Dr MMLH Wassen, Department of Obstetrics and Gynaecology, Maastricht University Medical Centre, P. Debyelaan 25, PO 5800, 6202 AZ, Maastricht, the Netherlands. Email firstname.lastname@example.org
Please cite this paper as: Wassen M, Zuijlen J, Roumen F, Smits L, Marcus M, Nijhuis J. Early versus late epidural analgesia and risk of instrumental delivery in nulliparous women: a systematic review. BJOG 2011;118:655–661.
Background The optimal timing of epidural analgesia during labour and delivery has been a controversial issue.
Objective Review of the literature regarding the relation between the timing of epidural analgesia and the rate of caesarean or instrumental vaginal deliveries.
Search strategy Pubmed, Embase and the Cochrane Library were searched for articles published until 31 July 2010.
Selection criteria Studies were selected in which the effects of early latent phase (defined as a cervical dilatation of 3 cm or less) epidural analgesia (including combined-spinal epidural) and late active phase epidural analgesia on the mode of delivery in nulliparous women at 36 weeks of gestation or more were evaluated.
Data collection and analysis Data extraction was completed by using a data-extraction form. Risk ratio and its 95% confidence intervals were calculated for caesarean delivery and instrumental vaginal delivery. Pooled data were calculated.
Main results The search retrieved 20 relevant articles, of which six fulfilled the selection criteria of inclusion. These six studies reported on 15 399 nulliparous women in spontaneous or induced labour with a request for analgesia. Risk of caesarean delivery (pooled risk ratio 1.02, 95% CI 0.96–1.08) or instrumental vaginal delivery (pooled risk ratio 0.96, 95% CI 0.89–1.05) was not significantly different between groups.
Authors’ conclusions This systematic review showed no increased risk of caesarean delivery or instrumental vaginal delivery for women receiving early epidural analgesia at cervical dilatation of 3 cm or less in comparison with late epidural analgesia.
Epidural analgesia (EA) is the most effective treatment for pain control during labour and delivery.1,2 The effect of regional analgesia on progress of labour and mode of delivery has often been debated. Results of randomised controlled trials (RCTs) and systematic reviews published between 2002 and 2004 did not demonstrate any difference in the rate of caesarean deliveries between women who had received EA and women who only received intravenous analgesia.1,3–6 A Cochrane review, published, in 2005, showed that EA was associated with an increased risk of instrumental vaginal birth (pooled risk ratio [RR] 1.38, 95% CI 1.24–1.53) compared with deliveries with nonepidural analgesia or no analgesia.2
A landmark study by Wong et al.,7 published in 2005, provided evidence that early epidurals in comparison with late epidurals do not cause an increased rate of caesarean deliveries and instrumental vaginal deliveries in nulliparous women with spontaneous labour. Recently, more studies have been performed on the timing effect of EA related to the mode of delivery. A systematic review in 2007 by Marucci et al.8 reported on the timing effects of neuraxial analgesia on the mode of delivery. That review included five RCTs, one impact cohort study and three retrospective cohort studies, of which one contains only data from an abstract. A similar rate for caesarean delivery (odds ratio [OR] 1.00, 95% CI 0.82–1.23) and instrumental vaginal delivery in the early neuraxial analgesia and control group (OR 1.00, 95% CI 0.83–1.21) was shown. However, these latter results may not be convincing because of the use of a too broad definition of the early group (before 4–5 cm dilatation) and the use of an incomparable control group (including parenteral opioid and/or late EA).8
Friedman9 analysed 500 nulliparous women in 1955 and divided the first stage into an early latent phase and an ensuing active phase. He described the start of the active phase of labour at 3–4 cm cervical dilatation. The introduction of a national guideline in the Netherlands in 2008 on the management of labour pain resulted not only in an increase in EA requests, but also those requests being made at an earlier stage in the course of labour.10 Therefore we are interested in the effect of early EA strictly defined as 3 cm or less in the latent phase on the mode of delivery.
The main objective of this report was to review recent literature on the influence of this stricter definition of early EA (including combined-spinal epidural) compared with late EA in nulliparous women at 36 weeks or more of gestation, on the rate of caesarean deliveries or instrumental vaginal deliveries.
Searching and selection
A search was made for RCTs, prospective cohort studies and retrospective cohort studies in which the effects of early EA (defined as cervical dilatation <4 cm) on the mode of delivery in nulliparous women have been studied.
We performed a search in the electronic databases Pubmed (Medline), EMBASE and the Cochrane Library until 31 July 2010. The following terms were used: ‘Analgesia, epidural’ [Mesh] AND [‘caesarean section’ [Mesh] OR ‘instrumental vaginal delivery’ OR ‘Vacuum Extraction, Obstetrical’ [Mesh]] AND [early OR timing OR ‘Labor Stage, First’ [Mesh]] Publication year or language restrictions on publication data were not applied. Cross-references of the selected studies were checked to identify other studies. Two authors (MW and JZ) independently performed the search and screened the abstracts of identified studies. Relevant abstracts were selected and full text articles were studied. Any disagreement was resolved with a third reviewer (FR) if necessary.
Study population and eligibility criteria
Women with a gestational age of at least 36 weeks, spontaneous or induced labour, with a singleton in vertex presentation were included in this review. Studies were included according to the following defined criteria:
1 RCTs and prospective and retrospective cohort studies.
2 Nulliparous women, 36 weeks or more of gestation with spontaneous or induced labour.
3 Comparison of early EA, defined as 3 cm or less, versus late EA (at least 4 cm).
4 Primary outcome was the rate of caesarean deliveries or instrumental vaginal deliveries.
Reviews and meta-analyses were excluded from analysis for this review. Only trials with a clear description of the type of analgesia used were included in the analysis. No publicly accessible protocol for this current systematic review is available. Primary outcome measure was the effect of early EA on the mode of delivery: instrumental vaginal delivery or caesarean delivery.
Assessment of methodological quality and data extraction
The systematic review was written in accordance with the PRISMA statement.11
The methodological quality of the RCTs was assessed by the Jadad criteria.12 A numerical score between zero and five is assigned as a rough measure of study reporting quality (zero being weakest and five being strongest).
Blinding of participants and caregiver as well as blinding of the outcome assessment was not possible because of the type of intervention. Therefore, the maximum score that could be given to these studies was three, according to the Jadad criteria.
Data extraction was independently completed by MW and JZ by using a data-extraction form. The following data were extracted from the selected RCTs: publication year, inclusion period, inclusion criteria, exclusion criteria, parity, gestational age, length, weight, randomisation procedure, description of cross-over, dropouts, withdrawals, power analysis, mode of delivery, indication for instrumental delivery and cervical dilatation at EA. No original or unpublished data were obtained from the investigators of the included studies.
Primary outcome measure was the effect of early EA on the mode of delivery: instrumental vaginal delivery or caesarean delivery. For each study separately, risk ratio and its 95% confidence interval were calculated for dichotomous variables. Subsequently, data of all studies were pooled using the Mantel–Haenszel method. The presence of statistical heterogeneity was determined using I² statistics. This I² describes the percentage of variation across studies that is the result of heterogeneity rather than chance. Heterogeneity was considered substantial when I² was above 50%.13,14 Confidence intervals around I² were calculated using the statistical package R.15 After exclusion of the heterogeneous studies, the pooled results were recalculated. Analysis was performed using Review Manager (RevMan) (Computer program, Version 5.0., Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2008).
The search identified a total of 327 studies including 124 duplicates. Of the remaining 203, we excluded 183 after reviewing the abstracts. The full text of the remaining 20 articles was examined in more detail. This resulted in the exclusion of another 14 articles for the reasons explained in Figure 1. Finally, six studies—five RCTs and one retrospective cohort study—appeared to be appropriate for review and fulfilled the inclusion criteria.
Description of studies and characteristics
The six selected studies involved a total of 15 399 (range 60–12 793 women per study) nulliparous women with a gestational age of at least 36 weeks and a singleton in vertex presentation. Of the selected studies, five concerned nulliparous women only.7,16–19 The retrospective cohort study of Ohel and Harats20 presented results for both nulliparous and multiparous women. Only nulliparity-related results were included in this review, however.
Description of the characteristics and interventions of the separate studies are presented in Table 1. Three RCTs included women in spontaneous labour,7,16,19 one RCT included women with induced labour,18 and one included both spontaneous and induced labour.17 The retrospective cohort study of Ohel did not report the start of labour.20
Nulliparity, ≥36 weeks of gestation, spontaneous or induced labour, analgesia request <3 cm dilatation
EA after first request: 3 ml lidocaine 2% test dose, followed by bolus of 10 ml of ropivacaine 0.2% plus 50 μg fentanyl, continuous infusion of ropivacaine 0.1% with fentanyl 2 μg/ml at 10 ml/hour rate. Boluses of 5–10 ml ropivacaine 0.2% upon request
<4 cm: intravenous pethidine and promethazine as required ≥4 cm EA, identical protocol
Ohel and Harats, 1994,20 retrospective cohort study
Nulliparity and multiparity, term, vertex position, vaginal delivery
≤3 cm continuous EA, bupivacaine 0.5%
>3 cm continuous EA bupivacaine 0.5%
In the study of Luxman et al.16 early and late EA were compared. Women who were randomised to late EA received no other form of analgesia before EA administration. Continuous EA was achieved by using bupivacaine 0.25%, 8 ml/hour.16
Wong et al.7,18 randomised women after the first request for analgesia to receive intrathecal fentanyl or systemic hydromorphone. Subsequently patient-controlled epidural analgesia was initiated in the intrathecal group at the second request and in the systemic group at a cervical dilatation of at least 4 cm or at the third request for analgesia.
In the late intervention group of the RCT from Wong et al.7 in 2005, protocol violations occurred in 11 women at the second request for analgesia and with a cervical dilatation of <4 cm. They refused the study drug and received EA. In the RCT from Wong et al. in 2009,18 35 women in the late group received EA at second request when cervical dilatation was <4 cm.
The study by Ohel et al.17 compared early EA with early opioid (pethidine and promethazine iv) and late EA (at least 4–5 cm dilatation). The EA was maintained by a continuous infusion of ropivacaine 0.1% with fentanyl 2 μg/ml at a rate of 10 ml/hour. Boluses of 5–10 ml ropivacaine 0.2% were given upon request. Significantly more women in the late group did not receive EA (13.6%) in comparison with the early group (4.5%; P = 0.0008).
Wang et al.19 randomised between early EA versus early opioid (25 mg meperidine intramuscularly) and late EA when the cervix was at least 4 cm dilated. A patient-controlled EA pump was given with a 10-ml bolus of 0.125% ropivacaine and 0.3 μg/ml sufentanil without background infusion. In the late group protocol violation occurred in 41 of the 6399 women and they received EA when the cervix was <4 cm dilated; 102 women were randomly assigned to the early group whose cervix was >4 cm dilated.
In the retrospective cohort study of Ohel and Harats20 early EA versus late EA was studied. Continuous EA was achieved by using bupivacaine 0.5%. No further details about the epidural administration were given. A similar prevalence of administration of opioids was described in the early and late EA groups.
Methodological quality and statistical heterogeneity
In three of the five RCTs women were randomised using a computer-generated random number list and an adequate description of withdrawals and dropouts was given.7,18,19 The methodological quality of the RCTs was assessed by the Jadad criteria and is summarised in Table 2. Because of slow study enrolment, the RCT by Wong et al.18 in 2009 stopped before the initial 1600 women were included; 806 participants were included in the analysis.
In the RCT of Luxman et al.16 an inappropriate method of randomisation was used because participants were divided according to their sequence of arrival into two groups.
In the RCT of Ohel et al.,17 randomisation was stratified according to the cause of labour onset (either spontaneous or induced). The RCTs of Luxman and Ohel did not report the rate of withdrawals, cross-over or dropouts.16,17
No statistical heterogeneity among the selected studies was detected for the overall rate of caesarean delivery (I2 = 0.0%, 95% CI 0.0–89.2), caesarean delivery for dystocia (I2 = 0.0%, 95% CI 0.0–88.7), caesarean delivery for other fetal status (I2 = 35.3%, 95% CI 0.0–96.0), spontaneous vaginal delivery (I2 = 0.0%, 95% CI −0.0–95.8). However, confidence intervals around the I2 estimates were wide because of the low number of studies included. Statistical heterogeneity was found for instrumental vaginal delivery (I2 = 57.8%, 95% CI 0.0–92.2). Because of the heterogeneity, the pooled result was recalculated after excluding the heterogeneous trial20 (I2 = 0.0%, 95% CI 0.0–86.9). Again, the precision of these estimates of I2 was low.
Mode of delivery
The rate of caesarean delivery (RR 1.02, 95% CI 0.96–1.08), Figure 2, instrumental vaginal delivery (RR 0.96, 95% CI 0.89–1.05) and spontaneous delivery (RR 1.01, 95% CI 0.98–1.03) was not significantly different between the early EA and the control groups (Table 3) and no significant difference was found in the indication for caesarean delivery in both groups. Because of significant heterogeneity of instrumental vaginal delivery, the pooled result was recalculated after excluding the heterogeneous trial.20 After this, still no significant difference was observed in the two groups in the rate of instrumental vaginal delivery (RR 0.94, 95% CI 0.87–1.02).
During the last few years, many studies have been published about the effects of the timing of EA on the outcome of labour. This review shows no increased pooled risk of caesarean (RR 1.02, 95% CI 0.96–1.08) or instrumental vaginal (RR 0.96, 95% CI 0.89–1.05) deliveries in nulliparous women at 36 weeks or more of gestation receiving early EA at <4 cm dilatation in comparison with EA given to women admitted when at least 4 cm dilated. The pooled risk ratio of caesarean delivery is based on the results of RCTs only.
These results are comparable with those of a systematic review about the effects of EA timing on the mode of delivery, published by Marucci et al. in 2007.8 There are, however, several differences between both studies. Marucci et al. defined early EA as a cervical dilatation <4–5 cm, whereas we used the stricter lower limit of <4 cm. These are the women suffering from severe pain in the latent first stage of labour, whose desire for regional analgesia should not be denied.21
Marucci et al.8 included five RCTs, three retrospective studies, and one impact cohort study. We excluded the data of two of these studies: one because only an abstract was available, and the second because it was an impact cohort study in which two separate time periods of labour pain management were compared, the early period (before the use of EA) with mainly usage of early parenteral opioid, and the late period with use of early EA.22,23 In the current review, we included two recent randomised trials resulting in the data including far more nulliparous women (15 399 versus 3320).24 Moreover, Marucci et al. compared early EA with early parenteral opioid, or late EA, or both early parenteral opioid in combination with late EA. We only compared early EA with late EA.
The most important limitation of this systematic review is the lack of methodological uniformity in the different studies. Different interventions in the early EA group as well as in the control group, different dose of anaesthetics and spontaneous as well as induced labours were included. However, all women allocated to the ‘late intervention’ group of the included trials, except one,16 received systemic opioid analgesia. Besides, statistical heterogeneity was not significant for the overall rate of caesarean delivery, caesarean delivery for dystocia, caesarean delivery for other fetal status and rate of spontaneous delivery. However, the broad range of 95% confidence intervals of the I2 cannot exclude statistical heterogeneity. This is because of the small number of studies included in the meta-analysis. Individual results of all RCTs showed no increased risk for caesarean delivery or instrumental vaginal delivery in the early EA group. These results suggest that it is unlikely that specific analgesia techniques are associated with an increased risk of caesarean or instrumental vaginal delivery. Only the retrospective cohort study of Ohel and Harats20 showed an increased risk of instrumental vaginal delivery (RR 1.43, 95% CI 1.10–1.87; data not shown in this review). However, in this study, traditional epidural with 0.5% bupivacaine was used. The Comparative Obstetric Mobile Epidural Trial (COMET) Study Group reported a significant increase of the rate of normal vaginal deliveries in the low-dose combined-spinal group and the low-dose infusion group compared with the traditional epidural group.25 The use of low-dose epidural techniques for labour analgesia has benefits for delivery outcome.25
In two of the included studies combined-spinal epidural was used.7,18 There is little difference between combined-spinal epidural and EA in labour despite a slightly faster onset of pain relief with combined-spinal epidural and fewer complaints of pruritus with EA. No difference in obstetric or neonatal outcome was observed.26
According to the results of this meta-analysis and the individual included studies, there is compelling evidence that earlier EA is not accompanied by an increased rate of caesarean or instrumental vaginal deliveries. Obviously, it is important to realise that the results of this review are not applicable to women who are not in labour or who have an undilated unfavourable cervix. Although all reviewed studies mentioned that they included women in (early) labour, this was only specified in the study of Wang et al.,19 who used a lower limit of cervical dilatation of 1 cm as an exclusion criterion. According to the National Institute for Clinical Excellence guidelines, early labour or the latent stage of labour is described as a period of time, not necessarily continuous, when there are painful contractions, and there is some cervical change, including cervical effacement and dilatation up to 4 cm.21
This systematic review of the literature showed no increased risk of caesarean delivery or instrumental vaginal delivery for women receiving early EA compared with late EA. Therefore, a woman’s request for EA early in labour cannot be rejected on the grounds of its presumed adverse influence on the mode of delivery. Consequently, the preference of the labouring women should be leading.
Disclosure of interests
No relevant financial, personal, political, intellectual or religious interests were disclosed.
Contribution to authorship
MW is the first and main author of the manuscript. JZ was the second reviewer; she independently screened titles and abstracts of identified studies and independently performed the data extraction using the data extraction form. FR participated as the third independent reviewer when disagreements occurred and contributed to the writing of the manuscript. LS contributed to the writing of the manuscript and the methodology and statistic co-analysis. As anaesthesiologist, MM contributed to the writing of the manuscript and the anaesthetic details. JN contributed an essential part to the writing of the manuscript.
Details of ethics approval
Not relevant for a systematic review.
There was no funding by a commercial company, charity or government department for writing this systematic review.
No specific acknowledgements have to be listed in the article.