowel het proces van zuurstof therapie als in intraveneuze therapie werden (ongeregistreerde) variaties in de dagelijkse praktijk geobserveerd. Het daadwerkelijke effect van
deze variaties is afhankelijk van meerdere factoren en is daarom niet goed te voorspellen.
Echter, het ligt in de verwachting dat onopgemerkte (en niet traceerbare) verschillen in
de zorg leiden tot verschillen in de lange termijn uitkomst van patinten. Ook beperken
deze variaties zeer waarschijnlijk het onderzoek naar optimale strategien. Het is daarom
goed mogelijk dat beide therapien gebaat zijn bij een meer gestandaardiseerde aanpak. Om
standaardisatie te bereiken, zouden zowel veranderingen in cultuur, inclusief scholing, als
technische verbeteringen kunnen helpen. Hoewel het effect van elke afzonderlijke (kleine)
verandering in cultuur en technologie vrijwel onmeetbaar zal zijn, is het te verwachten dat
alle kleine veranderingen tezamen de kwaliteit van de zorg op de neonatale intensive care
zullen verbeteren.
02-08-12 11:37
212
02-08-12 11:37
Dankwoord
02-08-12 11:37
214
02-08-12 11:37
Dankwoord
215
Bij de totstandkoming van dit proefschrift zijn vele mensen betrokken geweest. Ik ben iedereen dan ook erg dankbaar voor hun hulp en steun. Zonder anderen tekort te doen wil ik de
volgende mensen graag noemen.
Allereerst dank ik de kinderen en hun ouders die belangeloos meegewerkt hebben aan de
onderzoeken. Ondanks de zware periode die jullie beleefden in het ziekenhuis gaven jullie
mij toestemming om mijn onderzoek uit te voeren.
Mijn promotor prof. dr. J. Dankelman, beste Jenny, na mijn afstuderen heb je me de mogelijkheid gegeven om mijn promotieonderzoek bij jou uit te voeren. Het onderzoek verliep
niet altijd even soepel, maar na een overleg met jou ging ik altijd met een positief gevoel weer
aan de slag. Je hebt het vermogen om razendsnel de vinger op een zwakke plek te leggen, en
om vervolgens met dezelfde snelheid een mogelijke oplossing te suggereren. Ik dank je voor
alle kansen die ik gekregen heb om mezelf verder te ontwikkelen.
Mijn tweede promotor prof. dr. H.J. Simonsz, Huib, nog nooit ben ik zon gedreven persoon als jij tegengekomen. Je hebt een grenzeloos vertrouwen in wetenschap en techniek, en
werkt zeer gepassioneerd. Ik ben je dankbaar voor het initiren van het onderzoek, voor het
vinden van financiering, voor je vasthoudendheid en voor het vertrouwen dat je in me had.
Mijn copromotor dr. B.J. Smit, Bert, ik heb het gevoel dat wij gezamenlijk een brug over de
bekende kloof tussen artsen en ingenieurs hebben gebouwd. Onze bijeenkomsten leverden
dikwijls verrassende nieuwe inzichten op doordat we open stonden voor elkaars vakgebieden. Ik ben je dankbaar voor je onvoorwaardelijke steun, je integriteit en je betrokkenheid.
Beste overige leden van de commissie, prof. dr. ir. C.A. Grimbergen, prof. dr. S. Bambang
Oetomo, prof. dr. F. van Bel, prof. dr. ir. R.M. Verdaasdonk en prof. dr. ir. R.H.M. Goossens,
bedankt voor het lezen van mijn proefschrift en voor het plaatsnemen in mijn promotiecommissie.
Dit proefschrift was nooit geworden wat het nu is zonder de inzet van de medeauteurs:
Paul Eilers, bedankt voor je inspirerende hulp bij de statistische vraagstukken.
Arjan van der Plas en Carol van der Palen, jullie hulp bij het opzetten van het onderzoek
over intraveneuze therapie was erg waardevol.
Roland van Rens, je hebt me aangestoken met je onuitputtelijke enthousiasme voor alles wat
met intraveneuze therapie te maken heeft. Ik wens je alle goeds voor je eigen onderzoek.
02-08-12 11:37
216
Sandra Horsch, je bent medeauteur van n artikel, maar op de achtergrond heb je veel meer
voor mij betekend. Je nodigde me zelfs tijdens je verlof uit om bij jou thuis mijn artikelen te
bespreken. Bedankt voor alles!
Sander Schutte, je begon als afstudeerbegeleider en eindigde als vriend. We hebben mooie
tijden beleefd, zowel achter onze laptops als in de kroeg. Ik wens jou en Jasper succes met
jullie eigen bedrijf!
Kirsten Henken, je stond altijd klaar om me te helpen. Zo was je zelfs bereid om midden in de
nacht enqutes uit te delen aan de verpleegkundigen! Bedankt voor je al je support.
Denise Rook, vanaf het moment dat je me leerde dat ROS ook iets anders betekent dan
paard, zijn beide onderwerpen vrijwel dagelijks onderwerp van ons gesprek op SK2210,
hotelkamers, of via internet. Jij leefde intens mee met mijn onderzoek, en was altijd beschikbaar als ik je weer eens nodig had. Heel erg bedankt!
Lieve (oud-)SK2210-ers, vele uren hebben wij gezamenlijk doorgebracht in het onderzoekershok. Bedankt voor het delen van jullie bureau, jullie (medische) kennis, en jullie gezelligheid.
Hester, jouw accurate manier van onderzoek doen, en de manier waarop je met patinten
omgaat zijn een grote inspiratiebron voor me geweest. Lisha, dankzij jouw open mind verdwenen problemen als sneeuw voor de zon. Margriet, het was heerlijk om mijn typische
ingenieurs-problemen met jou te kunnen bespreken. Bedankt voor alle steun! Tom, je bent
inmiddels vaker in het Sophia aanwezig dan ik. Je bent een waardig opvolger, en ik heb er alle
vertrouwen in dat je je promotieonderzoek tot een goed einde zal brengen.
Alle medewerkers van de IC Neonatologie, bedankt voor de medewerking bij het uitvoeren
van de verschillende studies. Ik ben zeer dankbaar voor het feit dat jullie mij de mogelijkheid
hebben gegeven om cameras op te hangen op de ICs. Hans van Goudoever, Ren Kornelisse,
Irwin Reiss en Yvonne Kant, bedankt voor alle mogelijkheden die ik van jullie heb gekregen.
Leden van het FMEA-team, bedankt voor jullie enthousiaste inzet. Coby de Boer, bedankt
voor je hulp bij het opzetten van de enqute, en voor het doorlezen van mijn stukken.
De mannen van de medische technologie: Ab, Arjan, Arthur, Cees, Fons, Johan, Leo, Marco,
Ricardo, en in het bijzonder Arie & Arie: bedankt voor de talloze keren dat ik jullie hulp heb
gekregen bij het ontwikkelen van mijn onderzoeksopstellingen. Telkens als ik het niet meer
wist, vonden (of maakten!) jullie weer een geschikt stekkertje, snoertje of klemmetje.
Annelies, Ferry, Ineke, Jolanda en Wietse ik ben zeer dankbaar voor jullie gastvrijheid en
oprechte interesse. Ik waardeer jullie voortdurende toewijding om de veiligheid van intraveneuze therapie en overige medische instrumenten te vergroten.
02-08-12 11:37
Dankwoord
217
Dear employees of Philips Medical Systems in Beblingen, thank you for the interesting
visits to your company, and the open minded discussions.
Mijn collegas op de TU, bedankt voor de alles-kan-mentaliteit. De spontane filosofische
discussies bij het koffiezetapparaat, de interessante vakgroepbijeenkomsten en de verscheidene borrels hebben ervoor gezorgd dat ik met veel plezier naar de TU ging. Anouk,
Dineke en Diones, bedankt voor jullie ondersteuning. Hoewel ik helaas niet altijd aanwezig
kon zijn, heb ik erg genoten van de taarten in de kantoortuin, het schaatsen, het zeilen en de
lunches in de Delftse Hout.
Ik dank ook mijn collegas in het LUMC voor de manier waarop zij mij verwelkomd hebben
in mijn nieuwe baan, voor de fijne samenwerking en voor de steun tijdens de laatste loodjes
van mijn promotietraject.
Thom, onze koffiepauzes duurden niet zelden langer dan (door mij) gepland. Dat is ook niet
zo gek de wereldproblematiek heb je niet zomaar in een kwartiertje opgelost, laat staan
de dagelijkse dieptepunten van een onderzoeker. Maar gelukkig zijn kroegen tot diep in de
nacht open, en gaat het programmeren in Matlab beter met een wijntje erbij ;-) Bedankt dat
je opeens bij mijn bureau stond en niet meer weg bent gegaan!
Lieve clubgenootjes, jullie steun en onvoorwaardelijke vriendschap zijn geweldig! Het
clubeten, de vrimibos, de clubweekenden, de zeer frequente clubmailtjes, en alle andere activiteiten waren een welkome afwisseling op mijn promotieonderzoek.
(Oud-)StuD-ers, bedankt voor de top-tijden die we beleefd hebben op de borrels, de feesten,
en de extravagante superdeluxe StuD-weekenden. Op naar het derde lustrum!
Familie Hoogweg en familie Lems, bedankt voor jullie gastvrijheid. De rust en ruimte op
jullie erf maakten het mogelijk dat ik dagelijks samen met Wox mijn hoofd leeg kon maken.
Dankbaar ben ik ook alle (oud-)stalgenootjes. Bedankt voor de goede zorgen voor Wox, de
heerlijke buitenritten, en de eet/knip-avonden. Esther & Gerco, Marco & Mariska enorm
bedankt voor jullie vriendschap. Jullie openhartige gesprekken, adviezen, schitterende fotos
en groom-sessies zijn erg waardevol! Christa en Mirtro, bedankt voor jullie verhelderende
lessen en coaching. Louise, bedankt voor je betrokkenheid en hulp bij de laatste fase van het
proefschrift.
Ankie en Fred, bedankt voor jullie hulp bij de organisatie van de borrel.
02-08-12 11:37
218
Lieve Evert en Quita bedankt voor jullie interesse en betrokkenheid bij mijn onderzoek, de
(meerdaagse) uitjes, en de ontelbare keren dat Joost en ik spontaan konden aanschuiven voor
een heerlijk diner.
Lieve papa en mama, jullie waren een vaste steun en toeverlaat tijdens de afgelopen jaren.
Bij hoogte-en dieptepunten pakte ik direct de telefoon om jullie bij te praten. Bedankt voor
alle kansen en liefde die ik van jullie heb gekregen.
Lotte, ik ben bevoorrecht met jou als lieve kleine zus. Jij adviseerde me als ik weer eens
gillend een paskamer uitrende, en was altijd in voor een droog wit wijntje in de zon.
Mijn broertjes, wie zouden er op dit moment beter naast me kunnen staan om de brassende
menigte uit elkaar te houden dan jullie: Jelle de bijna arts en Hidde de bijna ingenieur.
Vergeten jullie niet de stellingen ff door te lezen?!
Joost, ik ben je ongelooflijk dankbaar voor alles wat je voor mij hebt gedaan. Bedankt voor je
eindeloze geduld, je zorgzaamheid, je zwijgende (maar veelzeggende) blik, je meedenken, je
brede schouders, en voor je onvoorwaardelijke liefde.
Anne
02-08-12 11:37
Dankwoord
219
02-08-12 11:37
220
02-08-12 11:37
About the author
02-08-12 11:37
222
02-08-12 11:37
About the author
223
Anne Catherine van der Eijk was born in Zwolle, the Netherlands on the 1st of August, 1981.
After completing secondary school at the Gymnasium Celeanum in Zwolle, she started
studying Industrial Design Engineering at the Delft University of Technology in September
1999. Three years later she interrupted her study for one year to become a full time member
of the board of StuD Studentenuitzendbureau in Delft.
In November 2004 Anne obtained her Bachelor of Science in Industrial Design Engineering and started with the master study Biomedical Engineering at the Faculty of Mechanical, Maritime and Materials Engineering in Delft. After finishing her master thesis under
supervision of Prof. dr. Jenny Dankelman she continued her graduation project as a PhD
student in 2007. The PhD research was performed in cooperation with the Erasmus Medical
Centre - Sophia Childrens hospital, Department of Neonatology, Rotterdam.
From April 2010 the PhD project was performed on a part-time basis to enable Anne to
work as a consultant for medical-technical issues in the Erasmus MC. Since October 2011,
she enjoys working in the Leiden University Medical Centre where she is concerned with
the quality, safety, and maintenance of medical equipment.
Anne is happily married with Joost Kroes, they live in Delft.
02-08-12 11:37
02-08-12 11:37

Proefschrift Definitief Lage Resolutie

Enviado por

Dados do documento

Direitos autorais

Formatos disponíveis

Compartilhar este documento

Compartilhar ou incorporar documento

Opções de compartilhamento

Você considera este documento útil?

Este conteúdo é inapropriado?

Direitos autorais:

Formatos disponíveis

Proefschrift Definitief Lage Resolutie

Enviado por

Direitos autorais:

Formatos disponíveis

On-ward observations

in neonatal intensive care:

Dit proefschrift is goedgekeurd door de promotoren:

The research presented in this thesis was partially supported by

Lay-out Louise de Kruijf

Part I Supplemental oxygen therapy

10.1 Main conclusions

About the Author

To be able to meet the specific needs of preterm

1.2 Preterm infants, definitions & prevalence

1.3 Prematurity, causes & outcome

1.4 Prematurity & oxidative stress related diseases

1.4.2 Infant respiratory distress syndrome

1.5 Supplemental oxygen therapy

1.5.1 A brief overview of history of supplemental oxygen therapy

Syringe pumps with

1.6.1 A brief overview of history of IV therapy

1.7 Problem statements

Manual control [%]

Dedicated control [%]

Urschitz106 Open loop

Urschitz106 Closed loop

Morozoff State machine

Morozoff123 Adaptive model

Morozoff123 Closed loop, PID

1.9 Thesis outline

14. Thompson J, Irgens LM, Rasmussen S, Daltveit AK.

42. Koch J, Hensley G, Roy L, Brown S, Ramaciotti C,

94. Klem S, Farrington J, Leff R. Influence of infusion

106. Urschitz MS, Horn W, Seyfang A, Hallenberger A,

117. Yu C, He WG, So JM, Roy R, Kaufman H, Newell

127. Azhar N, Karim U. Automatic feedback control of

Oxygenation in preterm infants

To prevent the negative effects of supplemental

2.2 Oxygen in the human body

= haemoglobin concentration (grams litre1)

Q = heart rate stroke volume

Oxygenation in preterm infants

2.2.1 The oxygen dissociation curve

Oxygen saturation (SO2) [%]

right shifted by:

Partial pressure of oxygen (PO2) [mmHg]

Oxygenation in preterm infants

2.3 Monitoring of oxygenation

2.3.1 Physical assessment

Oxygenation in preterm infants

Oxygenation in preterm infants

Oxygenation in preterm infants

2.4 Reference values for blood oxygen levels

Oxygenation in preterm infants

2.5 Monitoring oxygenation in preterm infants:

Oxygenation in preterm infants

42. Pollitzer MJ, Soutter LP, Reynolds EO. Continuous

Oxygenation in preterm infants

56. Saugstad OD. Is oxygen more toxic than currently

71. Chambrin MC. Alarms in the intensive care unit:

83. Naulaers G, Morren G, Van Huffel S, Casaer P,

96. Singh SA, Singhal N. Does end-tidal carbon dioxide

Oxygenation in preterm infants

109. Weinberger B, Laskin DL, Heck DE, Laskin JD.

122. Tin W, Milligan DWA, Pennefather P, Hey E. Pulse

New-generation pulse oximeters in ELBW infants

Pulse oximetry is a widely used

Softcare SCPR <1.5kg

14. Thompson J, Irgens LM, Rasmussen S, Daltveit AK.

42. Koch J, Hensley G, Roy L, Brown S, Ramaciotti C,

94. Klem S, Farrington J, Leff R. Influence of infusion

106. Urschitz MS, Horn W, Seyfang A, Hallenberger A,

117. Yu C, He WG, So JM, Roy R, Kaufman H, Newell

127. Azhar N, Karim U. Automatic feedback control of

2.5 Monitoring oxygenation in preterm infants:

42. Pollitzer MJ, Soutter LP, Reynolds EO. Continuous

56. Saugstad OD. Is oxygen more toxic than currently

71. Chambrin MC. Alarms in the intensive care unit:

83. Naulaers G, Morren G, Van Huffel S, Casaer P,

96. Singh SA, Singhal N. Does end-tidal carbon dioxide

109. Weinberger B, Laskin DL, Heck DE, Laskin JD.

122. Tin W, Milligan DWA, Pennefather P, Hey E. Pulse

14. Saugstad O. Bronchopulmonary dysplasia-oxidative

28. Fellman V, Hellstrm-Westas L, Norman M, Westgren

41. Elliott M, Tate R, Page K. Do clinicians know how to

12. Morozoff EP, Smyth JA. Evaluation of three automatic