Uterine rupture after induction of labour in women with previous caesarean section

Authors


Professor Z. Alfirevic, Liverpool Women's Hospital, Crown Street, Liverpool, L8 7SS, UK.

Abstract

Background  The issues related to safety of induction of labour in women with previous caesarean section remain controversial. The main adverse outcome fuelling this debate is a ‘small’ risk of uterine rupture that is potentially devastating for both the mother and the fetus.

Objective  To estimate the risk of uterine rupture or dehiscence in women who require induction of labour with previous caesarean sections.

Design  Five year retrospective review of computerised hospital records and case note review of index cases.

Setting  Large inner city teaching hospital.

Population  Two hundred and five women who had their labour induced with history of one lower segment caesarean section.

Methods  This study was conducted at Liverpool Women's Hospital, a tertiary referral centre, with approximately 6000 births per annum. We searched the hospital's computerised records of deliveries from June 1997 to June 2002 and reviewed all indications and outcomes of induction of labour in women with one previous caesarean section. Women with singleton pregnancy and cephalic presentation were then divided into three groups: those with one previous caesarean section and no previous vaginal deliveries, those whose last delivery was a caesarean section but had delivered vaginally before and those whose last delivery was by vaginal route, but had had one caesarean section in the past.

Main outcome measures  Uterine rupture or dehiscence, adverse neonatal outcome.

Results  Two hundred and five women were included. There were four cases of uterine rupture and one dehiscence (2.4%, 95% CI 0.8–5.6%). Two babies were profoundly acidotic at birth, but all five neonates were healthy when discharged from hospital with no long term morbidity. All five cases occurred in the group of women with no previous vaginal deliveries. The intrauterine pressure catheter recordings had contributed to the diagnosis of uterine rupture/dehiscence in three out of five cases.

Conclusion  In women with previous caesarean section and no vaginal deliveries, induction of labour carries a relatively high risk of uterine rupture/dehiscence despite all precautions, including intrauterine pressure monitoring.

INTRODUCTION

The Clinical Audit Unit of RCOG (1997)1 has recommended that women who have previously had a caesarean section should be actively considered for subsequent vaginal delivery. This advice is based on the meta-analyses that have examined indicators for the success of trial of labour and the morbidity and mortality data associated with vaginal birth after caesarean section. It is estimated that if women with a previous caesarean section scar are permitted to labour, 60–80% will achieve vaginal delivery. However, in the UK, more than 20% of pregnancies are being induced since the National Institute of Clinical Excellence/Royal College of Obstetricians and Gynaecologists guidelines2 (NICE/RCOG) suggested that induction of labour should be offered to all women after 41 weeks. As a consequence, clinicians are increasingly being faced with the need to induce labour in women who have undergone a previous caesarean section.

The issues related to the safety of induction of labour in women who have undergone a previous caesarean section remain controversial. The main adverse outcome fuelling this debate is a ‘small’ risk of uterine rupture that is potentially devastating for both the mother and the fetus. In 1984, MacKenzie et al.3 were the first to report a case series of 143 women with previous caesarean section who were induced with prostaglandins claiming that this intervention reduced the need for repeat caesarean section without evidence of undue risk of lower segment scar rupture. In 1999, Vause and Macintosh4 reviewed the literature regarding the use of prostaglandins to induce labour in women with previous caesarean sections that had accumulated since the report by MacKenzie et al. Having identified six observational study and one small randomised study, they concluded that there is a dearth of evidence from which to assess the risks and benefits of prostaglandin induction in women with previous caesarean section. Their best estimate of symptomatic rupture risk was no greater than 0.6%.

In 2001, Lydon-Rochelle et al.5 published a large population-based retrospective cohort study comparing the risk of uterine rupture in women with previous caesarean section and found that the risk of uterine rupture was the highest after induction of labour with prostaglandins (2.4%) compared with only 0.5% in women with spontaneous onset of labour. One of the concerns was the use of relatively high doses of misoprostol for induction of labour during the study period, which may have contributed to such an unexpected high rate of uterine ruptures. Unfortunately, clinical details, including type of prostaglandin used and management during labour, could not be determined.

Our hospital has a fairly liberal policy for induction of labour in women with previous caesarean section. Vaginal prostaglandins (dinoprostone gel) have been used routinely with continuous fetal monitoring and insertion of intrauterine pressure catheter when possible. In the light of the reports described above, we reviewed the effectiveness and safety of our current management protocol.

METHODS

This study was conducted at Liverpool Women's Hospital, which is a tertiary referral centre with approximately 6000 births per annum. We searched the hospital's electronic patient records from June 1997 to June 2002 and reviewed all indications and outcomes of induction of labour in women with one previous caesarean section. We included only women who had one lower segment uterine scar and singleton pregnancy with cephalic presentation. We excluded women with the history of hysterotomy, uterine rupture and inverted T uterine incision.

Where the rupture of uterus or dehiscence of scar was established, the notes were examined for details and detailed information were transferred to a data sheet. Uterine rupture was defined as an intra-operative finding of fetal parts within the abdominal cavity. Dehiscence was defined as a window in the lower segment with either membranes bulging or parts of the baby visualised through it. In order to ensure that all cases of uterine rupture/dehiscence were captured, we searched Adverse Clinical Event Reports, ongoing and previous legal cases, complaints and records from perinatal mortality/morbidity meetings within the study period.

All women with unfavourable cervix, irrespective of membrane status, were induced with 1 mg dinoprostone gel (Prostin); this was repeated only once 6 hours later if there was no cervical change. If artificial rupture of membranes was possible, oxytocin was commenced at 2 Montevideo units (MU)/minute and doubled every 30 minutes to the maximum dose of 16 mu/minute. The guidelines stipulated that an intrauterine pressure catheter should be used with the syntocinon infusion. Montevideo units (MU) were calculated in 10-minute intervals, aiming for maximum values of 160 MU. Poor uterine activity was considered if contractions less than 120 MU and oxytocin commenced or increased. The woman was reassessed 2 hours following the start of infusion. If uterine activity was adequate (>120 MU) with inadequate progress of labour, a caesarean section was recommended.

Eligible women were divided into three groups: those with one only previous delivery by caesarean section, those whose last delivery was a caesarean section but had delivered vaginally before and the third group whose last delivery was by vaginal route but had had one caesarean section in the past.

RESULTS

A total of 29,677 deliveries were recorded in the study period with 205 women who met our inclusion criteria (0.7%) (Table 1). Overall, 125 women delivered vaginally (61%) and 80 women by caesarean section (39%).

Table 1.  Maternal and labour characteristics of 205 women who had induction of labour after one previous caesarean section.
 PGE2 only (n= 97)PGE2 and oxytocin (n= 52)ARM only (n= 11)ARM and oxytocin (n= 45)
  1. ARM = artificial rupture of membranes; PROM = prelabour rupture of membranes; CS = caesarean section; IUPC = intrauterine pressure catheter.

Indications
Post-term4520411
PROM7807
Hypertension/pre-eclampsia8512
Maternal reasons1610212
Poor obstetric history6211
Bleeding0105
Fetal reasons15637
 
Obstetric history
No vaginal deliveries5332220
Vaginal delivery followed by CS1910315
CS followed by vaginal delivery2510610
IUPC2510610

Women with no previous vaginal deliveries were most likely to have failed induction. Out of 107 women in this group, only 44 delivered vaginally (41%) compared with 81/98 of women who had had a previous vaginal delivery (83%) (odds ratio 6.8; 95% CI 3.4–13.9) (Table 2).

Table 2.  Mode of delivery and perinatal outcome for women who were induced with previous caesarean section.
 PGE2 only (n= 97)PGE2 and oxytocin (n= 52)ARM only (n= 11)ARM and oxytocin (n= 45)
  • ARM = artificial rupture of membranes; CS = caesarean section.

  • *

    Adverse neonatal outcome defined as seizures, death, admission to NICU, Apgar <7 at 5 minutes, pH <7.05.

  • **

    Apgar <5 at 5 minutes.

  • pH < 7.05.

  • ††

    Admission to Neonatal Intensive Care Unit.

Spontaneous vaginal delivery45 (47%)20 (38.5%)8 (73%)28 (62%)
No vaginal deliveries previous14307
Vaginal delivery followed by CS99212
CS followed by vaginal delivery22869
 
Instrumental vaginal delivery10 (10%)8 (15.5%)06 (13.5%)
No vaginal deliveries6806
Vaginal delivery followed by CS4000
CS followed by vaginal delivery0000
 
Caesarean section42 (43%)24 (46%)3 (27%)11 (24.5%)
No vaginal deliveries332127
Vaginal delivery followed by CS6113
CS followed by vaginal delivery3201
 
Uterine dehiscence0001 (2%)
No vaginal deliveries0001
Vaginal delivery followed by CS0000
CS followed by vaginal delivery0000
 
Uterine rupture1 (1%)2 (4%)01 (2%)
No vaginal deliveries12††01
Vaginal delivery followed by CS0000
CS followed by vaginal delivery0000
 
Adverse neonatal outcome*0121
No vaginal deliveries01**21††
Vaginal delivery followed by CS0000
CS followed by vaginal delivery0000

Women induced with prostaglandins had unfavourable cervices and, as expected, a higher rate of caesarean sections compared with women induced only with artificial rupture of membranes and oxytocin (43%vs 25%; odds ratio 2.3, 95% CI 1.1–4.9). Overall use of intrauterine pressure catheter was lower than anticipated (25%) with only 19% use in the highest risk group (prostaglandin and syntocinon).

There were four cases of uterine rupture and one of dehiscence (2.4%, 95% CI 0.8–5.6%). Two babies were profoundly acidotic at birth, but all five neonates were healthy when discharged from hospital with no long term morbidity (follow up two to six years) (Table 3). Four women had an uneventful post-operative recovery and were discharged by day five. One woman had problems with recurrent urinary retention and was discharged home with a urinary indwelling catheter. The catheter was removed three days later and follow up urodynamic studies were normal.

Table 3.  Clinical characteristic of five women with uterine rupture/dehiscence.
 WeeksType of inductionLabour (hours)Oxytocin (hours)IUPCCTGAnalgesiaRupture suspectedNeonatal outcome
  1. ARM = artificial rupture of membranes; CS = caesarean section; IUPC = intrauterine pressure catheter; CTG = cardiotocography; NICU = Admission to Neonatal Intensive Care Unit.

Rupture
Case 134PGE2 and oxytocin1515YesNormalEpiduralNoNormal at 4 years of age
Case 240PGE2 and oxytocin66YesAbnormalEpiduralNoCord pH 6.83, NICU, Normal at 4 years of age
Case 340PGE24.5YesAbnormalEpiduralYesCord pH 6.8, Normal at 2 years of age
Case 440ARM and oxytocin163YesAbnormalEpiduralYesApgar 6 at 5 minutes, Normal at 6 years of age
 
Dehiscence
Case 140ARM and oxytocin106YesAbnormalEntonox/OpiodsYesNormal at 6 years of age

All five cases of uterine rupture/dehiscence occurred in the group of 107 women with no previous vaginal deliveries (4.7%). The intrauterine pressure catheter recordings have contributed to the diagnosis of uterine rupture/dehiscence in three out five cases (Table 3), with no false positive recordings. During the same five-year period, there were 720 women who had one previous caesarean section and went into spontaneous labour with only two cases of dehiscence (0.3%) and no uterine ruptures.

DISCUSSION

Our data show that, despite strict guidelines and liberal use of intrauterine pressure catheters, uterine rupture/dehiscence is a relatively common incident during induction of labour after caesarean section (2.4%). All cases of uterine rupture/dehiscence happened when prostaglandins and/or oxytocin were used in women without previous vaginal delivery. In fact, in this subgroup, the incidence of uterine rupture was 4.7%. We are conscious that confidence intervals around these figures are wide (1.6–10.8%) and should be interpreted with caution.

The overall incidence of uterine rupture following induction of labour in literature varies from 0.2% to 9%.6 However, external validity (generalisibility) of these estimates is limited because of the diverse sample of women, physician preferences, classification of rupture/dehiscence and varied management protocols.

Interestingly, most published studies have concentrated on the type of induction agent as a predictive factor for a successful vaginal delivery. In order to explore the impact of previous obstetric history on the risk of uterine rupture, we identified 16 relevant published case series using the same search strategy as Vausse and Macintosh4(Table 4). Unfortunately, only two other studies provided data for women who had had a successful vaginal delivery in the past. It was therefore not possible to confirm our observation that induction of labour after caesarean section carries a much greater risk for uterine rupture for women with no previous vaginal deliveries compared with women who delivered vaginally in the past. Additional factors may also be important including suturing technique to repair the lower segment,20 post-operative complications,21 time interval between first caesarean section and planned delivery,22 size of fetus23 and morbid maternal obesity.24

Table 4.  Published case series of induction of labour after caesarean section with data on uterine rupture.
Uterine ruptures* after labour induction with previous caesarean section
StudyNo previous vaginal deliveriesPrevious vaginal deliveriesCombined
  • *

    Cases of uterine dehiscence were excluded.

MacKenzie7Not statedNot stated1/439 (0.2%)
Goldberger et al.8Not statedNot stated0/19
Blanco et al.9Not statedNot stated0/25
Norman and Ekman100/260/40/30
Taylor et al.111/42No data1/42 (2.3%)
Leung et al.12Not statedNot stated70/8513 (0.8%)
Stone et al.130/94Excluded0/94
Williams et al.140/117Excluded0/117
Flamm et al.15Not statedNot stated6/453 (1.3%)
Zelop et al.1613/560 (2.3%)Excluded13/560 (2.3%)
McNally and Turner171/51 (2%)1/52 (1.9%)2/103 (1.9%)
Ravasia et al.188/575 (1.4%)Excluded8/575 (1.4%)
Blanchette et al.19Not statedNot stated12/754 (1.6%)
Lyndon-Rochelle et al.5Not statedNot stated24/2326 (1%)
Taylor et al.25Not statedNot stated14/790 (1.8%)
Delaney and Young26Not statedNot stated6/803 (0.7%)
Our series4/107 (3.7%)0/984/205 (2%)

We remain uncertain about the ‘protective’ role of an intrauterine pressure catheter in this clinical scenario. Rodriguez et al.27 found that an increase in baseline intrauterine pressure was observed in 4 out of 39 patients who had rupture of the uterus with an intrauterine pressure catheter in situ. This increase in baseline uterine pressure was associated with severe variable decelerations such that, by itself, the intrauterine pressure catheter added little to the diagnosis of uterine rupture. Despite strict guidelines advocating routine use, in our cohort intrauterine pressure catheter was used only in 25% of cases. The review of the clinical records failed to explain the reasons for the clinicians' reluctance to use intrauterine pressure catheter. Although, in two instances uterine rupture occurred despite intrauterine pressure catheter, one can argue that in three other cases intrauterine pressure catheter contributed to the timely diagnosis, prompt delivery and good perinatal outcome.

Having reviewed our data we are left with the following question: What is an acceptable risk of uterine rupture on a modern delivery suite? We believe that the risk for women with previous vaginal delivery is likely to be sufficiently low (<1%) for induction of labour to be the preferred option. Women with no previous vaginal delivery have to be counselled very carefully. The risk of uterine rupture is likely to be between 1% and 5% and is unlikely to be reduced by the use of modern technology. Although most of uterine ruptures will have a happy outcome for mother and baby, for women with no previous vaginal delivery who require induction of labour, caesarean section may be the preferred option.

Accepted 15 June 2004

Ancillary