Discontinuation of Epidural Analgesia Late in Labour For Reducing The Adverse Delivery Outcomes Associated With Epidural 1

Cochrane Database of Systematic Reviews
Discontinuation of epidural analgesia late in labour for

reducing the adverse delivery outcomes associated with
epidural analgesia (Review)
Torvaldsen S, Roberts CL, Bell JC, Raynes-Greenow CH
Torvaldsen S, Roberts CL, Bell JC, Raynes-Greenow CH.

Discontinuation of epidural analgesia late in labour for reducing the adverse delivery outcomes associated with epidural analgesia.
Cochrane Database of Systematic Reviews 2004, Issue 4. Art. No.: CD004457.
DOI: 10.1002/14651858.CD004457.pub2.
www.cochranelibrary.com
Discontinuation of epidural analgesia late in labour for reducing the adverse delivery outcomes associated with epidural analgesia
(Review)
Copyright © 2010 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
TABLE OF CONTENTS
HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
AUTHORS’ CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
Analysis 1.1. Comparison 1 Discontinued versus continued, Outcome 1 Instrumental delivery. . . . . . . . . 13
Analysis 1.2. Comparison 1 Discontinued versus continued, Outcome 2 Instrumental delivery by subgroup analysis. 14
Analysis 1.3. Comparison 1 Discontinued versus continued, Outcome 3 Caesarean section. . . . . . . . . . 15
Analysis 1.4. Comparison 1 Discontinued versus continued, Outcome 4 Caesarean section by subgroup analysis. . . 16
Analysis 1.5. Comparison 1 Discontinued versus continued, Outcome 5 Spontaneous vaginal delivery. . . . . . 17
Analysis 1.6. Comparison 1 Discontinued versus continued, Outcome 6 Spontaneous vaginal delivery by subgroup
analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
Analysis 1.7. Comparison 1 Discontinued versus continued, Outcome 7 Duration of 2nd stage. . . . . . . . 19
Analysis 1.8. Comparison 1 Discontinued versus continued, Outcome 8 Duration of 2nd stage by subgroup analysis. 20
Analysis 1.9. Comparison 1 Discontinued versus continued, Outcome 9 Fetal malposition. . . . . . . . . . 21
Analysis 1.10. Comparison 1 Discontinued versus continued, Outcome 10 Fetal malposition by subgroup analysis. . 22
Analysis 1.11. Comparison 1 Discontinued versus continued, Outcome 11 Inadequate pain relief. . . . . . . . 23
Analysis 1.12. Comparison 1 Discontinued versus continued, Outcome 12 Inadequate pain relief by subgroup analysis. 24
Analysis 1.13. Comparison 1 Discontinued versus continued, Outcome 13 Apgar score < 7 or < 8 at 1 minute. . . 25
Analysis 1.14. Comparison 1 Discontinued versus continued, Outcome 14 Apgar score < 7 or < 8 at 1 minute by subgroup
analysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
Analysis 1.15. Comparison 1 Discontinued versus continued, Outcome 15 Apgar score < 7 at 5 minutes. . . . . 27
Analysis 1.16. Comparison 1 Discontinued versus continued, Outcome 16 Umbilical arterial pH. . . . . . . . 27
Analysis 1.17. Comparison 1 Discontinued versus continued, Outcome 17 Umbilical arterial pH by subgroup analysis. 28
WHAT’S NEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
Discontinuation of epidural analgesia late in labour for reducing the adverse delivery outcomes associated with epidural analgesia i
(Review)
[Intervention Review]
Discontinuation of epidural analgesia late in labour for

reducing the adverse delivery outcomes associated with
epidural analgesia
Siranda Torvaldsen1 , Christine L Roberts2 , Jane C Bell3 , Camille H Raynes-Greenow4

1
Public Health Officer Training Program, School of Public Health and Community Medicine, University of New South Wales, Sydney,
Australia. 2 Kolling Institute for Medical Research, Northern Clinical School, E25 - Royal North Shore Hospital, The University of
Sydney, Sydney, Australia. 3 Centre for Perinatal Health Services Research, QEII Research Institute, Sydney, Australia. 4 Sydney School
of Public Health, University of Sydney, Syndney, Australia
Contact address: Siranda Torvaldsen, Public Health Officer Training Program, School of Public Health and Community Medicine, Uni-
versity of New South Wales, Room 317, Level 3, Samuels Building, Sydney, New South Wales, 2052, Australia. siranda@unsw.edu.au.
Editorial group: Cochrane Pregnancy and Childbirth Group.

Publication status and date: Edited (no change to conclusions), published in Issue 9, 2010.
Review content assessed as up-to-date: 28 October 2007.
Citation: Torvaldsen S, Roberts CL, Bell JC, Raynes-Greenow CH. Discontinuation of epidural analgesia late in labour for reducing
the adverse delivery outcomes associated with epidural analgesia. Cochrane Database of Systematic Reviews 2004, Issue 4. Art. No.:
CD004457. DOI: 10.1002/14651858.CD004457.pub2.
ABSTRACT
Background
Although epidural analgesia provides the most effective labour analgesia, it is associated with some adverse obstetric consequences,
including an increased risk of instrumental delivery. Many centres discontinue epidural analgesia late in labour to improve a woman’s
ability to push and reduce the rate of instrumental delivery.
Objectives
To assess the impact of discontinuing epidural analgesia late in labour on:
i) rates of instrumental deliveries and other delivery outcomes; and
ii) analgesia and satisfaction with labour care.
Search methods
We searched the Cochrane Pregnancy and Childbirth Group’s Trials Register (October 2007).
Selection criteria
Randomised controlled trials of epidurals discontinued late in labour compared with continuation of the same epidural protocol until
birth, in women who receive an epidural for analgesia in the first stage of labour.
Data collection and analysis
Two review authors independently assessed study eligibility and quality and extracted the data. We analysed categorical data using
relative risk (RR), and continuous data using weighted mean difference.
Discontinuation of epidural analgesia late in labour for reducing the adverse delivery outcomes associated with epidural analgesia 1
(Review)
Main results
We identified six studies, of which five were included (462 participants). Three of these were high-quality studies whilst the other
two were judged to be of lower quality because placebo was not used and the method of randomisation not described. All studies
used different epidural analgesia protocols (type of drug, dosage or method of administration). Overall, the reduction in instrumental
delivery rate was not statistically significant (23% versus 28%, RR 0.84, 95% confidence interval (CI) 0.61 to 1.15) nor was there
any statistically significant difference in rates of other delivery outcomes. The only statistically significant result was an increase in
inadequate pain relief when the epidural was stopped (22% versus 6%, RR 3.68, 95% CI 1.99 to 6.80).
Authors’ conclusions
There is insufficient evidence to support the hypothesis that discontinuing epidural analgesia late in labour reduces the rate of in-
strumental delivery. There is evidence that it increases the rate of inadequate pain relief in the second stage of labour. The practice of
discontinuing epidurals is widespread and the size of the reduction in instrumental delivery rate could be clinically important; therefore,
we recommend a larger study than those included in this review be undertaken to determine whether this effect is real or has occurred
by chance, and to provide stronger evidence about the safety aspects.
PLAIN LANGUAGE SUMMARY
Discontinuation of epidural analgesia late in labour for reducing the adverse delivery outcomes associated with epidural
analgesia
Not enough evidence to suggest that stopping an epidural late in labour lowers the risk of instrumental delivery or other unwanted
outcomes.
Epidurals are used for pain relief in labour, but they increase the risk of instrumental delivery (vacuum/forceps). Stopping epidurals
early aims to allow women to feel the pushing urge and so reduce the chance of having an instrumental birth and possible problems
associated with such a birth. There is not enough evidence from the five included trials, involving 462 participants, to show whether
stopping an epidural really does lower the risk of instrumental delivery or of any other unwanted outcome. The results show that
women whose epidural is stopped report more pain than women whose epidural is continued until the birth of the baby. More research
is needed.
BACKGROUND risks for mothers of vaginal/perineal trauma and anal sphinc-

ter damage, resulting in urinary incontinence, bowel and sexual
In high-income countries most women giving birth use some problems (Brown 1998; Eason 2000; Johanson 1999; MacArthur
method of pain relief during labour. Epidural analgesia pro- 1997; Sultan 1993).
vides the most effective labour analgesia and its use for labour
and childbirth has increased dramatically over the past 20 years Several mechanisms for the increase in instrumental deliveries as-
(Anim-Somuah 2005; Leighton 2002; Roberts 2002). However, a sociated with epidural analgesia have been postulated including
Cochrane Systematic Review shows that compared to other forms relaxation of the pelvic floor muscles, which may delay the rota-
of analgesia, epidurals are also associated with some adverse obstet- tion of the head, a weakened desire to push due to diminution of
ric consequences (Anim-Somuah 2005). These include a longer the bearing down reflex and reduced uterine activity (Bates 1985;
second stage of labour, increased use of oxytocin augmentation, Mayberry 1999). Discontinuing the epidural late in labour aims to
and an increased risk of instrumental delivery, urinary retention improve a woman’s ability to push during second stage and thereby
and fever. The increased risk of instrumental delivery for women reduce instrumental deliveries associated with epidural analgesia.
receiving epidural analgesia (relative risk 1.38, 95% confidence in- In surveys of delivery wards, the practice of letting the epidural
terval 1.24 to 1.53) is of particular concern (Anim-Somuah 2005). wear off has been reported as ranging from 42% to 89% (Paech
High rates of instrumental deliveries are associated with increased 1991; Page 1989; Roberts 2003; Vandendriesen 1998). A 1992
(Review)
systematic review of discontinuing epidural block at 8 cm dilata- Types of outcome measures
tion included only one trial (92 women) and concluded that al- 1. Incidence of instrumental delivery;
lowing the epidural to ’wear off ’ at the end of the first stage of 2. incidence of caesarean section;
labour caused a significant increase in pain during the second stage 3. incidence of spontaneous vaginal delivery;
but this was accompanied by a reduced incidence of instrumental 4. duration of the second stage of labour;
delivery (Howell 1995). 5. incidence of fetal malposition at delivery (occipitoposterior
Satisfaction with childbirth is not necessarily contingent upon the and occipitotransverse positions);
absence of pain (Ranta 1995). Many women are willing to experi- 6. reporting of inadequate pain relief during the second stage
ence pain in childbirth but do not want pain to overwhelm them. of labour;
7. maternal dissatisfaction with labour care;
Antenatally, 82% of women wish to see how labour progresses
and only want analgesia when pain becomes severe or intolerable 8. incidence of low Apgar score at one minute;
(Ranta 1995). Consequently, when assessing the effectiveness of 9. incidence of low Apgar score at five minutes;
labour analgesia, it may be as important to assess satisfaction with 10. incidence of admission to neonatal intensive care unit;
labour care as the level of analgesia or anaesthesia achieved. 11. incidence of low umbilical arterial pH;
12. incidence of urinary incontinence postpartum;
13. incidence of fecal incontinence postpartum;
14. incidence of sexual problems postpartum.
OBJECTIVES A priori subgroup analyses
1. By parity (first birth versus second or subsequent births);
1. To assess the effectiveness of discontinuing epidural 2. by type of epidural analgesia (intermittent bolus,
analgesia late in labour in reducing instrumental deliveries and continuous infusion or combined spinal epidural);
other adverse delivery outcomes associated with epidural 3. by type of analgesic agent (local anaesthetic alone versus
analgesia. local anaesthetic plus opioid).
2. To assess the impact of discontinuing epidural analgesia late

in labour on analgesia and satisfaction with labour care.
Search methods for identification of studies
METHODS Electronic searches

We searched the Cochrane Pregnancy and Childbirth Group’s Tri-
als Register by contacting the Trials Search Co-ordinator (October
2007).
Criteria for considering studies for this review
The Cochrane Pregnancy and Childbirth Group’s Trials Register
is maintained by the Trials Search Co-ordinator and contains trials
identified from:
Types of studies 1. quarterly searches of the Cochrane Central Register of
Randomised controlled trials. Quasi-randomised studies are not Controlled Trials (CENTRAL);
included. 2. monthly searches of MEDLINE;
3. handsearches of 30 journals and the proceedings of major
conferences;
Types of participants 4. weekly current awareness search of a further 36 journals
plus monthly BioMed Central email alerts.
Pregnant women of any age who receive an epidural for labour Details of the search strategies for CENTRAL and MEDLINE,
analgesia in the first stage of labour. Labour onset may be sponta- the list of handsearched journals and conference proceedings, and
neous or induced. the list of journals reviewed via the current awareness service can be
found in the ’Search strategies for identification of studies’ section
within the editorial information about the Cochrane Pregnancy
Types of interventions and Childbirth Group.
Any epidural (any dosage regimen, any method of administration) Trials identified through the searching activities described above
that is discontinued late in labour (at least 8 centimetre cervical are given a code (or codes) depending on the topic. The codes are
dilatation) and replaced by either placebo or no treatment com- linked to review topics. The Trials Search Co-ordinator searches
pared with continuation of the same epidural protocol until birth. the register for each review using these codes rather than keywords.
(Review)
We did not apply any language restrictions. except Johnsrud 1988 specified that the women were healthy or
had uncomplicated pregnancies, or both. All women had epidural
analgesia initiated in the first stage of labour. The epidural infu-
Data collection and analysis sion or top-ups were then either stopped or replaced with a study
solution (either the same solution or saline) late in labour. Four
Two review authors independently assessed trials for inclusion in studies specified that this occurred either at at least eight centime-
the review. We assessed the methodological quality according to tres or more than eight centimetres whilst Johnsrud 1988 stated
the criteria in the Cochrane Reviewers’ Handbook (Clarke 2003), that the study group were without infusion analgesia in the second
with a grade allocated to each trial on the basis of allocation con- stage but did not specify exactly when it was switched off.
cealment: A (adequate), B (unclear), C (clearly inadequate). For In three studies (Chestnut 1987a; Chestnut 1987b; Chestnut
all trials we documented details regarding randomisation method, 1990), the continuous epidural infusion solution syringe was sub-
completeness of follow up, use of placebo, blinding of outcome stituted with a syringe containing the study solution when the
measurement and analysis in randomised groups. woman’s cervix was eight or more centimetres dilated. These stud-
Two review authors independently extracted data using prepared ies were methodologically almost identical, all involving replacing
data extraction forms. We performed statistical analyses using the the continuous epidural infusion solution with a study solution
Review Manager software (RevMan 2003). We analysed categor- which contained either the same drug or a placebo (saline). The dif-
ical data using relative risk and continuous data using weighted ferences between these studies were the drug(s) used in the epidu-
mean difference. We used sensitivity analyses to evaluate the ef- ral solution. Chestnut 1987a used 0.75% lidocaine, Chestnut
fect of the trial quality. We measured heterogeneity using the I2 1987b used 0.125% bupivacaine and Chestnut 1990 used a low-
statistic. As there was no significant heterogeneity, we used a fixed- dose local anaesthetic plus an opioid (0.0625% bupivacaine plus
effect model to pool results. 0.0002% fentanyl). In one study (Johnsrud 1988) women receiv-
ing a continuous epidural infusion were also studied but, unlike the
Chestnut studies, there was no placebo. Only one study (Luxman
1996) used intermittent boluses, commonly referred to as ’top-
RESULTS ups’. The women in the study group did not receive any more
top-ups after the cervix was eight centimetres dilated, whilst the
women in the control group continued to receive top-ups until
Description of studies delivery. In this study, labour was managed actively, using rou-
tine amniotomy on admission for all women not presenting with
Six potentially eligible trials were identified, of which five were spontaneous rupture of membranes and oxytocin was given when
included. No potentially relevant on-going trials or trials awaiting cervical dilatation was less than one centimetre per hour.
assessment were identified.
Excluded study
One study (Phillips 1983) was excluded because for 27% of the Risk of bias in included studies
study participants, the reason for the epidural was obstetric or
medical rather than analgesia (see Table of ’Characteristics of ex-
cluded studies’). In addition, this trial was of poor methodological
quality. Participants were randomised by shuffling sheets and it
High-quality studies
was not placebo controlled. When the fetal head was below the
ischial spines, no further top-ups of bupivacaine 0.25% were given The high-quality studies were the three Chestnut studies (
to women in the study group. There was no significant difference Chestnut 1987a; Chestnut 1987b; Chestnut 1990) where the
in the total dose of bupivacaine received by women in the study study solutions were prepared and coded by the hospital pharma-
and control groups. cist according to a table of random numbers. The patient, anaes-
thesiologist, obstetrician, paediatrician, and nursing staff were un-
aware of the identity of the study solution. These three studies did
Included studies have exclusions which were due to protocol violations or because a
Five studies were included (Chestnut 1987a; Chestnut 1987b; caesarean section was performed after the initiation of the epidural
Chestnut 1990; Johnsrud 1988; Luxman 1996) with a total of but before substitution of the epidural infusion solution with the
462 participants (see Table of ’Characteristics of included studies’). study solution (between 11% and 18% of eligible study partici-
All participants were women in labour with their first baby, which pants). Four women in the Chestnut 1990 study were excluded
was a term singleton fetus of vertex presentation. All the studies for protocol violations postrandomisation.
(Review)
Lower-quality studies (4) Duration of the second stage of labour
The other two studies (Johnsrud 1988; Luxman 1996) were not All studies compared the length of second stage of labour between
placebo controlled and the method of randomisation was not de- the study and control group but only three studies (Chestnut
scribed. The Johnsrud 1988 study had exclusions for the same rea- 1987a; Chestnut 1987b; Luxman 1996), with 203 participants in
sons as the Chestnut studies but did not state when the protocol total, reported the mean duration in minutes with standard devi-
violations or caesarean sections occurred. Caesarean section is an ation. Meta-analysis of these three studies showed no statistically
important outcome measure but its incidence could not be calcu- significant or clinically important difference in the length of sec-
lated from this study and, because of this, neither could the inci- ond stage (weighted mean difference (WMD) -5.80 minutes, 95%
dence of spontaneous vaginal delivery. The Luxman 1996 study CI -12.91 to 1.30). It should be noted that the assumptions of the
had four postrandomisation exclusions for caesarean sections (two analysis may have been violated due to duration of second stage
in each group) which were not included in their analyses. of labour often being non-normally distributed as a result of a few
None of the longer-term outcomes (maternal satisfaction with women having very long second stages. In addition, there is some
labour care, incidence of postpartum urinary incontinence, fecal heterogeneity in this outcome (I-squared statistic = 41.8%) One
incontinence and sexual problems) were reported in any of the other study (Johnsrud 1988) reported no significant difference in
studies. mean length of second stage (37 minutes versus 38 minutes) but
could not be included in the meta-analysis because no standard
deviation was reported. The remaining study (Chestnut 1990) re-
Effects of interventions ported median length of second stage and found no significant
difference (63 minutes versus 53 minutes).
Five studies were included (Chestnut 1987a; Chestnut 1987b;
Chestnut 1990; Johnsrud 1988; Luxman 1996) with a total of
462 participants. (5) Incidence of fetal malposition at delivery
All studies except Luxman 1996 reported the number of persis-
(1) Incidence of instrumental delivery tent fetal malpositions (occipitoposterior and occipitotransverse
positions plus one case of deflection). Of the 388 participants in
All studies reported the incidence of instrumental delivery. The
whom this was measured, 36 (9%) had persistent malpositions.
overall incidence of instrumental delivery varied considerably
No studies showed any significant difference between the study
among the studies, from 15% in Luxman 1996 to 35% in
and control group with a pooled relative risk of 1.36 (95% CI
Chestnut 1987b. The meta-analysis showed no statistically signif-
0.73 to 2.56).
icant reduction in the incidence of instrumental deliveries (23%
versus 28%, relative risk (RR) 0.84, 95% confidence interval (CI)
0.61 to 1.15).
(6) Reporting of inadequate pain relief during the
second stage of labour
(2) Incidence of caesarean section All studies except Luxman 1996 reported a measure of pain relief.
All studies except Johnsrud 1988 reported the number of caesarean We defined inadequate pain relief as either ’poor’ or ’very poor’ on
sections performed postrandomisation. Meta-analysis showed no a five point scale or ’poor’ on a four point scale. Meta-analysis of
significant difference in the rate of caesarean sections (RR 0.98, the four studies (384 participants) found a statistically significant
95% CI 0.43 to 2.25). It should be noted that, overall, caesarean increase in inadequate pain relief in the study group (22% versus
section was an infrequent outcome, occurring in only 18 of the 6%, RR 3.68, 95% CI 1.99 to 6.80).
282 (6%) participants in whom it was reported.
(7) Maternal dissatisfaction with labour care

(3) Incidence of spontaneous vaginal delivery This outcome was not reported in any of the studies.
The three Chestnut studies reported the number of spontaneous
vaginal deliveries and in Luxman 1996 it was possible to calcu-
late this by subtracting the number of instrumental deliveries plus (8) Incidence of low Apgar score at one minute
caesarean sections from the total number of deliveries. Johnsrud All studies reported Apgar scores at one minute. Luxman 1996
1988 did not report the number of caesarean sections, therefore did not report the number of babies with low Apgar scores so we
we were unable to calculate the incidence of spontaneous vaginal were unable to include this study in our meta-analyses. However,
delivery for that study. Meta-analysis of spontaneous vaginal de- Luxman 1996 did report that there was no significant difference
liveries showed no significant difference (RR 1.11, 95% CI 0.95 in mean Apgar scores (9.6 versus 9.3). Of the four studies included
to 1.30). in the meta-analysis (388 participants), the three Chestnut papers
(Review)
reported the number of babies with Apgar scores less than seven (2) By type of epidural analgesia (intermittent bolus,
whilst Johnsrud 1988 reported the number of babies with Apgar continuous infusion or combined spinal epidural)
scores less than eight. Of these 388 participants, 57 (15%) had Four studies used a continuous epidural infusion (Chestnut 1987a;
low Apgar scores (defined as either less than 7 or less than 8) at one Chestnut 1987b; Chestnut 1990; Johnsrud 1988), one (Luxman
minute. Meta-analysis showed no statistically significant difference 1996) used intermittent boluses (top-ups), and none used com-
between the two groups (RR 1.55, 95% CI 0.94 to 2.55). bined spinal epidural. There was no evidence that the intervention
had different effects on these subgroups.
(9) Incidence of low Apgar score at five minutes

(3) By type of analgesic agent (local anaesthetic alone versus
Only the three Chestnut studies reported the number of babies
local anaesthetic plus opioid)
with low Apgar scores at five minutes. Of the 208 participants,
only three had an Apgar score of less than seven at five minutes Four studies used a local anaesthetic alone in their epidurals
(RR 3.92, 95% CI 0.45 to 34.21). Luxman 1996 reported no (Chestnut 1987a; Chestnut 1987b; Johnsrud 1988; Luxman
difference in mean Apgar scores between the two groups at five 1996), and one used a local anaesthetic plus an opioid (Chestnut
minutes (10 in both groups). Johnsrud 1988 did not report Apgar 1990). The only outcome which changed when the analysis was
scores at five minutes. restricted to epidurals using a local anaesthetic alone was the in-
cidence of low Apgar score at one minute, which increased in the
study group and just reached statistical significance (RR 1.77, 95%
CI 1.03 to 3.03).
(10) Incidence of admission to neonatal intensive care
unit
This outcome was not reported in any of the studies.
DISCUSSION
(11) Incidence of low umbilical arterial pH This review is limited by the small number of studies and par-
ticipants, and the only confident conclusion to be drawn is that
The three Chestnut studies measured umbilical arterial pH. There discontinuing epidural analgesia late in labour increases the rate
was no significant difference between the two groups in any study, of inadequate pain relief during the second stage of labour. Whilst
nor in the meta-analysis (208 participants, WMD -0.01, 95% CI it is possible that discontinuing epidural analgesia reduces the in-
-0.03 to 0.01). cidence of instrumental delivery, at present there is insufficient
evidence to accept or refute this. A 16% reduction in instrumental
delivery rates may be clinically important, but a much larger study
(12) Incidence of urinary incontinence postpartum than any of those included in this review would be needed to deter-
This outcome was not reported in any of the studies. mine the real impact of the intervention. The only study (Chestnut
1987b) in which the intervention group had a significantly shorter
second stage of labour was also the only study where the study
group had a significantly lower rate of instrumental delivery and a
(13) Incidence of fecal incontinence postpartum
significantly higher rate of inadequate pain relief, suggesting that
This outcome was not reported in any of the studies. there may be a trade-off between satisfactory analgesia during the
second stage and an increased risk of instrumental delivery. This
was also the study with the highest overall instrumental delivery
(14) Incidence of sexual problems postpartum rates.
This outcome was not reported in any of the studies. The management of women with epidural analgesia and the type
and dosage of drug varies between hospitals, countries and over
time. All the studies in this review used different epidural analgesia
A priori subgroup analyses protocols. Although our subgroup analyses did not show any sig-
nificant differences, the effect of discontinuing epidural analgesia
late in labour may not be the same under all conditions. The last
high-quality trial (Chestnut 1990) was undertaken 14 years ago,
(1) By parity
and practices have changed since then. Second stage management
All study participants were having their first baby so this subgroup strategies which have the potential to reduce instrumental deliv-
analysis was not undertaken. ery rates among women with epidural analgesia include the use of
(Review)
an upright position, delayed pushing, and no upper limit on the Implications for practice
length of second stage in the absence of maternal or fetal distress
At present there is insufficient evidence to support the hypothesis
and as long as progress is being made.
that discontinuing epidural analgesia reduces the incidence of in-
strumental delivery, but there is evidence that it increases women’s
The practice of discontinuing epidural analgesia for the second pain. Whenever possible, women should be informed of the risks
stage of labour is widespread, and many women may be experi- and benefits of discontinuing epidural analgesia, and encouraged
encing increased pain during their second stage of labour when to participate in the decision of whether to discontinue or not.
there is no evidence of any benefit, and there is the possibility of an
increased risk of low Apgar score at one minute. Whilst an increase Implications for research
in pain may be acceptable to some women if accompanied by a
A large, well conducted randomised controlled trial is required to
reduced risk of instrumental delivery or other adverse outcome,
determine whether discontinuing epidural analgesia reduces the
women are unlikely to find it acceptable when not accompanied
incidence of instrumental delivery, as well as being effective, safe
by any benefit. The effect of discontinuing epidurals on women’s
and acceptable to women and staff.
satisfaction with labour care is unknown as no study reported this
outcome.
ACKNOWLEDGEMENTS
Thanks to Sonja Henderson, Zarko Alfirevic and Lynn Hampson
AUTHORS’ CONCLUSIONS for their support in the production of this review.
REFERENCES
References to studies included in this review Acta Obstetricia et Gynecologica Scandinavica 1988;67(4):
355–8.
Chestnut 1987a {published data only}
Luxman 1996 {published data only}
Chestnut DH, Bates JN, Choi WW. Continuous infusion
Luxman D, Wolman I, Niv D, Cohen JR, Lottan M,
epidural analgesia with lidocaine: efficacy and influence
Pauzner D, et al. Effect of second-stage 0.25% epidural
during the second stage of labor. Obstetrics & Gynecology
bupivacaine on the outcome of labor. Gynecologic &
1987;69(3):323–7.
Obstetric Investigation 1996;42(3):167–70.
Chestnut 1987b {published data only}
∗
Chestnut DH, Vandewalker GE, Bates JN, Choi WW. References to studies excluded from this review
The influence of continuous epidural bupivacaine analgesia
on progress of the second stage of labor: a randomized, Phillips 1983 {published data only}
double-blind, placebo controlled study. Anesthesiology 1986; Phillips KC, Thomas TA. Second stage of labour with or
65(3A):A383. without extradural analgesia. Anaesthesia 1983;38:972–6.
Chestnut DH, Vandewalker GE, Owen CL, Bates JN, Choi
WW. The influence of continuous epidural bupivacaine Additional references
analgesia on the second stage of labor and method of
delivery in nulliparous women. Anesthesiology 1987;66(6): Anim-Somuah 2005
774–80. Anim-Somuah M, Smyth R, Howell C. Epidural versus
non-epidural or no analgesia in labour. Cochrane Database
Chestnut 1990 {published data only}
of Systematic Reviews 2005, Issue 4. [Art. No.: CD000331.
∗
Chestnut DH, Laszewski LJ, Pollack KL, Bates JN,
DOI: 10.1002/14651858.CD000331.pub2]
Manago NK, Choi WW. Continuous epidural infusion of
0.0625% bupivacaine-0.0002% fentanyl during the second Bates 1985
stage of labor. Anesthesiology 1990;72(4):613–8. Bates RG, Helm CW, Duncan A, Edmonds DK. Uterine
Chestnut DH, Pollack KL, Laszewski LJ, Bates JN, Choi activity in the second stage of labour and the effect
WW. Continuous epidural infusion of bupivacaine-fentanyl of epidural analgesia. British Journal of Obstetrics and
during the second stage of labor. Anesthesiology 1989;71 Gynaecology 1985;92(12):1246–50.
(3A):A841. Brown 1998
Johnsrud 1988 {published data only} Brown S, Lumley J. Maternal health after childbirth: results
Johnsrud ML, Dale PO, Løvland B. Benefits of continuous of an Australian population based survey. British Journal of
infusion epidural analgesia throughout vaginal delivery. Obstetrics and Gynaecology 1998;105(2):156–61.
(Review)
Clarke 2003 Paech 1991
Clarke M, Oxman AD, editors. Cochrane Reviewers’ Paech MJ, Godkin R. A survey of epidural analgesia practice
Handbook 4.2.0 [updated March 2003]. In: The Cochrane in Western Australian obstetric units. Anaesthesia and
Library, Issue 2, 2003. Oxford: Update Software. Updated Intensive Care 1991;19(3):388–92.
quarterly. Page 1989
Page IJ, Chapman MG. Hospital practice: a survey of the
Eason 2000
management of the second stage of labour with epidural
Eason E, Labrecque M, Wells G, Feldman P. Preventing
analgesia. Journal of Obstetrics and Gynaecology 1989;9:
perineal trauma during childbirth: a systematic review.
293–6.
Obstetrics & Gynecology 2000;95(3):464–71.
Ranta 1995
Howell 1995 Ranta P, Spalding M, Kangas-Saarela T, Jokela R, Hollmen
Howell CJ. Discontinuing epidural block at 8cm dilatation. A, Jouppila P, et al. Maternal expectations and experiences
In: Enkin MW, Keirse MJNC, Renfrew MJ, Neilson JP of labour pain - options of 1091 Finnish parturients. Acta
(eds.) Pregnancy and Childbirth Module of The Cochrane Anaesthesiologica Scandinavica 1995;39(1):60–6.
Database of Systematic Reviews, 1995 [updated 24 February RevMan 2003 [Computer program]
1995]. Available from BMJ Publishing Group: London. The Cochrane Collaboration. Review Manager (RevMan).
Johanson 1999 Version 4.2.2 for Windows. Oxford, England: The
Johanson RB, Heycock E, Carter J, Sultan AH, Walklate K, Cochrane Collaboration, 2003.
Jones PW. Maternal and child health after assisted vaginal Roberts 2002
delivery: five-year follow up of a randomised controlled Roberts CL, Algert CS, Douglas I, Tracy SK, Peat B. Trends
study comparing forceps and ventouse. British Journal of in labour and birth interventions among low-risk women in
Obstetrics and Gynaecology 1999;106(6):544–9. NSW. Australian and New Zealand Journal of Obstetrics and
Gynecology 2002;42(2):176–81.
Leighton 2002
Roberts 2003
Leighton BL, Halpern SH. The effects of epidural analgesia
Roberts CL, Raynes-Greenow CH, Upton A, Douglas ID,
on labor, maternal, and neonatal outcomes: a systematic
Peat B. The management of labour among women with
review. American Journal of Obstetrics and Gynecology 2002;
epidural analgesia. Australian and New Zealand Journal of
186(5 Suppl):S69–S77.
Obstetrics and Gynaecology 2003;43:78–81.
MacArthur 1997 Sultan 1993
MacArthur C, Bick DE, Keighley MR. Faecal incontinence Sultan AH, Kamm MA, Hudson CN, Thomas JM, Bartram
after childbirth. British Journal of Obstetrics and Gynaecology CI. Anal-sphincter disruption during vaginal delivery. New
1997;104(1):46–50. England Journal of Medicine 1993;329(26):1905–11.
Mayberry 1999 Vandendriesen 1998
Mayberry LJ, Wood SH, Strange LB, Lee L, Heisler DR, Vandendriesen NM, Lim W, Paech MJ. Labour ward
Neilson-Smith K. Managing second-stage labor: exploring midwifery staff epidural knowledge and practice. Anaesthesia
the variables during the second stage. AWHONN Lifelines and Intensive Care 1998;26(4):411–9.
1999 Dec–2000 Jan;3(6):28–34. ∗
Indicates the major publication for the study
(Review)
CHARACTERISTICS OF STUDIES
Characteristics of included studies [ordered by study ID]
Chestnut 1987a
Methods Centrally randomised according to a table of random numbers. Caregivers and participants blinded to
study solution
Participants Nulliparous women with term (>= 36 weeks) singleton fetuses of vertex presentation who were receiving a
continuous epidural infusion of 0.75% lidocaine. Exclusion criteria: women with pre-eclampsia, insulin-
dependent diabetes, intrauterine growth retardation or chorioamnionitis. 65 women enrolled, 12 excluded
Interventions When the cervix was dilated by 8 or more centimetres, the syringe containing 0.75% lidocaine was
substituted with a study solution. In the study group (n = 27) this contained saline and in the control
group (n = 26) this contained 0.75% lidocaine
Outcomes Instrumental delivery, oxytocin augmentation, fetal malposition, caesarean section, inadequate pain relief,
Apgar scores < 7 at 1 and 5 mins, umbilical arterial pH, duration of 1st and 2nd stage of labour
Notes Women encouraged to push from full dilatation.
Risk of bias
Bias Authors’ judgement Support for judgement
Allocation concealment? Low risk A - Adequate
Chestnut 1987b
study solution
Participants Healthy nulliparous women with term (>= 36 weeks) singleton fetuses of vertex presentation who were
receiving a continuous epidural infusion of 0.125% bupivacaine. Exclusion criteria: women with pre-
eclampsia or insulin-dependent diabetes. 110 women enrolled, 18 excluded
Interventions When the cervix was dilated by 8 or more centimetres, the syringe containing 0.125% bupivacaine was
substituted with a study solution . In the study group (n = 46) this contained saline and in the control
group (n = 46) this contained 0.125% bupivacaine
Notes
Risk of bias
(Review)
Chestnut 1987b (Continued)
Chestnut 1990
study solution
Participants Healthy nulliparous women with term (>= 36 weeks) singleton fetuses of vertex presentation who were
receiving a continuous epidural infusion of 0.0625% bupivacaine and 0.0002% fentanyl. Exclusion
criteria: women with pre-eclampsia or insulin-dependent diabetes. 75 women enrolled, 12 excluded
(including 2 in each group excluded postrandomisation for protocol violation)
Interventions When the cervix was dilated by 8 or more centimetres, the syringe containing 0.0625% bupivacaine 0.
0002% fentanyl was substituted with the study solution. In the study group (n = 34) this contained saline
and in the control group (n = 29) this contained bupivacaine -0.0002% fentanyl
Notes
Risk of bias
Johnsrud 1988
Methods Randomly assigned - method not described. Unblinded.
Participants Nulliparous women with term (exceeding 37 weeks) singleton fetuses of vertex presentation who were
receiving a continuous epidural infusion of 2.5 mg/ml bupivacaine. 200 women enrolled, 20 excluded,
unclear whether pre or postrandomisation
Interventions In the study group (n = 90) the epidural infusion was discontinued in 2nd stage whilst the same epidural
infusion was continued in the control group (n = 90)
Outcomes Instrumental delivery, fetal malposition, inadequate pain relief, Apgar scores < 8 at 1 and 5 mins, duration
of 1st and 2nd stage of labour
Notes Oxytocin augmentation given to all women. Active pushing time was not to exceed 1 hour
Risk of bias
(Review)
Johnsrud 1988 (Continued)
Allocation concealment? Unclear risk B - Unclear
Luxman 1996
Methods Randomly assigned - method not described. Unblinded.
Participants Primiparous women in spontaneous labour with uncomplicated pregnancies at term and receiving epidural
top-ups of 0.25% bupivacaine. 74 women enrolled, 4 postrandomisation exclusions
Interventions The study group (n = 37) was not given top-ups after cervical dilatation had reached 8 centimetres, whilst
the control group (n = 37) continued to receive top-ups until delivery
Outcomes Instrumental delivery, oxytocin augmentation, caesarean section, mean Apgar scores at 1 and 5 mins, rate
of dilatation, duration of 2nd stage of labour. No measure of pain relief reported
Notes Amniotomy performed on all women who did not present to the labour ward with spontaneous rupture
of membranes
Risk of bias
Allocation concealment? Unclear risk B - Unclear
mins: minutes
Characteristics of excluded studies [ordered by study ID]
Study Reason for exclusion
Phillips 1983 This study was excluded because in 16 of the 59 women (27%), the reason for the epidural was obstetric or medical
rather than for analgesia
(Review)
DATA AND ANALYSES
Comparison 1. Discontinued versus continued
No. of No. of
Outcome or subgroup title studies participants Statistical method Effect size
1 Instrumental delivery 5 462 Risk Ratio (M-H, Fixed, 95% CI) 0.84 [0.61, 1.15]
2 Instrumental delivery by 5 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
subgroup analysis
2.1 Continuous infusion 4 388 Risk Ratio (M-H, Fixed, 95% CI) 0.81 [0.58, 1.12]
2.2 Intermittant bolus 1 74 Risk Ratio (M-H, Fixed, 95% CI) 1.2 [0.40, 3.59]
2.3 LA alone 4 399 Risk Ratio (M-H, Fixed, 95% CI) 0.85 [0.62, 1.18]
2.4 LA + opioid 1 63 Risk Ratio (M-H, Fixed, 95% CI) 0.71 [0.24, 2.09]
3 Caesarean section 4 282 Risk Ratio (M-H, Fixed, 95% CI) 0.98 [0.43, 2.25]
4 Caesarean section by subgroup 4 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
analysis
4.2 Intermittant bolus 1 74 Risk Ratio (M-H, Fixed, 95% CI) 1.0 [0.15, 6.73]
5 Spontaneous vaginal delivery 4 282 Risk Ratio (M-H, Fixed, 95% CI) 1.11 [0.95, 1.30]
6 Spontaneous vaginal delivery by 4 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
subgroup analysis
6.2 intermittant bolus 1 74 Risk Ratio (M-H, Fixed, 95% CI) 0.97 [0.77, 1.22]
7 Duration of 2nd stage 3 203 Mean Difference (IV, Fixed, 95% CI) -5.80 [-12.91, 1.30]
8 Duration of 2nd stage by 3 Mean Difference (IV, Fixed, 95% CI) Subtotals only
subgroup analysis
8.1 Continuous infusion 2 133 Mean Difference (IV, Fixed, 95% CI) -15.12 [-35.42, 5.
18]
8.2 Intermittant bolus 1 70 Mean Difference (IV, Fixed, 95% CI) -4.5 [-12.09, 3.09]
9 Fetal malposition 4 388 Risk Ratio (M-H, Fixed, 95% CI) 1.36 [0.73, 2.56]
10 Fetal malposition by subgroup 4 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
analysis
11 Inadequate pain relief 4 384 Risk Ratio (M-H, Fixed, 95% CI) 3.68 [1.99, 6.80]
12 Inadequate pain relief by 4 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
subgroup analysis
13 Apgar score < 7 or < 8 at 1 4 388 Risk Ratio (M-H, Fixed, 95% CI) 1.55 [0.94, 2.55]
minute
14 Apgar score < 7 or < 8 at 1 4 Risk Ratio (M-H, Fixed, 95% CI) Subtotals only
minute by subgroup analysis
(Review)
15 Apgar score < 7 at 5 minutes 3 208 Risk Ratio (M-H, Fixed, 95% CI) 3.92 [0.45, 34.21]
16 Umbilical arterial pH 3 208 Mean Difference (IV, Fixed, 95% CI) -0.01 [-0.03, 0.01]
17 Umbilical arterial pH by 3 Mean Difference (IV, Fixed, 95% CI) Subtotals only
subgroup analysis
17.3 LA alone 2 145 Mean Difference (IV, Fixed, 95% CI) -0.02 [-0.05, 0.00]
17.4 LA + opioid 1 63 Mean Difference (IV, Fixed, 95% CI) 0.01 [-0.02, 0.04]
Analysis 1.1. Comparison 1 Discontinued versus continued, Outcome 1 Instrumental delivery.
Review: Discontinuation of epidural analgesia late in labour for reducing the adverse delivery outcomes associated with epidural analgesia
Comparison: 1 Discontinued versus continued
Outcome: 1 Instrumental delivery
Study or subgroup Discontinued Continued Risk Ratio Weight Risk Ratio

n/N n/N M-H,Fixed,95% CI M-H,Fixed,95% CI
Chestnut 1987a 9/27 8/26 12.8 % 1.08 [ 0.49, 2.38 ]
Chestnut 1987b 11/46 21/46 33.0 % 0.52 [ 0.29, 0.96 ]
Chestnut 1990 5/34 6/29 10.2 % 0.71 [ 0.24, 2.09 ]
Johnsrud 1988 23/90 23/90 36.1 % 1.00 [ 0.61, 1.65 ]
Luxman 1996 6/37 5/37 7.9 % 1.20 [ 0.40, 3.59 ]
Total (95% CI) 234 228 100.0 % 0.84 [ 0.61, 1.15 ]

Total events: 54 (Discontinued), 63 (Continued)
Heterogeneity: Chi2 = 3.72, df = 4 (P = 0.45); I2 =0.0%
Test for overall effect: Z = 1.10 (P = 0.27)
0.1 0.2 0.5 1 2 5 10

Favours discontinued Favours continued
(Review)
Analysis 1.2. Comparison 1 Discontinued versus continued, Outcome 2 Instrumental delivery by subgroup
analysis.
Outcome: 2 Instrumental delivery by subgroup analysis

1 Continuous infusion
Chestnut 1987a 9/27 8/26 13.9 % 1.08 [ 0.49, 2.38 ]
Chestnut 1987b 11/46 21/46 35.8 % 0.52 [ 0.29, 0.96 ]
Chestnut 1990 5/34 6/29 11.0 % 0.71 [ 0.24, 2.09 ]
Johnsrud 1988 23/90 23/90 39.2 % 1.00 [ 0.61, 1.65 ]
Subtotal (95% CI) 197 191 100.0 % 0.81 [ 0.58, 1.12 ]

Heterogeneity: Chi2 = 3.27, df = 3 (P = 0.35); I2 =8%
2 Intermittant bolus
Luxman 1996 6/37 5/37 100.0 % 1.20 [ 0.40, 3.59 ]
Subtotal (95% CI) 37 37 100.0 % 1.20 [ 0.40, 3.59 ]

Heterogeneity: not applicable
3 LA alone
Chestnut 1987a 9/27 8/26 14.3 % 1.08 [ 0.49, 2.38 ]
Chestnut 1987b 11/46 21/46 36.7 % 0.52 [ 0.29, 0.96 ]
Johnsrud 1988 23/90 23/90 40.2 % 1.00 [ 0.61, 1.65 ]
Luxman 1996 6/37 5/37 8.7 % 1.20 [ 0.40, 3.59 ]
Subtotal (95% CI) 200 199 100.0 % 0.85 [ 0.62, 1.18 ]

4 LA + opioid
Chestnut 1990 5/34 6/29 100.0 % 0.71 [ 0.24, 2.09 ]
Subtotal (95% CI) 34 29 100.0 % 0.71 [ 0.24, 2.09 ]

0.1 0.2 0.5 1 2 5 10

(Review)
Analysis 1.3. Comparison 1 Discontinued versus continued, Outcome 3 Caesarean section.
Outcome: 3 Caesarean section

Chestnut 1987a 1/27 0/26 5.0 % 2.89 [ 0.12, 67.96 ]
Chestnut 1987b 6/46 6/46 59.3 % 1.00 [ 0.35, 2.87 ]
Chestnut 1990 0/34 1/29 16.0 % 0.29 [ 0.01, 6.76 ]
Luxman 1996 2/37 2/37 19.8 % 1.00 [ 0.15, 6.73 ]
Total (95% CI) 144 138 100.0 % 0.98 [ 0.43, 2.25 ]

0.01 0.1 1 10 100

(Review)
Analysis 1.4. Comparison 1 Discontinued versus continued, Outcome 4 Caesarean section by subgroup
analysis.
Outcome: 4 Caesarean section by subgroup analysis

Chestnut 1987a 1/27 0/26 6.3 % 2.89 [ 0.12, 67.96 ]
Chestnut 1987b 6/46 6/46 73.9 % 1.00 [ 0.35, 2.87 ]
Chestnut 1990 0/34 1/29 19.9 % 0.29 [ 0.01, 6.76 ]
Subtotal (95% CI) 107 101 100.0 % 0.98 [ 0.39, 2.46 ]

Luxman 1996 2/37 2/37 100.0 % 1.00 [ 0.15, 6.73 ]
Subtotal (95% CI) 37 37 100.0 % 1.00 [ 0.15, 6.73 ]

3 LA alone
Chestnut 1987a 1/27 0/26 6.0 % 2.89 [ 0.12, 67.96 ]
Chestnut 1987b 6/46 6/46 70.5 % 1.00 [ 0.35, 2.87 ]
Luxman 1996 2/37 2/37 23.5 % 1.00 [ 0.15, 6.73 ]
Subtotal (95% CI) 110 109 100.0 % 1.11 [ 0.46, 2.68 ]

4 LA + opioid
Chestnut 1990 0/34 1/29 100.0 % 0.29 [ 0.01, 6.76 ]
Subtotal (95% CI) 34 29 100.0 % 0.29 [ 0.01, 6.76 ]

0.01 0.1 1 10 100

(Review)
Analysis 1.5. Comparison 1 Discontinued versus continued, Outcome 5 Spontaneous vaginal delivery.
Outcome: 5 Spontaneous vaginal delivery

Chestnut 1987a 17/27 18/26 20.1 % 0.91 [ 0.62, 1.34 ]
Chestnut 1987b 29/46 19/46 20.9 % 1.53 [ 1.01, 2.30 ]
Chestnut 1990 29/34 22/29 26.1 % 1.12 [ 0.88, 1.44 ]
Luxman 1996 29/37 30/37 32.9 % 0.97 [ 0.77, 1.22 ]
Total (95% CI) 144 138 100.0 % 1.11 [ 0.95, 1.30 ]

0.1 0.2 0.5 1 2 5 10

(Review)
Analysis 1.6. Comparison 1 Discontinued versus continued, Outcome 6 Spontaneous vaginal delivery by
subgroup analysis.
Outcome: 6 Spontaneous vaginal delivery by subgroup analysis

Chestnut 1987a 17/27 18/26 30.0 % 0.91 [ 0.62, 1.34 ]
Chestnut 1987b 29/46 19/46 31.1 % 1.53 [ 1.01, 2.30 ]
Chestnut 1990 29/34 22/29 38.9 % 1.12 [ 0.88, 1.44 ]
Subtotal (95% CI) 107 101 100.0 % 1.18 [ 0.97, 1.45 ]

2 intermittant bolus
Luxman 1996 29/37 30/37 100.0 % 0.97 [ 0.77, 1.22 ]
Subtotal (95% CI) 37 37 100.0 % 0.97 [ 0.77, 1.22 ]

3 LA alone
Chestnut 1987a 17/27 18/26 27.2 % 0.91 [ 0.62, 1.34 ]
Chestnut 1987b 29/46 19/46 28.2 % 1.53 [ 1.01, 2.30 ]
Luxman 1996 29/37 30/37 44.6 % 0.97 [ 0.77, 1.22 ]
Subtotal (95% CI) 110 109 100.0 % 1.11 [ 0.92, 1.34 ]

4 LA + opioid
Chestnut 1990 29/34 22/29 100.0 % 1.12 [ 0.88, 1.44 ]
Subtotal (95% CI) 34 29 100.0 % 1.12 [ 0.88, 1.44 ]

0.1 0.2 0.5 1 2 5 10

(Review)
Analysis 1.7. Comparison 1 Discontinued versus continued, Outcome 7 Duration of 2nd stage.
Outcome: 7 Duration of 2nd stage
Mean Mean
Study or subgroup Discontinued Continued Difference Weight Difference
N Mean(SD) N Mean(SD) IV,Fixed,95% CI IV,Fixed,95% CI
Chestnut 1987a 27 76 (48) 26 73 (63) 5.5 % 3.00 [ -27.24, 33.24 ]
Chestnut 1987b 40 94 (54) 40 124 (70) 6.7 % -30.00 [ -57.40, -2.60 ]
Luxman 1996 35 38.5 (15) 35 43 (17.3) 87.7 % -4.50 [ -12.09, 3.09 ]
Total (95% CI) 102 101 100.0 % -5.80 [ -12.91, 1.30 ]

Test for subgroup differences: Not applicable
-100 -50 0 50 100

(Review)
Analysis 1.8. Comparison 1 Discontinued versus continued, Outcome 8 Duration of 2nd stage by subgroup
analysis.
Outcome: 8 Duration of 2nd stage by subgroup analysis
Mean Mean
Chestnut 1987a 27 76 (48) 26 73 (63) 45.1 % 3.00 [ -27.24, 33.24 ]
Chestnut 1987b 40 94 (54) 40 124 (70) 54.9 % -30.00 [ -57.40, -2.60 ]
Subtotal (95% CI) 67 66 100.0 % -15.12 [ -35.42, 5.18 ]

Luxman 1996 35 38.5 (15) 35 43 (17.3) 100.0 % -4.50 [ -12.09, 3.09 ]
Subtotal (95% CI) 35 35 100.0 % -4.50 [ -12.09, 3.09 ]

Test for subgroup differences: Chi2 = 0.92, df = 1 (P = 0.34), I2 =0.0%
-100 -50 0 50 100

(Review)
Analysis 1.9. Comparison 1 Discontinued versus continued, Outcome 9 Fetal malposition.
Outcome: 9 Fetal malposition

Chestnut 1987a 2/27 2/26 13.4 % 0.96 [ 0.15, 6.34 ]
Chestnut 1987b 6/46 7/46 46.1 % 0.86 [ 0.31, 2.36 ]
Chestnut 1990 3/34 2/29 14.2 % 1.28 [ 0.23, 7.14 ]
Johnsrud 1988 10/90 4/90 26.3 % 2.50 [ 0.81, 7.68 ]
Total (95% CI) 197 191 100.0 % 1.36 [ 0.73, 2.56 ]

0.1 0.2 0.5 1 2 5 10

(Review)
Analysis 1.10. Comparison 1 Discontinued versus continued, Outcome 10 Fetal malposition by subgroup
analysis.
Outcome: 10 Fetal malposition by subgroup analysis

3 LA alone
Chestnut 1987a 2/27 2/26 15.6 % 0.96 [ 0.15, 6.34 ]
Chestnut 1987b 6/46 7/46 53.7 % 0.86 [ 0.31, 2.36 ]
Johnsrud 1988 10/90 4/90 30.7 % 2.50 [ 0.81, 7.68 ]
Subtotal (95% CI) 163 162 100.0 % 1.38 [ 0.70, 2.71 ]

4 LA + opioid
Chestnut 1990 3/34 2/29 100.0 % 1.28 [ 0.23, 7.14 ]
Subtotal (95% CI) 34 29 100.0 % 1.28 [ 0.23, 7.14 ]

0.1 0.2 0.5 1 2 5 10

(Review)
Analysis 1.11. Comparison 1 Discontinued versus continued, Outcome 11 Inadequate pain relief.
Outcome: 11 Inadequate pain relief

Chestnut 1987a 12/27 6/26 54.6 % 1.93 [ 0.85, 4.37 ]
Chestnut 1987b 17/44 2/44 17.9 % 8.50 [ 2.09, 34.62 ]
Chestnut 1990 7/34 1/29 9.6 % 5.97 [ 0.78, 45.73 ]
Johnsrud 1988 6/90 2/90 17.9 % 3.00 [ 0.62, 14.47 ]
Total (95% CI) 195 189 100.0 % 3.68 [ 1.99, 6.80 ]

0.01 0.1 1 10 100

(Review)
Analysis 1.12. Comparison 1 Discontinued versus continued, Outcome 12 Inadequate pain relief by
subgroup analysis.
Outcome: 12 Inadequate pain relief by subgroup analysis

3 LA alone
Chestnut 1987a 12/27 6/26 60.4 % 1.93 [ 0.85, 4.37 ]
Chestnut 1987b 17/44 2/44 19.8 % 8.50 [ 2.09, 34.62 ]
Johnsrud 1988 6/90 2/90 19.8 % 3.00 [ 0.62, 14.47 ]
Subtotal (95% CI) 161 160 100.0 % 3.44 [ 1.81, 6.53 ]

4 LA + opioid
Chestnut 1990 7/34 1/29 100.0 % 5.97 [ 0.78, 45.73 ]
Subtotal (95% CI) 34 29 100.0 % 5.97 [ 0.78, 45.73 ]

0.01 0.1 1 10 100

(Review)
Analysis 1.13. Comparison 1 Discontinued versus continued, Outcome 13 Apgar score < 7 or < 8 at 1
minute.
Outcome: 13 Apgar score < 7 or < 8 at 1 minute

Chestnut 1987a 9/27 1/26 4.6 % 8.67 [ 1.18, 63.69 ]
Chestnut 1987b 10/46 6/46 26.9 % 1.67 [ 0.66, 4.21 ]
Chestnut 1990 3/34 4/29 19.3 % 0.64 [ 0.16, 2.63 ]
Johnsrud 1988 13/90 11/90 49.2 % 1.18 [ 0.56, 2.50 ]
Total (95% CI) 197 191 100.0 % 1.55 [ 0.94, 2.55 ]

0.01 0.1 1 10 100

(Review)
Analysis 1.14. Comparison 1 Discontinued versus continued, Outcome 14 Apgar score < 7 or < 8 at 1
minute by subgroup analysis.
Outcome: 14 Apgar score < 7 or < 8 at 1 minute by subgroup analysis

3 LA alone
Chestnut 1987a 9/27 1/26 5.7 % 8.67 [ 1.18, 63.69 ]
Chestnut 1987b 10/46 6/46 33.3 % 1.67 [ 0.66, 4.21 ]
Johnsrud 1988 13/90 11/90 61.0 % 1.18 [ 0.56, 2.50 ]
Subtotal (95% CI) 163 162 100.0 % 1.77 [ 1.03, 3.03 ]

4 LA + opioid
Chestnut 1990 3/34 4/29 100.0 % 0.64 [ 0.16, 2.63 ]
Subtotal (95% CI) 34 29 100.0 % 0.64 [ 0.16, 2.63 ]

0.01 0.1 1 10 100

(Review)
Analysis 1.15. Comparison 1 Discontinued versus continued, Outcome 15 Apgar score < 7 at 5 minutes.
Outcome: 15 Apgar score < 7 at 5 minutes

Chestnut 1987a 2/27 0/26 50.5 % 4.82 [ 0.24, 95.88 ]
Chestnut 1987b 1/46 0/46 49.5 % 3.00 [ 0.13, 71.78 ]
Chestnut 1990 0/34 0/29 Not estimable
Total (95% CI) 107 101 100.0 % 3.92 [ 0.45, 34.21 ]

0.01 0.1 1 10 100

Analysis 1.16. Comparison 1 Discontinued versus continued, Outcome 16 Umbilical arterial pH.
Outcome: 16 Umbilical arterial pH
Mean Mean
Chestnut 1987a 27 7.22 (0.07) 26 7.24 (0.07) 30.4 % -0.02 [ -0.06, 0.02 ]
Chestnut 1987b 46 7.21 (0.12) 46 7.24 (0.06) 28.7 % -0.03 [ -0.07, 0.01 ]
Chestnut 1990 34 7.27 (0.06) 29 7.26 (0.07) 40.9 % 0.01 [ -0.02, 0.04 ]
Total (95% CI) 107 101 100.0 % -0.01 [ -0.03, 0.01 ]

Test for subgroup differences: Not applicable
-0.5 -0.25 0 0.25 0.5

(Review)
Analysis 1.17. Comparison 1 Discontinued versus continued, Outcome 17 Umbilical arterial pH by
subgroup analysis.
Outcome: 17 Umbilical arterial pH by subgroup analysis
Mean Mean
3 LA alone
Chestnut 1987a 27 7.22 (0.07) 26 7.24 (0.07) 51.4 % -0.02 [ -0.06, 0.02 ]
Chestnut 1987b 46 7.21 (0.12) 46 7.24 (0.06) 48.6 % -0.03 [ -0.07, 0.01 ]
Subtotal (95% CI) 73 72 100.0 % -0.02 [ -0.05, 0.00 ]

4 LA + opioid
Chestnut 1990 34 7.27 (0.06) 29 7.26 (0.07) 100.0 % 0.01 [ -0.02, 0.04 ]
Subtotal (95% CI) 34 29 100.0 % 0.01 [ -0.02, 0.04 ]

Test for subgroup differences: Chi2 = 2.61, df = 1 (P = 0.11), I2 =62%
-0.5 -0.25 0 0.25 0.5

WHAT’S NEW
Last assessed as up-to-date: 28 October 2007.
(Review)
Date Event Description
16 July 2010 Amended Contact details updated.
HISTORY
Protocol first published: Issue 4, 2003
Review first published: Issue 4, 2004
Date Event Description
13 February 2009 Amended Contact details updated.
12 February 2008 Amended Converted to new review format.
29 October 2006 New search has been performed Search updated. No new trial reports identified.
9 August 2004 New citation required and conclusions have changed Substantive amendment
CONTRIBUTIONS OF AUTHORS
Christine Roberts was responsible for the project proposal and was the primary author for the protocol. Christine Roberts and Siranda
Torvaldsen independently assessed trials for inclusion and extracted the data. Siranda Torvaldsen was responsible for revisions of the
protocol, data entry, analysis and was the primary author for the review. Jane Bell provided valuable Review Manager expertise. Christine
Roberts, Siranda Torvaldsen, Jane Bell and Camille Raynes-Greenow all gathered background information and provided editorial
assistance.
DECLARATIONS OF INTEREST
None known.
SOURCES OF SUPPORT
(Review)
Internal sources
• Centre for Perinatal Health Services Research, University of Sydney, Australia.
External sources
• Commonwealth Department of Health and Ageing (JC Bell), Australia.
• National Health and Medical Research Council (S Torvaldsen, CL Roberts and CH Raynes-Greenow’s positions are wholly or
partially funded), Australia.
INDEX TERMS
Medical Subject Headings (MeSH)

∗ Analgesia,
Epidural [adverse effects]; ∗ Analgesia, Obstetrical [adverse effects]; ∗ Labor, Obstetric; Delivery, Obstetric [methods]; Ran-
domized Controlled Trials as Topic
MeSH check words

Female; Humans; Pregnancy
(Review)

Discontinuation of Epidural Analgesia Late in Labour For Reducing The Adverse Delivery Outcomes Associated With Epidural 1

Enviado por

Dados do documento

Título original

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

Discontinuation of Epidural Analgesia Late in Labour For Reducing The Adverse Delivery Outcomes Associated With Epidural 1

Enviado por

Direitos autorais:

Formatos disponíveis

Cochrane Database of Systematic Reviews

Discontinuation of epidural analgesia late in labour for

Torvaldsen S, Roberts CL, Bell JC, Raynes-Greenow CH

Torvaldsen S, Roberts CL, Bell JC, Raynes-Greenow CH.

Discontinuation of epidural analgesia late in labour for

Siranda Torvaldsen1 , Christine L Roberts2 , Jane C Bell3 , Camille H Raynes-Greenow4

Editorial group: Cochrane Pregnancy and Childbirth Group.

PLAIN LANGUAGE SUMMARY

BACKGROUND risks for mothers of vaginal/perineal trauma and anal sphinc-

2. To assess the impact of discontinuing epidural analgesia late

METHODS Electronic searches

(7) Maternal dissatisfaction with labour care

(9) Incidence of low Apgar score at five minutes

Characteristics of included studies [ordered by study ID]

Notes Women encouraged to push from full dilatation.

Bias Authors’ judgement Support for judgement

Allocation concealment? Low risk A - Adequate

Bias Authors’ judgement Support for judgement

Allocation concealment? Low risk A - Adequate

Bias Authors’ judgement Support for judgement

Allocation concealment? Low risk A - Adequate

Methods Randomly assigned - method not described. Unblinded.

Bias Authors’ judgement Support for judgement

Allocation concealment? Unclear risk B - Unclear

Methods Randomly assigned - method not described. Unblinded.

Bias Authors’ judgement Support for judgement

Allocation concealment? Unclear risk B - Unclear

Characteristics of excluded studies [ordered by study ID]

Study Reason for exclusion

Comparison 1. Discontinued versus continued

Analysis 1.1. Comparison 1 Discontinued versus continued, Outcome 1 Instrumental delivery.

Comparison: 1 Discontinued versus continued

Outcome: 1 Instrumental delivery

Study or subgroup Discontinued Continued Risk Ratio Weight Risk Ratio

Chestnut 1987b 11/46 21/46 33.0 % 0.52 [ 0.29, 0.96 ]

Chestnut 1990 5/34 6/29 10.2 % 0.71 [ 0.24, 2.09 ]

Johnsrud 1988 23/90 23/90 36.1 % 1.00 [ 0.61, 1.65 ]

Luxman 1996 6/37 5/37 7.9 % 1.20 [ 0.40, 3.59 ]

Total (95% CI) 234 228 100.0 % 0.84 [ 0.61, 1.15 ]

0.1 0.2 0.5 1 2 5 10

Comparison: 1 Discontinued versus continued

Outcome: 2 Instrumental delivery by subgroup analysis

Study or subgroup Discontinued Continued Risk Ratio Weight Risk Ratio

Chestnut 1987b 11/46 21/46 35.8 % 0.52 [ 0.29, 0.96 ]

Chestnut 1990 5/34 6/29 11.0 % 0.71 [ 0.24, 2.09 ]

Johnsrud 1988 23/90 23/90 39.2 % 1.00 [ 0.61, 1.65 ]

Subtotal (95% CI) 197 191 100.0 % 0.81 [ 0.58, 1.12 ]

Subtotal (95% CI) 37 37 100.0 % 1.20 [ 0.40, 3.59 ]

Chestnut 1987b 11/46 21/46 36.7 % 0.52 [ 0.29, 0.96 ]

Johnsrud 1988 23/90 23/90 40.2 % 1.00 [ 0.61, 1.65 ]

Luxman 1996 6/37 5/37 8.7 % 1.20 [ 0.40, 3.59 ]

Subtotal (95% CI) 200 199 100.0 % 0.85 [ 0.62, 1.18 ]

Subtotal (95% CI) 34 29 100.0 % 0.71 [ 0.24, 2.09 ]

0.1 0.2 0.5 1 2 5 10

Comparison: 1 Discontinued versus continued

Outcome: 3 Caesarean section

Study or subgroup Discontinued Continued Risk Ratio Weight Risk Ratio

Chestnut 1987b 6/46 6/46 59.3 % 1.00 [ 0.35, 2.87 ]

Chestnut 1990 0/34 1/29 16.0 % 0.29 [ 0.01, 6.76 ]

Luxman 1996 2/37 2/37 19.8 % 1.00 [ 0.15, 6.73 ]

Total (95% CI) 144 138 100.0 % 0.98 [ 0.43, 2.25 ]

0.01 0.1 1 10 100

Comparison: 1 Discontinued versus continued