Detection of Bloodstream Infections and Prediction of Bronchopulmonary

Dysplasia in Preterm Neonates with an Electronic Nose

Tobias Rogosch, PhD
1
, Nina Herrmann
1
, Rolf F. Maier, MD
1
, Eugen Domann, PhD
2
, Akira Hattesohl, MS
3
,
Andreas Rembert Koczulla, MD
3
, and Michael Zemlin, MD
1
We show that smellprints of volatile organic components measured with an electronic nose (Cyranose 320; Smiths
Detection Group Ltd, Watford, United Kingdom) differ between tracheal aspirates from preterm neonates with or
without laboratory-conrmed bloodstream infections and with or without subsequent development of bronchopul-
monary dysplasia. Tracheal aspirate smellprints could be useful noninvasive diagnostic markers for preterm neo-
nates. (J Pediatr 2014;165:622-4).
A
cute infections and bronchopulmonary dysplasia
(BPD) are major causes of mortality and morbidity
in preterm neonates.
1
Measurements of inammatory
indicators and gene expression in tracheal aspirates (TAs) have
been used to identify preterm neonates with infections and/or
to predict the risk of BPD.
2-5
Exhaled breath condensate also
can be collected from ventilated infants for diagnostic pur-
poses.
6
Because of the technical difculties of sample collection
of exhaled breath condensates and/or their analysis, however,
they are not used routinely for diagnosis in preterm neo-
nates.
7,8
Novel biomarkers for the early diagnosis of inamma-
tory diseases are of interest for neonatologists.
9
We present a
novel rapid technique of measuring volatile organic compo-
nents directly in TA uids from preterm neonates.
Methods
Inclusion criteria were gestational age <37 weeks, need for
articial ventilation, and signed informed consent from par-
ents or caretakers, as approved by the local Ethics Committee
(Table). TAs were collected during routine suctioning after
instillation of 0.5 mL of sterile Ringers acetate. The suction
catheter was cleared of retained secretions by aspiration of
an additional 0.5 mL of Ringers acetate. One-half of each
sample was sent to microbiologic analysis, and the other
one-half was stored at 20

C for smellprint analysis.

Bacteria in TAs were identied by traditional culture and
by molecular techniques. Bacterial 16S ribosomal DNA was
amplied by polymerase chain reaction, and the products
were analyzed by the use of denaturing high-performance
liquid chromatography on a WAVE System (Transgenomic,
Omaha, Nebraska).
10,11
The denition of infection was adapted for neonates
12
from the Centers for Disease Control and Prevention deni-
tion of laboratory-conrmed bloodstream infection, which
requires coinciding laboratory ndings (positive blood cul-
ture, C-reactive protein >1 mg/dL, immature neutrophils:to-
tal neutrophils ratio >0.2, or leucocytopenia <4.000/mL) and
clinical criteria.
13
BPD was diagnosed according to the
criteria of the National Institutes of Health when the infant
required >21% oxygen for more than 28 days.
14
Smellprints of volatile organic compounds were measured
withthe Cyranose 320 detector according tothe manufacturers
instructions (Smiths Detection Group Ltd, Watford, United
Kingdom) and published protocols.
15-18
Then, 250 mL of TA
was heated to 37

C. Ambient air was used as a reference for a

10-second baseline. The snout of the electronic nose was then
placed above the surface of the TA for 10 seconds.
Data Analyses
Statistical analyses were performed with R 3.0.2 (http://www.
r-project.org) and GraphPad Prism 5.0 (GraphPad Software
Inc, La Jolla, California). Smellprint data were preprocessed
by centering and normalizing, followed by the classier linear
discriminant analysis, as described previously.
19
The linear
discriminant values were compared in a paired Student
t test or a Mann-Whitney U test as appropriate. A Mahalano-
bis distance between the groups greater than 1.96 was
considered signicant (condence level >95%, P < .05).
The cross-validation value (CVV) was calculated by perform-
ing a k-fold cross-validation (k = product of group sizes)
with the use of one data point of each group as test data.
Results
We collected 38 TAs from 28 intubated preterm neonates.
Smellprints of TAs of neonates with laboratory-conrmed
bloodstream infections (n = 8) were different from those of
From the
1
Department of Pediatrics, Philipps-University Marburg, Marburg,
Germany;
2
Institute of Medical Microbiology, Justus-Liebig-University, Giessen,
Germany; and
3
Division for Pulmonary Diseases, Department of Internal Medicine,
Philipps-University Marburg, Marburg, Germany
Supported by the Universities of Giessen and Marburg Lung Center (to M.Z. and
E.D.) and the Deutsche Forschungsgemeinschaft, Transregio 22 (TP A17 to M.Z.).
The authors declare no conicts of interest.
0022-3476/$ - see front matter. Copyright 2014 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jpeds.2014.04.049
BPD Bronchopulmonary dysplasia
CVV Cross-validation value
TA Tracheal aspirate
622
neonates without infection regardless of a positive (n = 14, P
< .0001, t test; CVV 62.5%, Mahalanobis distance 2.39) or
negative microbiologic analysis (n = 16, P < .0001, t test;
CVV 70.0%, Mahalanobis distance 2.55), whereas the latter
2 groups did not differ signicantly (P = .11, t test) (Figure).
BPD could only be diagnosed in the 23 neonates born <32
weeks of gestation.
14
Only the rst sample of these patients
was included in this analysis. Twelve of 23 neonates devel-
oped BPD, and their smellprints differed from those of neo-
nates without BPD by t test (P = .0017), but the Mahalanobis
distance was below the signicance threshold of 1.96 (Maha-
lanobis distance 1.91, CVV37.1%). The storage of 10 samples
for 7 days before reanalysis did not alter smellprints, conrm-
ing repeatability and reliability of the method (data not
shown), in agreement with published observations in exhaled
breath condensates.
20
Discussion
Smellprints of volatile organic compounds from TAs can
discriminate between preterm neonates with or without
laboratory-conrmed bloodstream infection. The observed
differences between smellprints of TAs from preterm neonates
who later developed BPD and those that did not indicate a
trend, which is encouraging further research. These results
are in agreement with numerous studies that described the
analysis of TAs or exhaled breath condensate by gas chroma-
tography and mass spectrometry,
21
bacterial culture,
3
gene
expression analysis,
4
or enzyme-linked immunosorbent assay
5
for the identication of inammation in preterm neonates.
TAs are easily accessible in patients undergoing mechanical
ventilation because secretions are removed routinely for clin-
ical reasons. Although neonates who developed BPD and neo-
nates with laboratory-conrmed bloodstream infection were
less mature (P < .01, respectively), samples of uninfected
immature neonates still clustered into the no infection group,
making a maturity related clustering of the smellprints un-
likely. However, larger studies are required to test the reliability
of TAsmellprints for the prediction of BPDand to rule out po-
tential biases causedby gestational age, sex, and other variables.
Smellprints of serial samples from 2 patients taken before and
shortly after the onset of laboratory-conrmed bloodstream
infection, of 3 uninfected patients who acquired colonization
of TA, and of one patient with unchanged conditions fell
Table. Patient characteristics
No. total
samples
Intubation
at age in days
(median, range)
Sample at
day after intubation
(median, range)
Gestational age
(median, range),
weeks + days
CRP (median,
quartiles), mg/L
I:T (median,
quartiles)
Invasive infection 8 1 (1-1) 2.5 (0-23) 25 + 2 (23 + 0, 28 + 1) 20 (<5, 32) 0.226 (0.168, 0.263)
No invasive infection but germs 14 1 (1-5) 1 (0-55) 28 + 3 (24 + 1, 36 + 5) <5 (<5, <5) 0.083 (0.072, 0.095)
No invasive infection, no germs 16 1 (1-2) 0 (0-23) 29 + 3 (25 + 1, 36 + 4) <5 (<5, 9) 0.095 (0.048, 0.125)
BPD 12 1 (1-1) 1 (0-7) 25 + 3 (24 + 1, 27 + 4) <5 (<5, 7) 0.134 (0.089, 0.188)
No BPD 11 1 (1-2) 0 (0-1) 29 + 2 (26 + 3, 31 + 2) <5 (<5, <5) 0.077 (0.061, 0.125)
CRP, C-reactive protein; I:T, immature neutrophils/total neutrophils ratio.
Figure. The plots of the linear discriminant analysis of smellprints measured with the electronic nose show that patterns of A,
samplesfromneonates withinvasiveinfectionswereseparablefromthosewithout andB, neonates withlater development of BPD
were separable fromneonates without. Patient numbers in the symbols are white in serial samples taken before and shortly after
onset of laboratory-conrmed bloodstreaminfection. inf+, smellprints of TAs of neonates with laboratory-conrmed bloodstream
infections; germ+ inf, neonates without infection but a positive microbiologic analysis in TA; germ inf, neonates without
infection and a negative microbiologic analysis in TA.
Vol. 165, No. 3 September 2014
623
into the appropriate no infection or infection clusters,
respectively. Thus, potential interindividual differences did
not interfere with the diagnostic accuracy.
This measurement yields blinded smellprints
22
and does
not allow the identication of individual substances.
23
Future
studies in which investigators use other techniques should
clarify whether individual volatile organic compounds can
be identied to monitor inammation and if exhaled breath
condensate and TAs yield comparable smell prints. Necro-
tizing enterocolitis and other diseases also would be inter-
esting candidates for studies of volatile organic compounds
in TA or other biosamples.
In conclusion, we present a novel measurement that com-
bines the relatively easy collection of TAs (compared with the
collection of exhaled breath condensate) with the relatively
simple method of measuring smellprints using a handheld
electronic nose. Theoretically, this technique can deliver re-
sults within minutes. Larger studies should be performed
because TA smellprints might be useful as a novel diagnostic
marker for preterm neonates. n
We gratefully thank Regina Stoehr, Sabine Jennemann, Ursula Boas,
and Silke Zechel-Gran for their excellent technical assistance.
Submitted for publication Nov 15, 2013; last revision received Mar 21, 2014;
accepted Apr 28, 2014.
Reprint requests: Michael Zemlin, MD, Department of Pediatrics, Philipps
University Marburg, Baldinger Str., D-35033 Marburg, Germany. E-mail:
zemlin@med.uni-marburg.de
References
1. Paananen R, Husa AK, Vuolteenaho R, Herva R, Kaukola T, Hallman M.
Blood cytokines during the perinatal period in very preterminfants: rela-
tionship of inammatory response and bronchopulmonary dysplasia. J
Pediatr 2009;154:39-43.e3.
2. Speer CP. Inammation and bronchopulmonary dysplasia. Semin Neo-
natol 2003;8:29-38.
3. Harwood DT, Darlow BA, Cheah FC, McNeill N, Graham P,
Winterbourn CC. Biomarkers of neutrophil-mediated glutathione and
protein oxidation in tracheal aspirates from preterm infants: association
with bacterial infection. Pediatr Res 2011;69:28-33.
4. Hikino S, Ohga S, Kinjo T, Kusuda T, Ochiai M, Inoue H, et al. Tracheal
aspirate gene expression in pretermnewborns and development of bron-
chopulmonary dysplasia. Pediatr Int 2012;54:208-14.
5. Hasan J, Beharry KD, Valencia AM, Strauss A, Modanlou HD. Soluble
vascular endothelial growth factor receptor 1 in tracheal aspirate uid
of preterm neonates at birth may be predictive of bronchopulmonary
dysplasia/chronic lung disease. Pediatrics 2009;123:1541-7.
6. Muller WG, Morini F, Eaton S, Peters M, Jaffe A. Safety and feasibility of
exhaled breath condensate collection in ventilated infants and children.
Eur Respir J 2006;28:479-85.
7. Rosias PP, Dompeling E, Hendriks HJ, Heijnens JW,
Donckerwolcke RA, Jobsis Q. Exhaled breath condensate in children:
pearls and pitfalls. Pediatr Allergy Immunol 2004;15:4-19.
8. Harrison CM, Andersen CC. Exhaled breath measures of inammation:
are they useful in neonatal chronic lung disease? Arch Dis Child Fetal
Neonatal Ed 2005;90:F6-10.
9. Mukhopadhyay S, Puopolo KM. Risk assessment in neonatal early onset
sepsis. Semin Perinatol 2012;36:408-15.
10. Domann E, Hong G, Imirzalioglu C, Turschner S, Kuhle J, Watzel C,
et al. Culture-independent identication of pathogenic bacteria and
polymicrobial infections in the genitourinary tract of renal transplant re-
cipients. J Clin Microbiol 2003;41:5500-10.
11. Imirzalioglu C, Hain T, Chakraborty T, Domann E. Hidden pathogens
uncovered: metagenomic analysis of urinary tract infections. Andrologia
2008;40:66-71.
12. Obladen M. Infektionen. In: Maier RF, Obladen M, eds. Neugeborene-
nintensivmedizin: Evidenz und Erfahrung. 8th ed. Berlin: Springer;
2011. p. 495.
13. Garner JS, Jarvis WR, Emori TG, Horan TC, Hughes JM. CDC deni-
tions for nosocomial infections, 1988. Am J Infect Control 1988;16:
128-40.
14. Ehrenkranz RA, Walsh MC, Vohr BR, Jobe AH, Wright LL, Fanaroff AA,
et al. Validation of the National Institutes of Health consensus denition
of bronchopulmonary dysplasia. Pediatrics 2005;116:1353-60.
15. Hattesohl AD, Jorres RA, Dressel H, Schmid S, Vogelmeier C,
Greulich T, et al. Discrimination between COPD patients with and
without alpha 1-antitrypsin deciency using an electronic nose. Respir-
ology 2011;16:1258-64.
16. Dragonieri S, Schot R, Mertens BJ, Le Cessie S, Gauw SA, Spanevello A,
et al. An electronic nose in the discrimination of patients with asthma
and controls. J Allergy Clin Immunol 2007;120:856-62.
17. Horvath I, Lazar Z, Gyulai N, Kollai M, Losonczy G. Exhaled biomarkers
in lung cancer. Eur Respir J 2009;34:261-75.
18. Fens N, van der Schee MP, Brinkman P, Sterk PJ. Exhaled breath analysis
by electronic nose in airways disease. Established issues and key ques-
tions. Clin Exp Allergy 2013;43:705-15.
19. Huttmann EM, Greulich T, Hattesohl A, Schmid S, Noeske S, Herr C,
et al. Comparison of two devices and two breathing patterns for exhaled
breath condensate sampling. PLoS ONE 2011;6:e27467.
20. Bofan M, Mores N, Baron M, Dabrowska M, Valente S, Schmid M,
et al. Within-day and between-day repeatability of measurements
with an electronic nose in patients with COPD. J Breath Res 2013;7:
017103.
21. Moser B, Bodrogi F, Eibl G, Lechner M, Rieder J, Lirk P. Mass spectro-
metric prole of exhaled breatheld study by PTR-MS. Respir Physiol
Neurobiol 2005;145:295-300.
22. Lewis NS. Comparisons between mammalian and articial olfaction
based on arrays of carbon black-polymer composite vapor detectors.
Acc Chem Res 2004;37:663-72.
23. Thaler ER, Hanson CW. Medical applications of electronic nose technol-
ogy. Expert Rev Med Devices 2005;2:559-66.
THE JOURNAL OF PEDIATRICS www.jpeds.com Vol. 165, No. 3
624 Rogosch et al