Misoprostol for second trimester pregnancy termination in women with prior caesarean: a systematic review

Authors

  • V Berghella,

    1. Division of Maternal-Fetal Medicine, Department of Obstetrics & Gynecology, Jefferson Medical College of Thomas Jefferson University, Philadelphia, PA, USA
    Search for more papers by this author
  • J Airoldi,

    1. Division of Maternal-Fetal Medicine, Department of Obstetrics & Gynecology, Jefferson Medical College of Thomas Jefferson University, Philadelphia, PA, USA
    Search for more papers by this author
  • AM O’Neill,

    1. Division of Maternal-Fetal Medicine, Department of Obstetrics & Gynecology, Jefferson Medical College of Thomas Jefferson University, Philadelphia, PA, USA
    Search for more papers by this author
  • K Einhorn,

    1. Division of Maternal-Fetal Medicine, Department of Obstetrics & Gynecology, Jefferson Medical College of Thomas Jefferson University, Philadelphia, PA, USA
    Search for more papers by this author
  • M Hoffman

    1. Department of Obstetrics and Gynecology, Christiana Care Health System, Newark, DE, USA
    Search for more papers by this author

V Berghella, 834 Chestnut Street, Suite 400, Philadelphia, PA 19107, USA. Email vincenzo.berghella@jefferson.edu

Abstract

Background  Second trimester pregnancy induction with misoprostol in women with prior caesarean delivery is not well studied.

Objective  To estimate the risk of uterine rupture using misoprostol as an induction agent for pregnancy termination in the second trimester of pregnancy in women with prior caesarean delivery.

Search strategy  Cases of women with a history of prior caesarean delivery and subsequent misoprostol induction for pregnancy termination in the second trimester (16–28 weeks) were obtained from two main data sources. First, a retrospective chart analysis was performed at Thomas Jefferson University Hospital and Christiana Hospital between 1998 and 2004. Second, multiple Medline, Scopus and POPLINE literature searches were performed.

Selection criteria  Case series and cohort studies of women with one or more prior caesarean delivery (of any type), and with a subsequent pregnancy with induction of labour for pregnancy termination at 16–28 weeks using misoprostol as the initial primary agent were included. Case reports were analysed separately.

Data collection and analysis  Total cases were analysed by type and number of prior caesarean delivery, for the primary outcome of uterine rupture.

Main results  The incidence of uterine rupture associated with second trimester misoprostol termination was 0.4% (2/461) in women with one prior low transverse, 0% (0/46) in those with two prior low transverse and 50% (1/2) in those with a prior classical caesarean delivery. One of the cases of uterine rupture in a woman with a prior low transverse caesarean required transfusion. None of the total eight cases (including case reports) of uterine rupture was associated with hysterectomy.

Conclusions  Second trimester misoprostol termination appears safe among women with one prior low transverse caesarean birth, as it is associated with incidences of uterine rupture of 0.4% (95% confidence interval 0.08–1.67%), of hysterectomy of 0% and of transfusion of 0.2%. There are insufficient data on risk with more than one prior caesarean birth or with prior classical caesarean birth.

Introduction

The incidence of caesarean birth in the US rose again in 2007, to an all-time-high of 31.8% of all births.1 Therefore, second trimester pregnancy induction in the scenario of a prior caesarean birth has become an increasingly common occurrence. Misoprostol has been used with increasing popularity for induction, as it has been shown to be the most effective agent for second trimester induction.2 The safety of the use of misoprostol has been questioned in women with a prior caesarean birth, given the elevated risk of uterine rupture in women undergoing term or near-term induction with misoprostol.3 The data on the safety of misoprostol induction before the third trimester in women with prior caesarean birth is limited. The purpose of this study was to estimate the risk of uterine rupture using misoprostol as an inducing agent in the second trimester of pregnancy in women with prior caesarean birth.

Sources

Only women with one or more prior caesarean birth (of any type), and with a subsequent pregnancy with induction of labour for pregnancy termination in the second trimester (16–28 weeks) using misoprostol as the initial primary agent were included.

Women without prior caesarean birth, those with induction after 28 weeks, and those with other or additional agents used for initiation of induction, other than laminaria or Foley bulb, were excluded.

Two types of data sources were used: (i) a retrospective chart analysis was performed for all women with one or more prior caesarean birth, who subsequently underwent misoprostol induction in the second trimester (16–28 weeks) for pregnancy termination at Thomas Jefferson University Hospital and Christiana Hospital between 1998 and 2004; (ii) multiple Medline, Scopus and POPLINE literature searches were performed, for information regarding uterine rupture in the setting of second trimester (16–28 weeks) labour induction for pregnancy termination using misoprostol only in women with one or more prior caesarean delivery. Two independent Medline searches were performed by two investigators (VB, JA) using the search terms ‘misoprostol’, ‘second trimester abortion’, ‘pregnancy, second trimester’, labour, induced’, ‘labour induction’, ‘vaginal birth after caesarean’, ‘caesarean delivery’, ‘uterine rupture’ and combinations of these terms. A professional librarian performed independently Medline, Scopus and POPLINE literature searches using the following search strategies:

Medline: ((second or mid) adj4 trimester?) AND ruptur$ AND (cesarean$ or caesarean$ or scarred) AND (Misoprostol (MESH) OR (misoprostol$ or cytotec or misotrol)) (textwords)

Scopus: (((TITLE-ABS-KEY(mid PRE/3 trimester)) OR (TITLE-ABS-KEY(mid-trimester)) OR (TITLE-ABS-KEY(second PRE/3 trimester)) OR (TITLE-ABS-KEY(second-trimester))) AND ((TITLE-ABS-KEY(misoprostol*) AND TITLE-ABS-KEY(rupture*)))) AND ((TITLE-ABS-KEY(cesarean*) OR TITLE-ABS-KEY(caesarean*)))

POPLINE: (misoprostol & caesarean & ruptur*)

The time period was from 1966 (1981 was date of first report on misoprostol) until April, 2008. All available studies were considered. All references of each study considered for inclusion were searched for possible inclusion. We excluded studies that did not report pregnancy outcome specifically for women with prior caesarean birth. Articles published in languages other than English were translated. Both abstracts and manuscripts were included. Authors were contacted if not all information sought was available from the publication. Information sought included study design, gestational age at induction, misoprostol dosing, indication for induction, type and number of prior caesarean birth. In addition, for cases of uterine ruptures, we sought potential maternal complications such as hysterectomy and blood transfusion.

We followed the recommended Meta-analysis of Observational Studies in Epidemiology (MOOSE) guidelines for Systematic Reviews of Observational Studies.4

Study selection

All prospective and retrospective study designs were eligible for inclusion. Case reports were also selected, but analysed separately, given the absence of a denominator. Data were selected and classified by two independent data coders (VB, JA).

The primary outcome was uterine rupture. Uterine rupture was defined as disruption or tear of the uterine muscle and visceral peritoneum, or separation of the uterine muscle with extension to bladder or broad ligament.5 All uterine ruptures that occurred after induction with misoprostol as the primary and initial agent were included. Attempts were made to evaluate any other factors that might have been related to the rupture, e.g. if rupture occurred during the ‘misoprostol only’ part of the induction, versus any that might have occurred with initiation of any additional agent (e.g. oxytocin) if misoprostol was initially ineffective.

Several different analyses were planned, including assessment of the primary outcome for different types and number of prior caesarean delivery, including groups of: one prior low-transverse, two prior low transverse, three prior low transverse, more than three prior low transverse, unknown scar and prior classical (vertical) caesarean delivery. To assess for publication bias, analyses stratified by study design were planned, as suggested by the MOOSE guidelines.4 A primary analysis of the literature excluding case reports, and a secondary one including these were carried out.

Statistical methods included cumulative meta-analysis of the data using proportions of uterine rupture to type and number of prior caesarean delivery. Confidence intervals of proportions were computed using the modified Wald method.6

Results

At our institution, 17 cases with prior caesarean birth and with subsequent second trimester misoprostol induction were identified. Their characteristics are included in Table 1. Each of the 13 women with one prior low transverse incision and the two with two prior low transverse incision had uncomplicated vaginal births. Of the two women with prior classical incisions, one had an uncomplicated vaginal birth, while the other developed a uterine rupture. (Table 2) This woman had one prior classical caesarean birth at 27 weeks for an abruption. In the next pregnancy, she was induced at 22 weeks because of fetal death from an abruption. Her induction was begun with a combination of a supracervical Foley bulb and 100 μg of misoprostol. Uterine rupture was diagnosed following onset of severe abdominal pain and sudden loss of fetal station. The fetus was delivered by laparotomy and the uterine rupture was repaired. The woman required a total of 5 units packed red blood cells, but otherwise had an uneventful postoperative course.

Table 1.   Characteristics of included studies
Author (reference)Year publicationStudy designGestational age (weeks)Dose (μg)/routeInterval of dosing (hours)Live babies/total cases (%)
  1. GA, gestational age; N/A, not available; *, including third trimester cases.

Blumenthal (9)2001Retrospective cohort (abstract only)‘Second trimester’200/vaginal4–80/10 (0)
Pongsatha (10)2003Case series‘Second trimester’‘mostly’ 400/either vaginal or oral3–125/21 (24)
Rouzi (11)2003Case series‘Second trimester’ (mean GA 19.8 weeks)200/vaginal60/10 (0)
Herabutya (12)2003Prospective cohort14–26600–800/vaginal6–12N/A
Aslan (14)2004Retrospective cohort‘Mean GA 27 weeks’*50/vaginal x 2, then 100/oral4N/A
Dickinson (15)2005Retrospective cohort14–28400/vaginal6101/101 (100)
Daskalakis (16)2005Retrospective cohort17–24400/vaginal and 400/oral6 (vaginal only)99/108 (92)
Tarim (19)2005Retrospective cohort‘Second trimester’ (18.9 ± 4.4)2001N/A
Pongsatha (20)2006Case series20.5 ± 2.8400/vaginal617/17 (100)
Daponte (21)2006Case series14–20400/vaginal685/85 (100)
Bhattacharjee (22)2007Retrospective cohort13–26200–400/vaginal or sublingual664/80 (80%)
Current study2009Case series16–28100–800/vaginal4–610/17 (59%)
Table 2.   Characteristics of case reports
Author (reference)Year publicationStudy designGestational age (weeks)Dose (μg)/routeInterval of dosing (hours)Live babies/ total cases (%)
Chen (7)1999Case report23200/vaginal– (one dose)1/1 (100)
Berghahn (8)2001Case report23400/vaginal, then 400 oral61/1 (100)
Petri (13)2003Case report1720041/1 (100)
Daskalakis (17)2005Case report23400/vaginal and 400/oral8 (vaginal only)1/1 (100)
Nayki (18)2005Case report2620031/1 (100)

Systematic review revealed 55 studies possibly eligible for inclusion. The reasons for the exclusion of 39 studies are shown in Figure 1. Characteristics of the 17 included studies (including ours) are shown in Table 1 (case series and cohort studies) and Table 2 (case reports).7–22 Additional information from one study14 could not be obtained, despite numerous email and phone attempts. The cases of uterine rupture analysed by incision number and type are shown in Table 3 (case series and cohort studies).

Figure 1.

 Flow diagram of meta-analysis.

Table 3.   Incidence of uterine rupture for each included study (case series and cohort studies) and in total, also stratified by type and number of prior caesarean delivery
Author (reference)1 LTCD# Rupture2 LTCD# Rupture
  1. LTCD, low transverse caesarean delivery; N/A, not available; *, at least two cases, probably around 20.

Blumenthal (9)7000
Pongsatha (10)21000
Rouzi (11)6040
Herabutya (12)45080
Aslan (14)N/A*2N/AN/A
Dickinson (15)780190
Daskalakis (16)960120
Tarim (19)12000
Pongsatha (20)16010
Daponte (21)85000
Bhattacharjee (22)80000
Current study13020
Totals (%) [95% CI]4612 (0.43%) [0.08–1.67]460 (0%) [0.00–9.20]

Of the 461 cases with one prior low transverse caesarean birth, the uterine rupture rate was 0.43% (95% confidence interval 0.08–1.67%). (Table 3) Three further cases of uterine rupture in women with one previous caesarean birth were reported as case reports.13,17,18 Including the cases in case reports, the total rupture rate was 1.1%. The other two uterine rupture cases were reported in a retrospective cohort study14 where it was difficult to determine the denominator (number of total second trimester misoprostol inductions). Table 4 provides details on all cases of rupture. One woman received initially 50 μg of intravaginal misoprostol, followed in the next 8 hours by two doses of 100 μg oral misoprostol. Oxytocin was then began and increased in the next 3 hours to 3 mU/hour. At this point vaginal examination could not palpate any longer the presenting part, with ultrasound revealing the fetus in the abdominal cavity. Laparotomy was performed, and ‘the woman did well and went home on the third postoperative day’.14 The second woman received initially 50 μg of intravaginal misoprostol, followed 4 hours later by another dose of 50 μg intravaginal misoprostol. Oxytocin was then begun at 1 mU/hour. Six hours later uterine rupture was suspected because of severe abdominal pain and confirmed at laparotomy by a fetal shoulder protruding although the prior caesarean incision. The defect was repaired, with an uncomplicated postoperative course.14 Another three cases13,17,18 were induced between 17 and 26 weeks, receiving total doses of misoprostol ranging from 800 to 1200 μg (Table 4).

Table 4.   Details of eight cases of uterine rupture reported in the literature after second trimester misoprostol induction in women with prior caesarean delivery
Type of prior caesarean deliveryNo. prior caesarean deliveryGestational age at inductionMisoprostol total dose (μg)Misoprostol routeLaparotomyHysterectomyBlood transfusionCommentsAuthor (reference)Case report
  1. LT, low transverse; NA, not available; pRBC, packed red blood cells.

LT126250VaginalxOxytocin addedAslan (14) 
LT126100VaginalxOxytocin addedAslan (14) 
LT1171000VaginalxPetri (13)x
LT1231200Vaginal/oralxDaskalakis (17)x
LT126800Vaginalx2 units pRBCNayki (18)x
LT223200VaginalxChorioamnionitisChen (7)x
LT223800Vaginal/oralx4 units pRBCsLaminaria; D&EBerghahn (8)x
Classical122100Vaginalx5 units pRBCAbruptio; FoleyCurrent study 

No uterine rupture took place in the 46 cases with two prior low transverse caesarean deliveries. Two uterine ruptures were reported as single case reports.7,8 Both occurred in women with two prior caesarean deliveries who presented at 23 weeks. The first woman presented with chorioamnionitis and probable long-standing rupture of membranes. Uterine rupture occurred 15 hours after the single 200 μg intravaginal misoprostol insertion. The total blood loss was 200 ml, with neither transfusion or hysterectomy necessary, and an otherwise uncomplicated course.7 The second woman presented for termination of a trisomy 21 fetus. She had outpatient placement of three laminaria in the cervix, followed by her own insertion of 400 μg of intravaginal misoprostol, followed by 400 μg of buccal misoprostol 6 hours later. The uterine rupture was diagnosed after a dilatation and evacuation procedure, performed while the woman appeared stable, revealed omentum. Laparotomy revealed the fetus in the upper abdomen. The estimated blood loss was 3000 ml, requiring transfusion of 4 units of packed red blood cells. The uterus was repaired, not necessitating hysterectomy, and no coagulopathy developed.8

Of the seven cases with three prior low transverse caesarean delivery, none had uterine rupture. Of the two cases with a prior classical caesarean delivery, one, previously described, had a rupture. None of all the eight cases of uterine rupture was associated with maternal mortality. Five7,8,13,17,18 of the eight cases of uterine rupture occurred during induction of live fetuses.

Conclusion

The incidence of uterine ruptures among women with one prior low transverse caesarean birth and subsequent second trimester misoprostol termination was 0.43%. The addition of ruptures reported only as cases increases the rate to 1.1%. Three of the five total uterine ruptures occurred at 26 weeks, of which two had oxytocin added to the initial misoprostol dosing. Both cases which occurred before 26 weeks were associated with total misoprostol dosing ≥1000 μg. None of the five ruptures necessitated hysterectomy, and only 1 (20%) a blood transfusion.

No uterine rupture was reported in cohort studies or series of women with two prior low transverse incisions, although two individual cases were reported. Both of these cases had risk factors associated chorioamnionitis in one,7 while the other occurred during curettage.8 There are only seven reported cases where the woman had three prior caesarean births12,15 and three cases where the type of caesarean birth was not specified;9 no ruptures were reported. There are, therefore, insufficient data to comment on the safety of second trimester termination in women with two or more low-transverse caesarean births, or with prior classical incisions. Larger studies are still needed to determine the true incidence of uterine rupture in these patient populations.

A woman with a prior caesarean birth has three options for mode of birth in the subsequent pregnancy if termination is desired or indicated in the second trimester: a medically induced trial of labour to try to achieve a vaginal birth after caesarean; a repeat caesarean birth or a dilatation and extraction (or D&E, i.e. D&X). There are no randomised trials comparing the safety (maternal and fetal/neonatal morbidity and mortality) or the effectiveness of these options.

Regarding safety of trial of labour after caesarean birth, which was the primary outcome of interest in our study, uterine rupture is the main complication. The risk of uterine rupture with a policy of trial of labour at term after caesarean delivery depends on several factors including number of prior caesarean births, prior classical (vertical) scar, prior rupture, induction or augmentation, fetal macrosomia, and possibly interval between delivery <18 months, maternal age >30 year old and fever around prior caesarean delivery.23 The scar rupture rate during VBAC where there is spontaneous onset of labour is 0.4–0.5%.24,25 In term or third trimester women with a prior caesarean birth and with a misoprostol induction, a 5.6% incidence of uterine rupture has been reported.26 Misoprostol is contraindicated for induction of labour in women after 28 weeks of gestation with prior caesarean birth.

In second trimester termination in women with prior caesarean birth using other induction agents, the incidence of uterine rupture in 79 women with one prior low transverse caesarean delivery and second trimester induction with agents such as prostaglandin E2, oxytocin, or both, was 3.8%, with an 11.4% incidence of related blood transfusion.27 Uterine rupture has been reported also in similar women induced with prostaglandin F2a and intra-amniotic urea.28

Given these results, misoprostol appears a safe induction agent for women with one prior low transverse caesarean requiring second trimester termination, although the use of oxytocin as an additional agent for labour augmentation in these women should be minimised.28

The option of advanced D&E or D&X has not been compared with medical induction in women with a prior caesarean delivery and subsequent need for second trimester termination. D&E or D&X in the second trimester are safe only when performed by experienced operators and are, therefore, not widely available.29,30 The option of laparotomy for women with a prior caesarean delivery needing termination in the second trimester for a non-viable or unwanted pregnancy means a repeat caesarean delivery, which is associated with significant maternal complications,5 and can be avoided with the use of misoprostol.

Strengths of our study include the completeness of the systematic review, done through multiple individual searches, including professional ones. We also followed the recommended QUOROM guidelines.4 As all systematic review of observational studies, a weakness is publication bias. Most uterine ruptures were reported in case reports, which increased the incidence of this complication when they were included in our analyses. It is difficult to judge publication bias as both preferential reporting of adverse outcomes or of good outcomes could have occurred.

While no method is risk-free, misoprostol for second trimester termination appears safe and effective at least for women with one prior low transverse caesarean birth. Regarding maternal risks from uterine rupture in women with any prior caesarean birth and subsequent second trimester induction with misoprostol, this report found minimal maternal morbidity (need for laparotomy, risk of blood transfusion and need for hysterectomy) and no maternal mortality. Second trimester termination is carried out for either fetal death or voluntary termination, so that fetal effects of uterine rupture do not usually play a role in clinical management. Therefore, there is an argument for its use where there has been more than one prior caesarean birth, as this review found no cases of rupture in nearly 50 inductions with misoprostol, and no serious morbidity even when case reports were included. Appropriate counseling including risks as described should be provided to the woman with a prior caesarean birth deciding on subsequent mode of birth. The woman has the ultimate choice.

To minimise risks, as uterine rupture can occur, an experienced obstetrician, anesthesia, nursing and operating personnel should be immediately available at all times.

Disclosure of interest

None pertinent.

Contribution to authorship

Vincenzo Berghella conceived the study, and contributed to every aspect, including study design, data retrieval, analysis of data, writing/editing as well as final approval of manuscript. James Airoldi contributed to study design, data retrieval, analysis of data and writing/editing as well as final approval of manuscript. Anna Marie O’Neill, Kim Einhorn and Matthew Hoffman contributed to data retrieval, analysis of data and editing as well as final approval of manuscript.

Thomas Jefferson University Institutional Review Board (Ethic) Approval #04U-415.

Funding

None.

Acknowledgements

The librarian Karen Krasznavolgyi for the professional literature searches.

Ancillary