- Top of page
Objective Recent studies have shown that among women with uterine scars from previous caesarean section of any type, induction of labour is associated with increased risk of uterine rupture compared with spontaneous labour. We have assessed the risk of uterine rupture in a cohort of women with a previous low transverse caesarean section in whom induction and management of labour were performed according to a strict protocol.
Setting University Hospital.
Population All women with a singleton pregnancy and a previous low transverse caesarean section requiring induction of labour from 1/1/1992 to 12/30/2001 (n= 310) were compared with a control cohort during the same study period constituted of women with a previous low transverse caesarean section in spontaneous labour (n= 1011).
Methods Clinical characteristics and rate of uterine rupture of women with previous caesarean section undergoing induction of labour were compared with those of women with previous caesarean section in spontaneous labour.
Main outcome measure Incidence of uterine rupture.
Results Uterine rupture occurred in 0.3% in the previous caesarean section—induction group versus 0.3% in the previous caesarean section—spontaneous labour group (P= 0.9). Logistic regression analysis showed no significant difference in the rate of uterine rupture between the induction and spontaneous labour group (P= 0.67) after controlling for maternal age, parity, duration of labour, gestational age at delivery and birthweight.
Conclusion Among women with a previous low transverse caesarean section, induction of labour is not associated with significantly higher rates of uterine rupture compared with spontaneous labour, provided a consistent protocol with strict criteria for intervention is adopted.
- Top of page
The epidemic of caesarean birth, first originated in the United States, has subsequently extended to Europe, with rates that have doubled in the past 10 years. The main indication for caesarean section has become repeat caesarean section. Documentation that the rate of uterine rupture is only marginally increased among women undergoing a trial of labour than among those undergoing an elective repeat caesarean section (0.4%vs 0.2%, OR = 2.1), with a significant decrease in need for transfusion (OR = 0.57) or hysterectomy (OR 0.39),1 has led authorities to encourage vaginal births after caesarean.2 However, nearly a quarter of women who are candidates for a trial of labour require induction of labour. In 2001, an American population-based retrospective cohort analysis on over 20,000 women3 reported that the risk of uterine rupture in women with one prior caesarean section undergoing induction of labour was significantly higher than among those with spontaneous onset of labour (0.7%vs 0.5%), particularly if induction was achieved with use of a prostaglandin (2.4%). However, the study did not provide information about the specific types and dosages of prostaglandin used. The authors concluded that the overall effect of induction of labour with prostaglandin on uterine rupture is still unclear and may vary according to the preparation used, the regimen and the degree of cervical ripeness for induction. An independent group of investigators4 confirmed that among women with a previously scarred uterus, induction of labour is associated with an increased risk of uterine rupture compared with spontaneous labour (2.3%vs 0.7%, P= 0.001). However, logistic regression analysis showed that the risk was greater with use of oxytocin induction (OR = 4.6, 95% CI 1.5–14.1) than with prostaglandin E2 gel (OR = 3.2, 95% CI 0.9–10.9) or oxytocin augmentation (OR = 2.3, 95% CI 0.8–7.0), the last two methods not achieving statistical significance.
Despite the inconsistencies in the published series, their results made the pendulum swing away from induction of labour in women with previous caesarean section, with a consequent rise in rates of caesarean section. The appropriate management protocols for vaginal birth after caesarean is now a major area of obstetric controversy. Different methods of labour induction may play an important role in the risk of uterine rupture, as may the timing of emergency caesarean section for failure to progress in labour.5
The objective of this study was to assess the risk of uterine rupture following induction of labour in women with a uterine scar from a previous low transverse caesarean section in a setting where induction of labour was performed according to a strict protocol for cervical ripening and subsequent management of labour, and with rigid criteria for definition of labour dystocia.
- Top of page
We reviewed the obstetric database containing information on all deliveries at ≥24 weeks at the University Hospital San Gerardo, University of Milano-Bicocca, from January 1992 to December 2001. During the study period, there were 27,179 deliveries at the Institution. Included in the study were all singleton pregnancies with a live fetus admitted for induction of labour (n= 310) or spontaneous labour (n= 1011) (Fig. 1). A history of one previous low transverse caesarean section was present in 1296, two previous caesarean sections in 24, and three previous caesarean sections in 1. The rates of multiple prior caesarean sections were similar in the two groups. During the same period, there were 446 elective caesarean sections in pre-caesarean section patients; the indications are displayed in Table 1.
Table 1. Indications for elective repeat caesarean section. Values are presented as n (%).
|Previous multiple caesarean delivery||102 (22.8)|
|Maternal request||56 (12.5)|
|Non-reassuring fetal assessment||70 (15.7)|
|Maternal indications||66 (14.8)|
|Non-cephalic presentation||60 (13.4)|
|Placental abruption or placenta praevia||29 (6.5)|
|Previous caesarean delivery with unknown or classical scar||25 (5.6)|
|Fetal macrosomia||11 (2.5)|
|Suspected fetopelvic disproportion||11 (2.5)|
|Fetal anomalies||4 (0.9)|
Trial of labour was not attempted in women with previous uterine scar from classical caesarean section, and induction with prostaglandins was not attempted in women with a history of two or more caesarean sections (n= 2) or uterine surgery with entry into the cavity. In women with previous caesarean section, closure of the uterine scar was in a double layer in approximately 50% of cases, the remainder had been closed in a single layer.
Labour was induced according to a consistent protocol, which differed from that of women with an unscarred uterus (Fig. 2). In women with a uterine scar from a previous caesarean section, cervical ripening was achieved with intracervical application of prostaglandin E2 gel 0.5 mg every 8 hours or intravaginal application of prostaglandin E2 gel 1 mg every 12 hours until a Bishop score >6 or for a maximum of four doses. If the cervix remained unripe after four doses, a caesarean section was performed. When the Bishop score was >6, membranes were ruptured; oxytocin infusion was started 2 hours later if labour was not established. Augmentation of labour was allowed only if cervical dilatation was >3 cm and the presenting part was engaged. Oxytocin infusion for induction or augmentation of labour was started at 1.3 to 2.6 mU/minute and was doubled every 40 minutes until an adequate labour pattern was achieved or up to a maximum of 42 mU/minute. The oxytocin infusion was decreased by half in the presence of uterine hyperstimulation, defined as uterine contractions more frequent than every 2 minutes for 15 minutes. Vaginal examinations were repeated every 2 hours during the active phase of labour, defined as a cervical dilation of at least 4 cm. When cervical dilatation was arrested for 2 hours or progressed at a rate below the 5th centile of the Friedman curve for more than 2 hours, despite appropriate augmentation of labour with amniotomy and/or oxytocin, a caesarean section was performed. The decision to perform an induction or to begin oxytocin infusion for augmentation of labour in patients with a scarred uterus was always taken in consultation with a senior obstetrician. Misoprostol was never used.
The use of epidural analgesia and intrauterine pressure catheters was negligible (<1%) in the overall population. In our labour ward, a midwife is assigned for support to the labouring woman in a one-to-one relationship and non-pharmacological methods of relief of pain during labour are employed, including maternal movement and positioning, touch and massage, and baths.
The main outcome variable was uterine rupture, defined as full-thickness separation of the uterine wall associated with at least one of the following: haemorrhage, defined as presence of any amount of intraperitoneal blood at laparotomy; extrusion of placenta or fetal parts; or severe fetal heart rate abnormalities, defined as stage 3 according to the classification of Boylan6 (tachycardia >180 beats/minute, reduced variability, bradycardia <100 beats/minute lasting longer than 10 minutes, repetitive late decelerations or severe variable decelerations). All charts coded as uterine rupture or symptomatic uterine dehiscence were reviewed to determine whether the rupture had occurred according to the above definition. Cases of asymptomatic dehiscence were excluded from the analysis.
The clinical characteristics and the rate of uterine rupture in women with a previous caesarean section and induction were compared with those of women with a previous caesarean section in spontaneous labour using Fisher's exact test and χ2 test for categorical data and Student's t test and one-way analysis of variance for numerical data (SPSS 11.5, SPSS, Chicago, Illinois). Logistic regression analysis was performed to assess whether the difference in rate of uterine rupture between the two groups was independent of other confounders. A P < 0.05 or an odds ratio (OR) with 95% confidence interval (CI) not inclusive of unity was considered significant.
- Top of page
A total of 310 women with uterine scar from previous low transverse caesarean section and undergoing induction of labour fulfilled the study criteria; their characteristics and obstetric outcome were compared with those of women with a previous caesarean section who went into labour spontaneously (Table 2).
Table 2. Population characteristics among women with previous low transverse caesarean section. Values are presented as mean [standard deviation] or n (%).
|Induction of labor (n= 310)||Spontaneous labor (n= 1011)|
|No previous vaginal delivery||232 (74.8)||790 (78.1)||0.3|
|Maternal age (years)||33.1 [5.6]||32.7 [4.0]||0.1|
|Gestational age (weeks)||39.3 [2.1]||39.2 [1.5]||0.3|
|Premature rupture of membranes||73 (23.5)||380 (37.6)||<0.001|
|Birthweight (g)||3178 ||3284 ||0.001|
|Apgar score <7 at 5 minutes||1/302 (0.3)||6/1006 (0.6)||0.9|
|Umbilical artery pH <7||0/288 (0)||3/905 (0.3)||0.9|
The mean (standard deviation, SD) Bishop score at induction was 4.1 (2.2), prostaglandin gel was used for cervical ripening in 240 cases (77.4%), in which the number of prostaglandin gel administrations was 1.8 (1.1) per patient (median 1 dose, range 1–4). The indications for induction of labour were post-term pregnancy (n= 52), hypertension (n= 25), diabetes (n= 9), gestational cholestasis (n= 17), chorioamnionitis (n= 2), fetal growth restriction (n= 14), oligohydramnios (n= 73), meconium-stained fluid at amnioscopy (n= 22), polyhydramnios (n= 4), Rhesus isoimmunisation (n= 2), non-reassuring fetal assessment (n= 6), premature rupture of membranes >24 hours (n= 60), abnormal vaginal bleeding (n= 2) and others (n= 22).
Women undergoing induction of labour had a significantly lower rate of premature rupture of membranes and a lower birthweight. As expected, they were three times more likely to receive oxytocin administration, had a longer duration of labour and significantly lower rates of vaginal delivery (Table 3). The rates of uterine rupture were not significantly different between the two groups. Logistic regression analysis showed that the lack of a significant association between induction of labour and uterine rupture persisted after controlling for maternal age, gestational age at delivery, previous vaginal delivery, duration of labour, use of oxytocin and birthweight (P= 0.67, OR = 0.56, 95% CI 0.04–8.11). Power analysis revealed that our sample population had an 80% power (α= 0.05) to detect a difference of at least 2% in the rupture rates between the two groups.
Table 3. Labor and delivery characteristics among women with previous caesarean section. Values are presented as mean [standard deviation] (range) or n (%).
| ||Induction of labour (n= 310)||Spontaneous labour (n= 1011)||P value or OR (95% CI)|
|Oxytocin for induction or augmentation||124 (40.0)||169 (16.7)||3.3 (2.5–4.4)|
|Duration of active labour (hours)||4.26 [4.34] (0.38–11.6)||3.49 [2.20] (0.33–14.6)||0.002|
|Uterine rupture||1 (0.3)||3 (0.3)||1.1 (0.1–11.7)|
|Blood loss (mL)||354 ||349 ||0.7|
|Total||220 (71.0)||825 (81.6)||0.6 (0.4–0.7)|
|Spontaneous||208 (67.1)||788 (77.9)||0.6 (0.4–0.8)|
|Operative||12 (3.9)||37 (3.6)||1.1 (0.5–2.1)|
|Total||90 (29.0)||186 (18.4)||1.8 (1.3–2.5)|
|For dystocia||19 (6.1)||61 (6.0)||1.1 (0.6–1.8)|
|For non-reassuring fetal testing||28 (9.0)||56 (5.5)||1.7 (1.1–2.8)|
|Other indications||43 (13.9)||69 (6.8)||2.2 (1.4–3.4)|
There were four cases of uterine rupture among women with previous caesarean section. One occurred following induction of labour in a woman with a previous caesarean section for breech presentation at 40 weeks performed three years earlier with uterine closure in a single layer. Induction was accomplished with one dose of intracervical prostaglandin E2 at 42 weeks. Uterine rupture was diagnosed by non-reassuring fetal heart rate tracing before the onset of labour. The remaining three cases of rupture followed a spontaneous onset of labour. One occurred at 6 cm of dilatation during oxytocin augmentation in a woman with two previous caesarean sections at 30 and 41 weeks (four and two years earlier), with uterine closure in two layers at both caesarean sections. Of note, both previous caesarean sections were complicated by puerperal infection. The second rupture occurred at 4 cm of dilatation during oxytocin augmentation in a woman with a previous caesarean section at 40 weeks performed three years earlier. The third rupture occurred at 8 cm of dilatation without oxytocin augmentation in a patient with clinical chorioamnionitis and a history of previous caesarean section at 40 weeks (two years earlier) with uterine closure in a single layer. Maternal outcome was favourable in all four cases of uterine rupture, none of which required hysterectomy. All four neonates had 5-minute Apgar scores of 7 or above.
- Top of page
We have found that among women with previous low transverse caesarean section, induction of labour can be associated with similar rates of uterine rupture as occur in spontaneous labour, provided a specific protocol for induction of labour is implemented. This protocol differed from the one used for induction of labour in a woman with an intact uterus in several important aspects. Oxytocin was not administered on an unripe cervix; we allowed a longer time for cervical ripening (up to 40 hours vs 28 hours among women with unscarred uterus) and a longer interval between subsequent prostaglandin gel administrations (12 hours vs 6 hours). We adopted a strict definition of labour dystocia in the group with a scarred uterus and the average duration of active labour in our series is lower than in other reported studies. Labour was augmented only if cervical dilatation was >3 cm and the presenting part was engaged. Misoprostol was never used for induction of labour, and prostaglandins were not used in women with multiple uterine scars, multiple gestations or uterine surgery other than low transverse caesarean section. Thus, our conclusions cannot be extrapolated to inductions of labour conducted with other protocols.
Our findings differ from those of Lydon-Rochelle et al.,3 who found increased rates of uterine rupture following induction of labour. However, their study used the International Classification of Diseases (ICD) discharge codes to classify women into groups and to evaluate the risks of obstetric complications, including uterine rupture. Because the hospital records were not reviewed, the authors could not confirm that the diagnostic code was correct, nor could they establish if they were dealing with uterine rupture or scar dehiscence. ICD coding on hospital discharge forms has been shown to be correct in only 72% of cases of induction of labour7 and in an unknown proportion of cases with uterine rupture.8 The lack of reliability of ICD codes in that study was supported by the authors' report that 267 women in the trial of labour group had breech presentation and 44 had placenta praevia. Both conditions are generally considered contraindications for vaginal delivery, particularly in women with previous uterine scar. In our study, all charts were reviewed for accuracy. Lydon-Rochelle et al.3 also reported significantly higher rates of uterine rupture following induction of labour with prostaglandin. Once again, because the use of prostaglandin was based upon ICD codes, the authors could not verify the accuracy of the statement or the type of prostaglandin used or the interval between prostaglandin administration and delivery. In our study, prostaglandin use, at the dosage and intervals described, was not associated with a significant increase in rate of uterine rupture.
Our findings are consistent with those of other studies,9,10 which reported no significant difference in rates of uterine rupture between women with previous caesarean section who were induced and those allowed to spontaneously labour. Although only few additional studies have been published on this topic in the English literature,11–13 one important observation can already be made. Whereas the rates of uterine rupture following spontaneous onset of labour fall within a narrow range, from 0%9 to 0.7%,4,9,10,12 those following induction of labour range from 0.3% (current study) to 4.6%.12 This variability in the latter group is most likely to reflect differences in patient selection and in the protocols used for induction of labour. For example, differences in management may explain why rates of uterine rupture are lower in European than in US studies. For this reason, meta-analysis of the published evidence is unlikely to provide conclusions that are clinically relevant. Instead, comparison of studies with similar patient selection criteria and induction strategies can provide valuable information. For example, use of misoprostol for labour induction in a scarred uterus has been consistently associated with an unacceptably high rate of uterine rupture and it is now considered a contraindication for induction of labour with a scarred uterus.14 Patients with a history of uterine sepsis may be at higher risk of uterine rupture. A recent series reported a 38% chance of such a history of uterine sepsis in women with uterine rupture during a subsequent trial of labour.15 In our series, two of the four cases of uterine rupture had a history of uterine infection after caesarean section or uterine surgery. So improved selection of women suitable for induction may also reduce the risk of uterine rupture. The minimal use of epidural anaesthesia or analgesia in our population may also have played a role in the low rate of uterine rupture associated with induction. Indeed, a review of 29,064 deliveries after caesarean section found that the use of epidural anaesthesia was significantly more frequent among the 70 women who experienced uterine rupture than among those who did not (24%vs 8.4%).16 Moreover, the lowest reported rates of uterine rupture with induction of labour after caesarean section (0% and 0.3%) are in series with negligible use of epidural anaesthesia.11 A recent review on the topic concluded that epidural anaesthesia is associated with a longer labour.17 Because duration of labour may be an important risk factor for uterine rupture,5 use of epidural anaesthesia may indirectly contribute to this complication. Alternative hypotheses to explain the association between epidural anaesthesia and uterine rupture are that epidural analgesia may mask the pain of early uterine diastasis before overt uterine rupture occurs, and it may interfere with the assessment of the severity of labour pain, which is an important symptom of obstructed labour. Whether epidural anaesthesia plays a causative role in uterine rupture remains controversial. Future studies should thus assess which aspects of the process of induction of labour contribute to the causation of uterine rupture and should identify the criteria for selection of the optimal candidate women.