Rupture of the uterine scar during term labour: contractility or biochemistry?


Dr C. Buhimschi, Department of Obstetrics, Gynecology and Reproductive Sciences, Yale University, 333 Cedar Street, LCI 804, New Haven, Connecticut 06520, USA.


Objective  Vaginal birth after a prior low transverse caesarean section (VBAC) is advocated as a safe and effective method to reduce the total caesarean section rate. However, the risk of uterine rupture has dampened the enthusiasm of practising clinicians for VBAC. Uterine rupture occurs more frequently in women receiving prostaglandins in preparation for the induction of labour. We hypothesised that similar to the cervix, prostaglandins induces biochemical changes in the uterine scar favouring dissolution, predisposing the uterus to rupture at the scar of the lower segment as opposed to elsewhere.

Design  We tested aspects of this hypothesis by investigating the location of uterine rupture associated with prostaglandins and compared it with the sites of rupture in the absence of prostaglandins.

Settings  Two North American University Hospitals.

Population  Twenty-six women with a prior caesarean section, experiencing uterine rupture in active labour.

Methods  Retrospective review of all pregnancies complicated by uterine rupture at two North American teaching hospitals from 1991 to 2000.

Main outcome measure  Site of the uterine rupture.

Results  Thirty-four women experienced rupture after a previous caesarean section with low transverse uterine incision. Ten of the women who ruptured (29%) received prostaglandins for cervical ripening (dinoprostone: n= 8 or misoprostol: n= 2) followed by either spontaneous contractions (n= 3) or oxytocin augmentation during labour (n= 7). In 16 women (47%), oxytocin alone was sufficient for the induction/augmentation of labour. Eight (23%) women ruptured at term before reaching the active phase of labour in the absence of pro-contractile agents or attempted VBAC. There were no differences among the groups in terms of age, body mass index, parity, gestational age, fetal weight or umbilical cord pH measurements. Women treated with prostaglandins experienced rupture at the site of their old scar more frequently than women in the oxytocin-alone group whose rupture tended to occur remote from their old scar (prostaglandins 90%vs oxytocin 44%; OR: 11.6, 95% CI: 1.2–114.3).

Conclusion  Women in active labour treated with prostaglandins for cervical ripening appear more likely to rupture at the site of their old scar than women augmented without prostaglandins. We propose that prostaglandins induce local, biochemical modifications that weaken the scar, predisposing it to rupture.


Caesarean section is the most common major surgical intervention in the United States,1 and its frequency is likely to increase further for non-recurring indications such as breech presentation.2 One proposed method to reduce the overall rate of caesarean delivery3 is to encourage women with a prior low transverse caesarean section to deliver vaginally (VBAC) during their next pregnancy, as approximately 80% of the women who attempt VBAC will be successful.4–6

Despite extensive study into the safety and efficacy of VBAC, controversy continues on how best to conduct a trial of labour, as uterine rupture is life threatening to both mother and child.7 Analyses of multiple confounding variables suggest the maternal morbidity and mortality rates associated with the elective caesarean delivery of women with a uterine scar is two to four times lower than those who attempt vaginal delivery.8,9 The most commonly suggested cause of uterine rupture in women attempting VBAC—increased uterine activity—is consistent with the well-documented early and frequent requirement for oxytocin augmentation during the active phase of labour.10 However, Phelan et al.11 noted that uterine activity patterns and oxytocin use are not categorically associated with uterine rupture, and suggested that other factors might be relevant. Several recent studies report an increased risk of uterine scar disruption in women receiving prostaglandin in preparation for VBAC.12,13

Prostaglandins are commonly used before induction of labour to ‘ripen’ the cervix, and their use is associated with uterine tachysystole and hypertonus.14 The difference can be dramatic. Uterine rupture rates after prostaglandins induced ripening/labour approximate 24.5 per 1000 compared with 5.2 per 1000 for spontaneous labour and 7.71 per 1000 for induced labour without prostaglandins.15 Despite contractions being caused by prostaglandin F breakdown, the high incidence of scar rupture in prostaglandins associated VBAC has not been satisfactory explained. We suggest prostaglandins induce ultrastructural changes in the extracellular matrix of the lower uterine segment scar, predisposing to dissolution during labour.16 If this hypothesis is correct, women with a prior caesarean who labour after treatment with prostaglandin will rupture at the site of the old scar more frequently than women undergoing VBAC without prostaglandin. We tested aspects of this hypothesis by determining the impact of prostaglandin treatment on the location of uterine rupture in patients with a prior low transverse uterine incision.


We reviewed medical records of women diagnosed with uterine rupture at two North American institutions (Wayne State University, Detroit, Michigan and University of Maryland, Baltimore, Maryland) over a 10-year period of June 1991–June 2000. The International Classification of Disease codes for uterine rupture were used to identify patients from hospital medical records and patient discharge data. The search was supplemented by a review of perinatal databases maintained at both institutions. Uterine rupture was considered the primary outcome and defined as a transmural myometrial defect extending into the peritoneal cavity with fetal heart abnormalities and requiring either surgical repair or hysterectomy.

The records were abstracted for demographic information, obstetric history and maternal and fetal outcome. Information on the course of labour was available for all patients. The site of uterine rupture with its attendant complications was determined by reviewing the operative reports, discharge summaries and intrapartum notes. Acknowledging that myometrial scar cannot be recognised visually at laparotomy in 46% of women,17 we assumed that the scar was the site of rupture. We defined uterine rupture remote from the old uterine scar as rupture distant from the lower uterine segment (left/right lateral side of the uterus, posterior uterine wall medial to the uterine fallopian tube, midline vertical incision of the uterine body occurring away from lower uterine segment). Most of these patients required a separate uterine incision to complete either delivery of the fetus or placenta, followed by a separate repair of the defect.

The fetal heart rate monitor strips were reviewed when available (11 out of 26 instances of rupture during augmented labour). Uterine activity was quantitated independently by two investigators (CSB, SP) who reviewed the details of the fetal heart rhythm and intrauterine pressure tracings for at least one hour prior to rupture. In the 15 patients for whom the monitor strips were either lost or not obtained, the nursing records and labour progress notes were reviewed to ascertain the number and maximum amplitude of uterine contractions for at least one hour prior to the catastrophic event.

Statistical analysis included Student's t test for continuous variables followed by Mann–Whitney Rank Sum Test, one-way repeated ANOVA and a two-tailed Fisher's exact test for categorical variables. A P value <0.05 was considered to indicate significant differences.


A total of 62,746 vaginal deliveries and 14,350 caesarean sections were performed at Wayne State University—Hutzel Hospital during June 1991–June 2000, whereas 16,268 vaginal deliveries and 3226 caesarean sections were performed at the University of Maryland during the same period. The Wayne State University perinatal database identified attempted VBAC as a separate entity beginning in 1995. The success rate for VBAC from 1995 to 2001 was 87%; and the overall incidence of uterine rupture was 0.94%. Similar calculations were not possible from the available data at the University of Maryland.

Thirty-four women experienced rupture of their uterus after a previous caesarean section. Twenty-six women in the active phase of labour (Table 1) with a low transverse uterine scar experienced uterine rupture. Ten women (29%) (prostaglandin/oxytocin group) ruptured after prostaglandin administration for cervical ripening (dinoprostone, n= 8, or misoprostol, n= 2) followed by either spontaneous contractions (n= 3) or oxytocin augmentation (n= 7). In 16 women (47%), the induction/augmentation required oxytocin alone (referred further as oxytocin group). Prostaglandin/oxytocin and oxytocin groups are the subject of our analysis.

Table 1.  Maternal, fetal and labour characteristics of the study group.
  1. NA = not applicable; SD = standard deviation.

Maternal characteristics
Age: years, mean [SD]28 [4]28 [5]
Parity, median [range]2 [1–5]3 [1–9]
BMI, mean [SD]30 [10]31 [7]
Gestational age: weeks, mean [SD]38.6 [2.3]39.5 [1.6]
One previous caesarean section, n (%)7 (75)14 (87.5)
Time interval prior caesarean section–rupture: months, median (SD) [range]57 (24) [24–96]55 (32) [12–134]
Labour characteristics
Oxytocin augmentation (%)716
Maximum dose of oxytocin: mIU/minute, median [range]7 [1–20]8 [2–28]
Prostaglandin gel doses, median [range]2.5 [1–4]NA
Time interval last prostaglandin gel–rupture: hours, median [range]9 [2.5–30]NA
Time interval oxytocin–rupture: hours, median [range]8 [2–31]7 [2–30]
Contractions/10 minutes, median [range]3.5 [2–8]3.0 [2–5]
Montevideo Units/10 minutes, median [range]247 [140–380]205 [160–300]
Site of rupture: scar (%)9 (90)7 (47), OR: 11.6, 95% CI 1.2–114.3

There were no differences in age, body mass index (BMI), parity, gestational age, fetal weight and umbilical cord blood gases among the groups (Table 1). Similar doses of oxytocin were used to augment/induce labour (median [range]: prostaglandin/oxytocin group 7 mIU/minute [1–20] vs oxytocin group: 8 mIU/minute, range [2–28]; P= 0.536). There was no difference in either the number of months since the prior caesarean section among groups (average [SD]: prostaglandin/oxytocin group 57 [24] months, oxytocin group 55 [32] months, or in the number of prior caesarean sections (median [range]: prostaglandin/oxytocin group 1 [1–2], oxytocin group 1 [1–2]; P= 0.299).

Women treated with prostaglandin in preparation for VBAC ruptured at the site of the old scar more often than women in the oxytocin group whose rupture typically occurred remote from the scar (prostaglandin/oxytocin group 9 out of 10 vs oxytocin group 7 out of 16, OR: 11.6, 95% CI 1.2–114.3). In one woman treated with prostaglandins who did not rupture at the old scar, the surgeon described an 8-cm vertical tear in the anterior wall of the uterus away from the lower segment. When myometrial contractility was stimulated/augmented solely by oxytocin, myometrial rupture on the lateral wall typically extended into the broad ligament and vaginal sidewall (n= 4). A vertical midline rupture in the anterior wall of the uterine body was identified in four women. Rupture of the posterior wall of the lower uterine segment was described only once. One woman required hysterectomy because of extensive myometrial damage. When rupture occurred at term before active labour, it occurred most often at the old scar (7/8, 88%).

There was no significant difference in uterine activity assessed by the number of contractions in 10 minutes among women who received prostaglandin/oxytocin or oxytocin alone (prostaglandin/oxytocin: median 3.5 [range 2–8] vs oxytocin: 3 [range 2–5]; P= 0.406) or Montevideo Units (prostaglandin/oxytocin: median 247 [range 140–380] vs oxytocin: 205 [range 160–300] MVU; P= 0.759). There were no differences in cervical dilation at the time of rupture between women who received prostaglandins or oxytocin (prostaglandin/oxytocin: median 4 cm [range 1.5–10] vs oxytocin: 7 cm [range 3.5–10]; P= 0.112).


The management of labour in women attempting VBAC remains controversial especially if there is need to hasten the timing of delivery or augment the labour. The use of contractile agents for augmentation and induction of labour complicates any evaluation of the safety and efficacy of VBAC. Some investigators believe prostaglandins are contraindicated in women who attempt VBAC,18 and oxytocin remains the most commonly used agent for the induction and augmentation of labour in the United States.19 It has been accepted since the early reports of uterine rupture that excess uterine activity reflecting the abuse of oxytocic drugs is the most frequent cause of rupture.10,20 Oxytocin induction was in one study associated with more than 50% of uterine ruptures in women with an unscarred uterus.10 But more recent studies do not support this conclusion. One review of oxytocin utilisation in patients attempting VBAC concluded there was no clear contraindication for oxytocin.10 Furthermore, fetal monitoring strips of women attempting VBAC who rupture do not show higher rates of hyperstimulation than VBAC patients who deliver vaginally.11 These observations suggest ‘other factors’ may predispose the uterine rupture.

Current knowledge of myometrial repair and regeneration after caesarean section is extremely limited. Schwartz and Paddock21 were the first to investigate myometrial healing in an animal model. They observed that the uterine incision of the guinea pig heals by fibroblastic proliferation. The first step in the initial healing process is fibroplasia, followed by maturation and reorganisation of the myometrial tissue. There is no scar formation in the guinea pig according to these investigators. A more recent study in humans led the investigators to a different conclusion, demonstrating granulation tissue and fibrosis within 18 days of the caesarean section.17 Thus, the hysterotomy defect is filled by connective tissue, not myometrium, with resulting scar formation. We hypothesise that similar to its effect on the cervix, prostaglandin ‘ripens’ the lower uterine segment facilitating disruption of the uterine scar.22 If this hypothesis is correct, the site of uterine rupture in women treated with prostaglandin should differ from those that were not exposed to prostaglandin. The findings of our present investigation support the working hypothesis.

One clear limitation of the current study is its retrospective design and the absence of operative reports in some patients from the prior surgeries. It is possible that a low transverse incision could have extended laterally into the sidewall of the uterus or a T-incision performed during a prior caesarean delivery without appropriate warning of the patient or poor documentation in the operative report. Further, we cannot exclude the possibility bias between the prostaglandin/oxytocin and oxytocin-only groups that could explain the results without a prostaglandin effect on the uterine scar. The women treated with prostaglandin were undergoing labour induction despite an unfavourable cervix. The oxytocin group either entered labour spontaneously or underwent induction in the setting of a more favourable cervix. Thus, contractile agents may have been administered for longer periods in the prostaglandin/oxytocin group despite the similar doses in the two groups. Other potential bias includes different surgeons, multiple techniques (one vs two layer) for uterine incision closure, different surgical materials, an inability to ascertain between different types and doses of prostaglandin and mechanical or hormonal techniques of cervical ripening. In light of the overall rarity of uterine rupture, a study with adequate power to exclude most of these variables will require a multicentre approach.

The debate on the safety and cost effectiveness of VBAC is not over. The most commonly used agents for the non-mechanical induction of cervical ripening and labour are prostaglandins of the F and E series. A recent study attributed an increased risk of rupture in women who receive prostaglandin agents for either the priming of the cervix, induction of labour or both.12 The increased risk could not be explained by uterine tachysystole. The ideal method of labour induction and cervical ripening in women planning VBAC remains to be identified.

In conclusion, women undergoing cervical ripening with prostaglandins prior to VBAC appear more likely to rupture at the site of the old scar than women augmented by oxytocin. We propose that prostaglandins induce biochemical modifications of the lower uterine segment scar that predispose it to rupture during labour. An understanding of the mechanisms controlling uterine scar formation could lead to new therapies promoting ‘better healing’ at the incision site.


Wayne State University is acknowledged. We thank Mr Michael Kruger (WSU) and Dr Edward Kuczynski (Yale University) for their generous support and help with statistical analysis. Dr Emmanuel Bujold facilitated identification of patients with uterine rupture at Wayne State University.

Accepted 25 April 2004