Uterine rupture after previous caesarean section

Authors

  • I Al-Zirqi,

    1. Faculty of Medicine, University of Oslo, Oslo, Norway
    2. National Resource Centre for Women’s Health, Division of Obstetrics and Gynaecology, Rikshospitalet, Oslo University Hospital, Oslo, Norway
    Search for more papers by this author
  • B Stray-Pedersen,

    1. Faculty of Medicine, University of Oslo, Oslo, Norway
    2. National Resource Centre for Women’s Health, Division of Obstetrics and Gynaecology, Rikshospitalet, Oslo University Hospital, Oslo, Norway
    Search for more papers by this author
  • L Forsén,

    1. Norwegian Institute of Public Health, Nydalen, Oslo, Norway
    Search for more papers by this author
  • S Vangen

    1. National Resource Centre for Women’s Health, Division of Obstetrics and Gynaecology, Rikshospitalet, Oslo University Hospital, Oslo, Norway
    Search for more papers by this author

Errata

This article is corrected by:

  1. Errata: Corrigendum Volume 117, Issue 8, 1041, Article first published online: 8 June 2010

Dr I Al-Zirqi, Division of Obstetrics and Gynaecology, Rikshospitalet, Oslo University Hospital, Oslo 0027, Norway. Email iqbal.al-zirqi@rikshospitalet.no

Abstract

Please cite this paper as: Al-Zirqi I, Stray-Pedersen B, Forsén L, Vangen S. Uterine rupture after previous caesarean section. BJOG 2010;117:809–820.

Objective  To determine the risk factors, percentage and maternal and perinatal complications of uterine rupture after previous caesarean section.

Design  Population-based registry study.

Population  Mothers with births ≥28 weeks of gestation after previous caesarean section (n = 18 794), registered in the Medical Birth Registry of Norway, from 1 January 1999 to 30 June 2005.

Methods  Associations of uterine rupture with risk factors, maternal and perinatal outcome were estimated using cross-tabulations and logistic regression.

Main outcome measure  Odds of uterine rupture.

Results  A total of 94 uterine ruptures were identified (5.0/1000 mothers). Compared with elective prelabour caesarean section, odds of rupture increased for emergency prelabour caesarean section (OR: 8.63; 95% CI: 2.6–28.0), spontaneous labour (OR: 6.65; 95% CI: 2.4–18.6) and induced labour (OR: 12.60; 95% CI: 4.4–36.4). The odds were increased for maternal age ≥40 years versus <30 years (OR: 2.48; 95% CI: 1.1–5.5), non-Western (mothers born outside Europe, North America or Australia) origin (OR: 2.87; 95% CI: 1.8–4.7) and gestational age ≥41 weeks versus 37–40 weeks (OR: 1.73; 95% CI: 1.1–2.7). Uterine rupture after trial of labour significantly increased severe postpartum haemorrhage (OR: 8.51; 95% CI: 4.6–15.1), general anaesthesia exposure (OR: 14.20; 95% CI: 9.1–22.2), hysterectomy (OR: 51.36; 95% CI: 13.6–193.4) and serious perinatal outcome (OR: 24.51 (95% CI: 11.9–51.9). Induction by prostaglandins significantly increased the odds for uterine rupture compared with spontaneous labour (OR: 2.72; 95% CI: 1.6–4.7). Prelabour ruptures occurred after latent uterine activity or abdominal pain in mothers with multiple or uncommon uterine scars.

Conclusion  Trial of labour carried greater risk and graver outcome of uterine rupture than elective repeated caesarean section, although absolute risks were low. A review of labour management and induction protocol is needed.

Introduction

Uterine rupture is a rare peripartum complication associated with severe maternal and perinatal morbidity and mortality,1,2 and is one of the most common clinical causes of medical litigation in the developed world.3 Previous caesarean section is the main risk factor for uterine rupture.1,4–6 Because the rate of caesarean section is increasing worldwide, we are dealing with an increasing number of mothers with previous caesarean section, with consequent higher risk of uterine rupture.7 The percentage of uterine rupture after trial of labour (TOL) remains low (<1%),8 with usually favourable maternal and perinatal outcome.1,2 However, dramatic reduction of TOL9 has been observed after reports of a worrying increase in rupture-related serious perinatal outcomes.10,11 This resulted in an increase in the rate of caesarean section and associated short-term and long-term complications.7,8 A meta-analysis found that uterine rupture may be twice as common after TOL than at elective repeat caesarean section.12 The majority of studies compared total TOL with elective repeated caesarean section without looking specifically into ruptures at emergency prelabour caesarean section, and without comparing rupture rates after each of spontaneous and induced labour with elective repeated caesarean section.

Norway has an increasing rate of caesarean section, from 1.8% in 1967 to 17.1% in 2008.13,14 However, there is a relatively high rate of TOL with vaginal births in 51% of mothers with previous caesarean section.15 This gives a good basis for studying uterine rupture among mothers with different starts to birth after previous caesarean section. Such information is available through the Medical Birth Registry of Norway (MBRN) where data about all births are routinely recorded.13

The aim of the present study was to determine the risk factors, percentage and maternal and perinatal outcomes of uterine rupture after previous caesarean section. A further aim was to study the impact of different induction methods on uterine rupture and perinatal outcomes.

Methods

Design and study population

The study was a population-based registry study. We used data from the MBRN. It is mandated by Norwegian law to notify the MBRN of all births in Norway. Midwives or physicians attending a birth complete a standardised form within 7 days of delivery. The form contains information on maternal health before and during pregnancy, detailed information about delivery and complications occurring intrapartum or postpartum, and information about the newborn.

The target population of the present study was all mothers giving birth from 1 January 1999 to 30 June 2005 (n = 365 107). This study sample comprised mothers with previous caesarean section and subsequent births at gestational age ≥28 weeks. We excluded antepartum stillbirths, and births with missing information on maternal age (n = 4), parity (n = 19), gestational age (n = 56) and ethnic origin (n = 161). The final sample included 18 794 mothers who gave birth to a total of 19 057 infants. History of previous caesarean section was specified in the MBRN data in a ticked box without specifying the numbers of previous caesarean sections. According to national guidelines, mothers attempting TOL should have only one previous caesarean section.16

We additionally performed a review of the medical records of mothers with uterine rupture detected at prelabour caesarean section during the study period.

Variables

Main outcome measure

The main outcome measure was uterine rupture, coded as ‘yes’ or ‘no’. Uterine rupture was identified through diagnostic code ICD-10: O71.0 (uterine rupture prior to labour start) and O71.1 (uterine rupture during labour) in the registration form. The International Classification of Diseases coding does not differentiate between complete or incomplete uterine rupture.17

Explanatory variables

These included the following variables:

Start of birth categorised into: ‘Elective prelabour caesarean section’ (reference), defined as planned caesarean section performed before onset of labour, ‘Emergency prelabour caesarean section’, defined as emergency caesarean section before onset of labour, ‘Spontaneous labour onset’, and ‘Induced labour onset’. Elective and emergency prelabour caesarean section constituted prelabour caesarean section, whereas spontaneous and induced labour constituted TOL. Definition of TOL was ‘an attempt of labour that either resulted in vaginal birth after previous caesarean section (VBAC), or resulted in emergency caesarean section after labour start (failed VBAC). Information about start of birth was complete.

Gestational age in weeks calculated by ultrasound at 18 weeks, categorised into: ‘24–36’, ‘37–40’ (reference), and ‘≥41’.

Demographic factors including Maternal age in years categorised into: ‘<30’ (reference), ‘30–39’, ‘≥40’; Parity categorised as ‘para 1’ (one previous delivery) (reference) and ‘para 2+’ (≥2 previous deliveries); Maternal ethnic origin, categorised into ‘Western’ (reference) defined as: Europe, North America and Australia, and ‘Non-Western’ defined as outside Europe, North America and Australia.

Specific induction methods were categorised into: ‘Spontaneous labour onset’ (reference), induction by ‘Prostaglandins with or without amniotomy’, ‘Prostaglandins, amniotomy and oxytocin’, ‘Oxytocin with or without amniotomy’, and ‘Mechanical methods’ defined as amniotomy alone or other non-medical induction methods.

In line with national guidelines,16 the prostaglandins used were Prostin E2 (Dinoproston) intravaginal tablets 3 mg repeated after 6 hours if needed, with maximum dose of three tablets. Oxytocin infusion consisted of 10 IU oxytocin in 100 ml NaCl 9 mg/ml, using an IVAC infusion pump, with a rate starting from 3 ml/hour when used for induction, and from 1 ml/hour when used for augmentation. Infusion rate was increased by 3 ml/hour every 30 minutes if there was no established labour, with maximum infusion rate of 18 ml/hour, not exceeding 6 hours. Fetal heart was monitored by cardiotocography (CTG) in all mothers undergoing induction. The infusion was terminated or reduced if contractions were well established. It was also terminated if no contractions were established within 8 hours from start of infusion.

All variables were identified in a ticked box on the birth registration form.

Secondary outcome measures

Secondary outcome measures consisted of maternal and perinatal outcomes.

Maternal outcomes

Severe postpartum haemorrhage, moderate postpartum haemorrhage, hysterectomy, and exposure to general anaesthesia were secondary maternal outcome measures. Other maternal outcomes such as postpartum sepsis, injuries to other organs and maternal deaths were not detected after uterine rupture, and were therefore not included.

Severe postpartum haemorrhage was defined as a visually estimated blood loss of >1500 ml within 24 hours postpartum, or the need for blood transfusion regardless of the amount of blood loss, coded as ‘yes’ or ‘no’;

Moderate postpartum haemorrhage was defined as a visually estimated blood loss of 500–1500 ml within 24 hours postpartum. Severe and moderate postpartum haemorrhage and exposure to general anaesthesia were identified in a tick box on the birth registration form.

Hysterectomy was identified through diagnostic coding (The Nordic Medico-Statistical Committee Classification of Surgical Procedures: LCD 00; LCC96; LCC00; MCA30; MCA33).18 Information on maternal deaths was collected from The Causes of Death Registry, Statistics Norway.

Perinatal outcome

Perinatal outcome was categorised into four mutually exclusive groups:

Perinatal deaths defined as intrapartum fetal deaths ≥28 weeks of gestation, and neonatal deaths seven or more days after birth, not related to congenital causes. Antepartum stillbirths were excluded from the present study because they were not delivery-related.

Post-hypoxic encephalopathy defined clinically as cerebral irritation, cerebral depression or seizures in the presence of severe asphyxia. Cerebral irritation was defined by national consensus as a stage 1 encephalopathy, and cerebral depression as stage 2 encephalopathy as described by Sarnat and Sarnat.19

Severe asphyxia efined by diagnostic coding ICD-10: P21.0 (falling or steady pulse <100/minute at birth, absent or gasping respiration, poor colour, absent tone; asphyxia with 1-minute Apgar sore 0–3), without encephalopathy.

Other complications defined as any neonatal problem with or without admission to neonatal intensive care unit, excluding perinatal deaths, severe asphyxia and post-hypoxic encephalopathy (PHE) described above. Common problems included prematurity, moderate and non-specific asphyxia, hypoglycaemia, respiratory distress syndrome and transient tachypnoea of the newborn.

Serious perinatal outcome was a computed variable, defined as a composite of perinatal deaths, severe asphyxia and PHE.

Statistical analysis

Frequency analysis was used to quantify the percentage of uterine rupture among all mothers with previous caesarean section, and among mothers with different starts of birth. The association of uterine rupture with different starts of birth, demographics and obstetric risk factors was analysed using cross-tabulation and logistic regression. All explanatory variables with a significance level of ≤ 0.20 in bivariate analyses were included in the multivariate analysis. The association between uterine rupture and maternal and perinatal outcomes was assessed using cross-tabulations and bivariate logistic regression models at different starts of birth. Fisher’s exact test was used when the outcome measure was of insufficient sample size. A multiple logistic regression model was used to measure the association of induction methods with uterine rupture risk among mothers attempting TOL. The association of induction of labour with perinatal outcomes was measured using cross-tabulations and multiple logistic regressions. The level of significance in multivariate analysis was set at < 0.05 and the data were analysed using SPSS, version 15 (Chicago, IL, USA). The influence of observations on the fit of the model was tested for each coefficient in the final models by using the influence test DfBeta in the logistic regression in SPSS.20

Results

A total of 11 954 mothers of 18 794 with previous caesarean section attempted TOL, giving TOL rate of 63.6%, whereas 6840 had prelabour caesarean section (36.4%) (Figure 1). There were 94 uterine ruptures detected among 18 794 mothers (5/1000). Ruptures were identified in 80 mothers after TOL and in 14 mothers with prelabour caesarean sections. Although the percentage of uterine rupture was lowest for elective prelabour caesarean section (0.7/1000), it was much higher when prelabour caesarean section was an unplanned emergency section (7.1/1000). Ruptures were significantly higher after TOL compared with prelabour caesarean section (6.7/1000 versus 2.0/1000; P = 0.000). Failed VBACs were associated with the highest rupture rates. Trial of labour resulted in 80.1% VBACs and 19.9% failed VBACs. Failed VBAC was significantly more likely after induced labour than spontaneous labour (Figure 1, Table 1). Vaginal births complicated by ruptures were five spontaneous vaginal and five vacuum deliveries.

Figure 1.

 Uterine rupture (n, %) at prelabour caesarean section (CS) and at trial of labour among mothers with previous CS.

Table 1.   Risk factors for uterine rupture among mothers with previous caesarean section
 Total no.Uterine rupture, n (no./1000)Unadjusted OR (95% CI)Adjusted OR* (95% CI)
  1. OR, odds ratio; 95% CI, 95% confidence intervals.

  2. *Adjusted for all variables in table.

  3. **Analysing only mother attempting trial of labour: OR: 3.02; 95% CI: 1.3–7.2.

  4. ***Analysing only mother attempting trial of labour: OR: 0.47; 95% CI: 0.3–0.8.

Total18 79494 (5.0)  
Start of birth
Prelabour caesarean section684014 (2.0)  
 Elective prelabour caesarean section (ref.)54424 (0.7)11
 Emergency prelabour caesarean section139810 (7.1)9.79 (3.1–31.2)8.63 (2.6–28.0)
Trial of labour11 95480 (6.7)  
 Spontaneous labour onset923951 (5.5)7.54 (2.7–20.9)6.65 (2.4–18.6)
 Induced labour onset2715     29 (10.7)14.67 (5.2–41.8)12.60 (4.4–36.4)
Maternal age in years
<3013 60866 (4.8)11
30–39 (ref.)433921 (4.8)0.99 (0.6–1.6)1.19 (0.7–1.9)
≥408477 (8.3)1.71 (0.6–7.6)2.48 (1.1–5.5)**
Ethnic origin
Western (ref.)16 86272 (4.3)11
Non-Western1932     22 (11.4)2.68 (1.7–4.3)2.87 (1.8–4.7)
Parity
1 (ref.)1031763 (6.1)11
≥2847731 (3.6)0.59 (0.4–0.9)0.61 (0.4–0.9)***
Gestational age in weeks
28–3617049 (5.3)1.34 (0.7–2.7)1.10 (0.5–2.3)
37–40 (ref.)1399655 (3.9)11
≥41309430 (9.7)2.48 (1.6–3.9)1.73 (1.1–2.7)

Risk factors of uterine rupture

Start of birth was a significant risk factor for uterine rupture. Compared with elective prelabour caesarean section, the odds of rupture were highest for induction (OR: 12.6; 95% CI: 4.4–36.4), followed by emergency prelabour caesarean section (OR: 8.6; 95% CI: 2.6–28.0) and spontaneous labour (OR: 6.6; 95% CI: 2.4–18.6) (Table 1).

Trial of labour (spontaneous and induced labour) significantly increased the odds for uterine rupture versus elective caesarean section (OR: 7.96; 95% CI: 2.9–22.0) (data not shown in table). Although a maternal age ≥40 years versus <30 years was not a significant risk factor in bivariate analysis, it significantly doubled the odds for rupture in multivariate analysis after adjusting for start of birth. The odds of rupture at older age became even greater after excluding mothers with prelabour caesarean sections. Non-Western maternal origin doubled rupture odds compared with maternal Western origin. Gestational age ≥41 weeks versus 37–40 weeks significantly increased the odds for uterine rupture, but this increase was marginal in multivariate analysis after adjusting for start of birth. Parity ≥2 showed a protective effect against uterine rupture, especially in the TOL group where the risk was halved, indicating a protective effect of previous vaginal delivery. There was no major influence of single individuals on the estimated coefficients after using the influence test DfBeta.

Ruptures at prelabour caesarean section

Clinical details of 11 of 14 mothers with ruptures at prelabour caesarean section were available and are shown in Table 2. All infants were inside the uterine cavity. Three mothers had elective caesarean sections, revealing complete rupture without any contractions or CTG changes. Eight mothers had emergency prelabour caesarean sections after infrequent contractions and/or abdominal pain, revealing five complete ruptures and three dehiscences. CTG changes were observed in only two mothers, and both had complete ruptures. Low vertical uterine scars or transverse scars in the lower part of the corpus were present in five mothers; one of the five mothers had had previous complete rupture of an inverted T scar following spontaneous contractions at 37 weeks; she presented in the index pregnancy with prelabour dehiscence at 28 weeks. Multiple uterine scars were present in eight mothers, and two presented with acute abdomen revealing complete ruptures of myomectomy scars. All but two neonates had Apgar scores ≥7 at 1 minute. Severe postpartum haemorrhage occurred in one mother.

Table 2.   Characteristics of births in 11 mothers with uterine ruptures at prelabour caesarean section
 Obstetric historyClinical presentationOperative findingOutcome
  1. AS, Apgar score; CS, caesarean section; CTG, cardiotocograph; PPH, postpartum haemorrhage.

Mother 1Para 2; two previous CS; first at 25 weeks (vertical uterine incision); second at 33 weeks (low transverse incision)38 weeks; no contractions; normal CTG; elective CSComplete rupture of transverse lower segment scar; intact amniotic membranes1 minute AS ≥7
Mother 2Para 1; previous CS at 28 weeks (vertical uterine incision)38 weeks; no contractions; normal CTG; elective CSSmall complete rupture in uterine fundus; intact amniotic membranes1 minute AS ≥7
Mother 3Para 2; two previous CS; second at term, through transverse incision in lower part of corpus due to adhesions.38 weeks; twin pregnancy; no contractions; normal CTG; elective CSComplete rupture of upper scar; intact amniotic membrane1 minute AS ≥7
Mother 4Para 3; first: vaginal delivery; second: emergency CS at 29 weeks (inverted T incision); thrid: rupture of scar at 37 weeks after contractions, resulting in stillbirth28 weeks; infrequent contractions; normal CTG; ultrasound showed uterine wall could not be seen; emergency prelabour CSDehiscence: absence of muscles in isthmus and vertical midline up to fundus; infant parts palpated under a fibrous membraneSevere maternal PPH
1 minute AS ≥7
Mother 5Para 4; four previous CSs at term.33 weeks; twin pregnancy; constant pain and tenderness on uterus; normal CTG; ultrasound showed thin isthmus; emergency prelabour CSDehiscence (lower segment scar). Infant seen moving under transparent serosa1 minute AS ≥7
Mother 6Para 1; previous CS at term due to prolonged labour41 weeks; infrequent contractions; borderline CTG; emergency prelabour CSComplete rupture (lower segment scar); intact amniotic membranes1 minute AS ≥7
Mother 7Para 1; previous CS at term due to prolonged labour38 weeks; infrequent contractions; normal CTG; emergency prelabour CSComplete rupture (lower segment scar); placenta protruding through rupture1 minute AS ≥7
Mother 8Para 1; previous CS; previous laparoscopic myomectomy37 weeks; acute abdomen; no contractions; normal CTG; emergency prelabour CSComplete rupture in uterine fundus; amniotic fluid filling peritoneal cavity1 minute AS ≥7
Mother 9Para 2; two previous CS; one at 26 weeks (vertical incision)36 weeks; infrequent contractions; emergency prelabour CSDehiscence (transverse lower segment scar); infant seen moving under transparent serosa1 minute AS 4; 5 minute AS ≥7
Mother 10Para 2; two previous CS due to prolonged labour38 weeks; infrequent contractions; borderline CTG; emergency prelabour CSComplete rupture (lower segment scar); amniotic membranes: not intact1 minute AS ≥7
Mother 11Para 1; previous CS; previous laparoscopic myomectomy on posterior uterine wall.35 weeks; constant pain and tenderness on uterus; normal CTG; emergency prelabour CSComplete scar rupture on posterior uterine wall; amniotic membrane: not intact1 minute AS 3; 5 minute AS ≥7

Maternal outcome after uterine rupture

Uterine ruptures among all mothers with previous caesarean section resulted in a three-fold, seven-fold, ten-fold and 23-fold increased risk of moderate postpartum haemorrhage, severe postpartum haemorrhage, general anaesthesia exposure and peripartum hysterectomy, respectively (Table 3). The increase in the odds of these morbidities was significant only if ruptures occurred after TOL (Table 3). However, no maternal death, injuries to other organs or postpartum sepsis was identified. Almost half of the mothers diagnosed with uterine rupture after TOL, developed moderate postpartum haemorrhage and almost half were exposed to general anaesthesia, while 15% developed severe postpartum haemorrhage and 3.8% needed peripartum hysterectomy.

Table 3.   Maternal outcomes after uterine ruptures at different starts of birth following previous caesarean section (n = 18 794 mothers)*
Start of birthSevere postpartum haemorrhage (>1500 ml blood loss)Moderate postpartum haemorrhage (500–1500 ml blood loss)HysterectomyGeneral anaesthesia
n (%)OR (95% CI)n (%)OR (95% CI)n (%)OR (95% CI)n (%)OR (95% CI)
  1. CS, caesarean section; NA, not applicable; OR, odds ratio; 95% CI, 95% confidence intervals.

  2. *No injuries to other organ, postpartum sepsis or maternal deaths detected following uterine ruptures.

  3. **Spontaneous and induced labour.

Elective prelabour CS (n = 5442)115 (2.1) 1144 (26.5) 11 (0.2) 377 (6.9) 
No rupture (ref.) (n = 5438)115 (2.1)11143 (26.5)111 (0.2)1377 (6.9)1
Rupture (n = 4)0 (0.0)NA1 (25.0)0.92 (0.1–8.9)0 (0.0)NA0 (0.0)NA
Emergency prelabour CS (n = 1398)52 (3.7) 351 (25.1) 6 (0.4) 250 (17.9) 
No rupture (ref.) (n = 1388)51 (3.7)1350 (25.2)16 (0.4)1246 (17.7)1
Rupture (n = 10)1 (10.0)2.90 (0.4–23.4)1 (10.0)0.33 (0.04–2.6)0 (0.0)NA4 (40.0)3.09 (0.8–11.8)
Trial of labour** (n = 11 954)253 (2.1) 2098 (17.6) 12 (0.1) 749 (6.3) 
No rupture (ref.) (n = 11 874)241 (2.0)12059 (17.3)19 (0.1)1711 (6.0)1
Rupture (n = 80)12 (15.0)8.51 (4.6–15.1)39 (48.8)4.53 (2.9–7.0)3 (3.8)51.36 (13.6–193.4)38 (47.5)14.20 (9.1–22.2)
All mothers (n = 18 794)420 (2.2) 3893 (20.7) 29 (0.2) 1376 (7.3) 
No rupture (ref.) (n = 18 700)407 (2.2)13852 (20.7)126 (0.1)11334 (7.1)1
Rupture (n = 94)13 (13.8)7.21 (3.9–13.0)41 (43.6)2.98 (1.9–4.5)3 (3.2)23.67 (4.0–79.6)42 (44.7)10.51 (6.9–15.8)

Perinatal outcomes after uterine rupture

Among 19 057 births, 97 were complicated by uterine ruptures (81 after TOL and 16 at prelabour caesarean section). All were single births except for two sets of twins, delivered by prelabour caesarean sections, and one twin set after TOL. Uterine ruptures resulted in 56 neonates with perinatal complications (0.29% of total births), of which nine were serious (Table 4). There were no deaths later than 7 days following ruptures. Uterine rupture resulted in a significantly larger number of perinatal complications when it occurred after TOL. Perinatal complications occurred in 48 neonates out of 81 ruptures after TOLs (59.3%).

Table 4.   Perinatal outcomes after uterine ruptures at different starts of birth following previous caesarean section (= 19 057 births)
Start of birthSerious perinatal outcomesSerious perinatal outcomesOther perinatal complicationsTotal perinatal complications
Perinatal deathSevere asphyxiaPost-hypoxic encephalopathy
 n (%)Pn (%)Pn (%)Pn (%)Pn (%)Pn (%)P
  1. NA, not applicable.

  2. *Odds ratio: 24.51 (95% confidence interval: 11.9–51.9).

Elective prelabour CS (n = 5555)5 (0.1) 0 (0.0) 5 (0.1) 0 (0.0) 2033 (36.6) 2038 (36.7) 
No rupture (ref.) (n = 5550)5 (0.1) 0 (0.0) 5 (0.1) 0 (0.0) 2029 (36.6) 2034 (36.6) 
Rupture (n = 5)0 (0.0)1.0000 (0.0)NA0 (0.0)1.0000 (0.0)0.0634 (80.0)0.0634 (80.0)0.064
Emergency prelabour CS (n = 1450)17 (1.2) 6 (0.4) 9 (0.6) 2 (0.1) 771 (53.2) 788 (54.3) 
No rupture (ref.) (n = 1439)17 (1.2) 6 (0.4) 9 (0.6) 2 (0.1) 767 (53.3) 784 (54.5) 
Rupture (n = 11)0 (0.0)1.0000 (0.0)1.0000 (0.0)1.0000 (0.0)1.0004 (36.4)0.2624 (36.4)0.241
Trial of labour (n = 12 052)69 (0.6) 17 (0.1) 38 (0.3) 14 (0.1) 4544 (37.7) 4613 (38.3) 
No rupture (ref.) (n = 11 971)60 (0.5) 14 (0.1) 35 (0.3) 11 (0.1) 4505 (37.6) 4565 (38.1) 
Rupture (n = 81)9 (11.1)0.000*3 (3.7)0.0003 (3.7)0.0023 (3.7)0.00039 (48.1)0.05248 (59.3)0.000
All births (n = 19 057)91 (0.5) 23 (0.1) 52 (0.3) 16 (0.1) 7348 (38.6) 7439 (39.0) 
No rupture (ref.) (n = 18 960)82 (0.4) 20 (0.1) 49 (0.3) 13 (0.1) 7301 (38.5) 7383 (38.9) 
Rupture (n = 97)9 (9.3)0.0003 (3.1)0.0003 (3.1)0.0023 (3.1)0.00047 (48.5)0.04556 (57.7)0.000

One intrapartum death followed rupture and placental abruption after induction by prostaglandins at 40 weeks of gestation. The other resulted after inverted T scar rupture at 37 weeks following spontaneous contractions. Early neonatal death occurred on the second day following uterine rupture after spontaneous labour, augmented by oxytocin, at 41 weeks of gestation. All three were delivered by emergency caesarean section because of threatened fetal hypoxia. Neonates with PHE comprised one born after spontaneous labour at 41 weeks, and two born at term after induction with prostaglandins. Two were delivered by emergency caesarean section and one by vacuum extraction because of threatened fetal hypoxia. Neonates with severe asphyxia without encephalopathy were delivered at term by emergency caesarean section because of threatened fetal hypoxia. These comprised one after spontaneous labour, one after induction with prostaglandins and one after mechanical induction.

In contrast, ruptures at elective or emergency prelabour caesarean sections had no significant impact on perinatal outcome. These ruptures resulted in eight neonates with minimal complications, mainly because of prematurity and respiratory distress syndrome.

The impact of induction method

Among mothers attempting TOL, induction significantly increased the odds for uterine rupture versus spontaneous labour, only when using prostaglandins with or without amniotomy (OR: 2.7). Induction by mechanical methods resulted in a rupture rate similar to the rate after spontaneous labour and lower than the rupture rate after prostaglandins (Table 5).

Table 5.   Association of induction of labour with uterine rupture and perinatal outcomes after previous caesarean section
 Total no. (mothers)*Uterine rupture no. (per 1000)OR (95% CI)OR** (95% CI)
  1. OR, odds ratio; 95% CI, 95% confidence interval.

  2. No major influence of single individuals on the estimated coefficients was found in the model after using the influence test DfBeta.

  3. *Mothers attempting trial of labour.

  4. **Adjusted for gestational age.

  5. ***Prostaglandins versus mechanical method: (OR: 2.75: 95% CI: 1.1–7.4).

Total  11 95480 (6.6)  
Spontaneous labour onset (ref.)923951 (5.5) 1 1
Induction methods271529 (10.7)  
Prostaglandins +/− amniotomy113018 (15.9)2.91 (1.7–5.0)***2.72 (1.6–4.7)
Prostaglandins, amniotomy and oxytocin1682 (11.9)1.30 (0.5–3.6)1.22 (0.4–3.4)
Oxytocin +/− amniotomy5554 (7.2)2.17 (0.5–8.9)2.01 (0.5–8.3)
Mechanical methods8625 (5.8)1.05 (0.4–2.6)***1.04 (0.4–2.6)
 Total no. (births)Serious perinatal outcome no. (%)OR (95% CI)OR** (95% CI)
Total12 05269 (0.6)  
Spontaneous labour, no rupture (ref.)925945 (0.5) 1 1
Spontaneous labour, rupture524 (7.7)17.06 (5.9–49.3)15.40 (5.3–44.8)
Induced labour, no rupture271215 (0.6)1.13 (0.6–2.0)1.08 (0.6–1.9)
Induced labour, rupture295 (17.2)42.65 (15.6–116.8)41.77 (15.1–115.8)

Compared with spontaneous labour without rupture, uterine ruptures after induction increased the odds for serious perinatal outcomes by 41-fold, whereas ruptures after spontaneous labour increased the odds by 15-fold.

Discussion

Main results

Uterine rupture occurred in 5/1000 mothers with previous caesarean section. Trial of labour, especially after induction using prostaglandins, significantly increased the odds for uterine rupture compared with elective prelabour caesarean section. Ruptures at emergency prelabour caesarean section occurred mostly after latent uterine activity or abdominal pain in mothers with multiple or uncommon uterine scars. Ruptures occurring after TOL were associated with significantly higher odds for severe maternal morbidity and serious perinatal outcomes. The greatest odds for serious perinatal outcomes followed ruptures after induction of labour.

Strengths and weaknesses of the study

The strength of this study was the population-based design, covering a large sample, allowing analysis of a rare complication like uterine rupture. We included all births after previous caesarean section regardless of the size of the maternity unit so a selection bias due to unit resources was avoided.

The weaknesses of this study include the use of ICD coding not distinguishing between complete and incomplete uterine rupture. However, we performed a small validation study of coding practice of intrapartum ruptures during the year 2003 in Oslo University Hospital. We found three complete and nine incomplete ruptures. The diagnostic codes O71.0 or O71.1, used to identify uterine ruptures in the MBRN file, had only been given to the three complete ruptures; the nine dehiscences were coded Z03.8 (threatened uterine rupture), which was not used to identify ruptures in the data file. Moreover, the percentage of uterine rupture in this study, especially after TOL, was similar to percentages reported for complete uterine ruptures in other studies.21,22 The worse maternal and perinatal outcomes after these ruptures indicate that they were probably complete ruptures.

As a result of the small number of events (ruptures), some of the results were associated with wide confidence intervals. However, meta-analyses may achieve higher precision.21–24

The impact of start of birth on uterine rupture

The overall percentage of uterine rupture of 0.5% in mothers with previous caesarean section was low, similar to that reported in previous studies in the developed world.8,23 Our rupture percentage of 0.7% after TOL was under 1%, in agreement with previous reports.22,23 Although TOL increased the odds of uterine rupture eight-fold compared with elective prelabour caesarean section, the absolute number was low. The percentage of rupture of 1.1% after induced labour lies within the range reported earlier between 0.3%25 and 4.6%.26 The low rupture rate at elective prelabour caesarean section (0.07%) was consistent with that of 0.16% reported in a meta-analysis.12 The significant increase of ruptures at failed VBACs indicates the importance of careful selection of mothers for TOL, and optimising labour management.

Obstetric history and ruptures at prelabour caesarean section

Complete uterine ruptures were detected at elective caesarean sections in two mothers with low vertical uterine scars and one with a transverse scar in the lower corpus. This indicates that the type of scar plays a role in prelabour ruptures.27,28 We propose that mothers with these scars should be scheduled for elective caesarean section not later than 38 weeks. However, one should be aware that the risk of iatrogenic prematurity and respiratory problems would be increased in infants born by caesarean sections at earlier gestations; a balance between uterine rupture and prematurity risk should be assessed for each individual pregnancy. The high percentage of ruptures identified at emergency prelabour caesarean section in this study has not been addressed previously. The majority of these mothers had multiple uterine scars or other risk factors in their obstetric history. With the exception of two borderline CTGs, there were no other abnormal CTG changes. This indicates that mothers presenting with infrequent contractions and normal CTG should not be left without quick intervention if they had either multiple uterine scars, vertical scars, previous uterine ruptures or myomectomy scars.29 Ultrasound evaluation of lower segment thickness in the third trimester might help to prevent such ruptures.30 The prelabour dehiscences and ruptures could have resulted in catastrophic outcomes if mothers had been left longer without intervention.

The impact of gestational age, maternal age and ethnic origin

There is a controversy regarding the role of advanced gestational age in increasing rupture risk among mothers with previous caesarean section. Earlier studies, mostly of small samples, found an increased rupture risk for gestational age beyond term, possibly as result of decreased thickness of the lower uterine segment as the pregnancy progresses.31–34 Other studies of larger samples, showed that advanced gestational age was not a significant risk factor.35 In this study there was a significant risk associated with advanced gestational age, but this was mostly related to the higher induction rate at this gestational age. A maternal age ≥40 years was associated with increased risk of uterine rupture, a finding in accordance with previous studies.33,36 Many factors may contribute to the increased uterine rupture at older age, including increasing dysfunctional labour because of decreasing strength of the myometrium, and defective healing of the uterine scar.37 Mothers of non-Western origin had increased risk of uterine rupture, in keeping with previous studies indicating greater risk of maternal and perinatal morbidity and mortality in these mothers.38 Communication problems and missing information on obstetric history might play a role in this regard. Finally, the protective effect of previous vaginal delivery is in agreement with previous studies.21,39

The impact of uterine rupture on outcome

The minimal maternal morbidity after uterine ruptures at prelabour caesarean section, in contrast to ruptures after TOL, indicates the significant risks of labour after previous caesarean section. The absence of maternal death, injuries to other organs, and postpartum sepsis may be related to prompt diagnosis and management in hospital settings.

The significant impact of uterine rupture after TOL on perinatal outcomes was similar to other studies,23,24 reflecting the serious nature of such ruptures, even in hospital settings. However, the absolute numbers of serious perinatal outcomes were relatively low.

The lack of significant impact of ruptures at prelabour caesarean section may indicate a shorter duration of fetal hypoxia versus ruptures after TOL, a less catastrophic nature in the absence of strong contractions, or that unnecessary delay in delivering the infant was avoided.40

The impact of induction

Induction of labour, especially with prostaglandins, increased the risk of uterine rupture in this sample as demonstrated elsewhere.26,33,41–46 Buhimschi et al.46 found that women induced with prostaglandins were more likely to rupture at the site of the uterine scar than those induced with oxytocin, suggesting that prostaglandins could induce local, biochemical modifications that weaken the scar, predisposing to rupture. Amniotomy and other mechanical methods carried lower risk of rupture than prostaglandin. This might reflect the strategy of using prostaglandins for an unripe cervix, whereas mechanical methods and oxytocin are used when the cervix is riper, resulting in shorter duration of induction. According to our protocol of induction, we were using a relatively higher dose of prostaglandins and oxytocin compared with other countries with a stricter protocol.25 The results suggest a need to review our induction protocol. Mechanical methods such as intracervical Foley catheter for preinduction ripening might offer safer induction. 47 Induction, especially with prostaglandins, should be used with extreme caution among mothers with previous caesarean section.39,48–50 Larger, ideally prospective, studies with information on dose and type of induction agent, duration of labour and underlying risk factors are needed.

Conclusion

Among mothers with previous caesarean section, TOL, especially after induction, had a significantly increased risk for uterine rupture compared with elective prelabour caesarean section. Ruptures occurred even with latent uterine activity, particularly in women with multiple or uncommon uterine scars. Uterine ruptures were significantly associated with severe maternal morbidity and serious perinatal outcome when occurring after trial of labour. Not recommending trial of labour to mothers with higher risk for failed attempt at VBAC might reduce the rate of uterine rupture. A review of induction protocols is needed to reduce uterine rupture risk and associated serious perinatal outcome. Further studies covering a larger sample over a longer interval of time are highly recommended.

Disclosure of interest

None declared.

Contribution to authorship

I. A. designed the study and wrote the paper. S. V. and B. S. contributed to the study design and drafting, revising it critically and approving the final article draft. L. F. helped in statistical analyses, interpretation of the results and producing the tables.

Details of ethics approval

The Regional Ethical Committee for Medical Research, the Norwegian Data Inspectorate and the Norwegian Directorate of Health approved the study.

Funding

We would like to thank the Norwegian Foundation for Health and Rehabilitation and the Norwegian Women’s Public Health Association for funding the study.

Acknowledgements

Special thanks are due to Pernille Frese, from the National Resource Centre for Women’s Health in Rikshospitalet for her help in drawing the figure provided. We are also grateful to our Consultant colleagues for their great effort in compiling the clinical details from case records of mothers with uterine ruptures in their respective departments: Dr Anne Flem Jacobson from Ullevål hospital, Oslo University Hospital, and Dr Per E Børdahl from Haukeland Hospital, Bergen.

Journal club

Discussion points

1. Objective and background: Compare the risks of trial of labour/attempt at vaginal birth after previous caesarean section with the general risks of repeat caesarean section.

 What are the main complications of uterine rupture? How is this different for women undergoing induction of labour for intrauterine death?

2. Methods: Describe the design of this study and discuss its potential limitations.

 The Equator website (http://www.equator-network.org) provides links to guidelines that facilitate both the writing and the critique of scientific studies. Which guideline would be most useful for this study? Review the methods of this study with reference to the relevant checklist. Discuss the advantages and disadvantages of deviating from the structure described in the checklist.

3. Results and implications: Was there any difference in maternal or perinatal outcome between repeat caesarean and attempt at vaginal birth (trial of labour)? Which were the other risk factors for adverse outcome? Were there any ‘protective’ factors? Discuss the implications for clinical practice.

 How are women with previous uterine scar(s) and latent uterine activity (early labour) managed in your practice/unit? What are the potential risks according to this study?

 The risk of uterine rupture is increased with trial of labour as compared to repeat caesarean section in this study but is still relatively low. What is your interpretation of the level of risk of rupture and adverse outcomes associated with induction of labour, especially with prostaglandins? What is your current recommendation for women with an unfavourable cervix who wish for a vaginal birth? What is the guidance in your unit/country?

D Siassakos

Southmead Hospital, Bristol, UK

Email jsiasakos@gmail.com

Ancillary