Risk factors for uterine rupture and neonatal consequences of uterine rupture: a population-based study of successive pregnancies in Sweden
M Kaczmarczyk, Department of Epidemiology, Emory University, School of Public Health, 1820 Marlbrook Drive, Atlanta, GA 30307, USA. Email email@example.com
Objective Uterine rupture is a rare but a catastrophic event. The aim of the present study was to explore the risk factors for uterine rupture and associated neonatal morbidity and mortality among a cohort of Swedish women attempting vaginal birth in their second delivery.
Design Population-based cohort study.
Population A total of 300 200 Swedish women delivering two single consecutive births between 1983 and 2001.
Methods Swedish population-based registers were used to obtain information concerning demographics, pregnancy and birth characteristics, and neonatal outcomes. Logistic regression was used to analyse potential risk factors for uterine rupture and risk of neonatal mortality associated with uterine rupture. Odds ratios were used to estimate relative risks using 95% CI.
Main outcome measure Uterine rupture and neonatal mortality in the second pregnancy.
Results Compared with women who delivered vaginally in their first birth, women who underwent a caesarean delivery were, during their second delivery, at increased risk of uterine rupture (adjusted OR 41.79; 95% CI 29.73–57.00). Induction of labour, high (≥4000 g) birthweight, postterm (≥42 weeks) births, high (≥35 years) maternal age, and short (≤164 cm) maternal stature were also associated with increased risk of uterine rupture. Uterine rupture was associated with a substantially increased risk in neonatal mortality (adjusted OR 65.62; 95% CI 32.60–132.08).
Conclusion The risk of uterine rupture in subsequent deliveries is not only markedly increased among women with a previous caesarean delivery but also influenced by induction of labour, birthweight, gestational age, and maternal characteristics.
Uterine rupture is a catastrophic event, most often resulting from the tearing of a previous caesarean scar during labour.1–3 In addition to previous caesarean, known or suspected risk factors for uterine rupture include induction of labour, maternal age, height, body mass index (BMI), education, cigarette smoking, birthweight, gestational age, instrumental vaginal delivery, and interpregnancy interval.1–13 However, effect sizes of these risk factors remain unclear due to methodological differences between studies, such as small sample sizes, lack of population-based studies, varying inclusion criteria, the potential recall bias in case–control studies, and discrepancies in the definition of uterine rupture.8,14 In addition, relatively few studies have examined the risk factors for uterine rupture in all women, regardless of history of caesarean.3,4,10,11
Although the incidence of uterine rupture is low, with an average incidence of approximately 1% among women attempting vaginal birth after a caesarean delivery,1,2,15 the increased maternal and neonatal morbidity and mortality associated with this condition2,12,16,17 serve to highlight the importance of prevention. Moreover, the rate of caesarean sections, most notably those that are elective, is increasing in developed countries.17–19 Consequently, it is possible that rates of uterine rupture will increase in these populations as well.
To further elucidate the risk factors for uterine rupture, we examined data from a nationwide prospective cohort study in Sweden of women attempting vaginal birth in their second delivery. We further assessed the risks of neonatal mortality among women who experienced uterine rupture.
The population-based Swedish Birth Register contains information on 98–99% of all births in Sweden. Starting with the first antenatal visit, information was prospectively collected, including maternal socio-demographic characteristics, reproductive history, and complications during pregnancy and delivery. This information was forwarded to the Swedish Birth Register through copies of standardised antenatal, obstetric, and paediatric records. Complications during pregnancy and delivery were registered when the woman was discharged from the hospital, and they were classified according to the Swedish versions of the International Classification of Diseases (ICD), ninth (ICD-9; used 1987–96) and tenth (ICD-10; used from 1997 and onwards) revisions. Information about onset of delivery (spontaneous, vaginally induced, or caesarean delivery before onset of labour) can from 1992 be obtained from checkboxes and mode of delivery (vaginal noninstrumental, vaginal instrumental [vacuum extraction or forceps] or caesarean) can be obtained from checkboxes or operation codes. Using the unique national registration number assigned to all Swedish residents, it is possible to link successive births of each woman, as well as to obtain information from other registries. Information on maternal education level was obtained from the Swedish Education Register, and information about neonatal death was obtained from the Swedish Cause of Death Register. Information about the Swedish Birth Register is also available online (at http://www.sos.se/fulltext/112/2003-112-3/2003-112-3.pdf).
In all, there were 327 700 women who delivered first births beginning in 1983 and second consecutive live single births from 1992 through 2001. We excluded 18 101 women who had a caesarean second delivery performed before onset of labour. Women for whom onset of labour in the second delivery could not be determined (n = 9399) were also excluded. Thus, the study population included 300 200 women with live single births in second pregnancy. The study was approved by the Research Ethics Committee of Karolinska Institutet.
Definition of exposures
From the first delivery, we only included information on mode of delivery, which was categorised as either a vaginal or a caesarean delivery. For each woman, we defined the interpregnancy interval as the time that elapsed between the birth of the first child and the estimated conception date of the second child. The interval was calculated in days, which were then converted to completed months (30 days was assumed to equal 1 month). Interpregnancy intervals were categorised as <12 complete, 12 to <36, and ≥36 months. Information about onset of second delivery was stratified into spontaneous or induced. Mode of second delivery was categorised into vaginal instrumental delivery or not. From second delivery, we also collected information about birthweight (<2499, 2500–3999, and ≥4,000 grams) and gestational age (≤36, 37–41, and ≥42 weeks). Maternal age was recorded at second delivery and categorised as ≤24, 25–29, 30–34, and ≥35 years. We also obtained information about maternal smoking, BMI, and height collected at time of registration to prenatal care in the second pregnancy. Information about maternal smoking was collected in a standardised manner using checkboxes and was dichotomised into nondaily smoker and daily smoker. Information about BMI was based on measured weight (in kilograms) at first prenatal care visit and self-reported height (in centimetres) and was calculated as weight in kilograms per height in square metres. Women were categorised as normal (BMI ≤ 24.9 kg/m2), overweight (BMI 25.0–29.9 kg/m2), and obese (BMI ≥ 30.0 kg/m2). Height measurements were categorised into the increments ≤159, 160–164, 165–169, and ≥170 cm. Years of completed formal education as of 31 December 2001 were categorised as ≤11, 12, and ≥13 years.
Definition of outcomes
First, we studied risk of uterine rupture (ICD-9 codes 665A and 665B and ICD-10 codes O710 and O711) in the second pregnancy. Thereafter, we studied risks of neonatal mortality (defined as an infant death within the first 27 completed days of life) and a low Apgar score (0–6) at 5 minutes in the second pregnancy.
Risk of uterine rupture in the second pregnancy was studied using the following exposure variables: mode of first delivery (caesarean or vaginal), interpregnancy interval, onset of second delivery (spontaneous or vaginally induced), vaginal instrumental second delivery (yes/no), birthweight, gestational age, and maternal characteristics in second pregnancy, including maternal age at delivery, maternal education, smoking habits, BMI, and height. When we studied risks of neonatal mortality and postneonatal mortality, we also used uterine rupture as an explanatory variable. Analyses of infant outcomes were restricted to live births. Logistic regression was used to produce odds ratios and 95% CI for each outcome. Percent population-attributable risk (PAR) was calculated using the number of caesarean sections, total number of births, and the adjusted odds ratio for the risk of uterine rupture after a previous caesarean. SAS software (version 9.1; Cary, NC, USA) was used for all analyses.
In our study, among the population of 300 200 women who attempted vaginal birth in the second delivery, 288 038 (95.9%) delivered vaginally and 12 162 (4.1%) by caesarean section. Of the 24 876 women who attempted vaginal birth after a caesarean delivery, 24.7% had an emergency caesarean in the second delivery, while the corresponding ratio among women who were vaginally delivered at first birth was 2.2%. Thus, compared with women who delivered vaginally in the first birth, women who delivered by caesarean section in the first birth were more likely to undergo an unplanned caesarean delivery during their second delivery (adjusted OR 13.33; 95 % CI 12.74–13.94).
The overall rate of uterine rupture among women with an attempted vaginal birth in their second delivery was 0.91/1000 (274/300 200). The rate of uterine rupture among women who attempted vaginal birth after a caesarean section in their first delivery was 9.00/1000 compared with a uterine rupture rate of 0.18/1000 among women without a history of caesarean delivery (Table 1). Thus, compared with women who had delivered vaginally in the first birth, women who had a caesarean had a substantially increased risk of uterine rupture in their second delivery (adjusted OR 41.79; 95% CI 29.73–58.74). Compared with women who experienced a spontaneous onset of delivery, women whose second delivery was induced faced a doubled increase in risk of uterine rupture (adjusted OR 2.06; 95% CI 1.48–2.86). Induction of labour was associated with a doubled risk of uterine rupture both among women with a previous caesarean (adjusted OR 2.02; 95% CI 1.41–2.89) and among women who did not have previous caesarean (adjusted OR 2.20; 95% CI 0.99–4.92).
Table 1. Risk factors for uterine rupture in the second delivery
|Caesarean section in first delivery|
|Yes||24 876||224||9.00||41.79 (29.73–58.74)|
|Onset of second delivery|
|Induced||22 297||63||2.83||2.06 (1.48–2.86)|
|Vaginal instrumental second delivery|
|Birthweight second delivery (g)|
|4000+||72 025||100||1.39||1.76 (1.32–2.35)|
|Gestational age in second delivery (weeks)|
|42+||19 253||46||2.39||1.58 (1.09–2.28)|
|Mother’s age at second delivery (years)|
|≤24||45 359||22||0.49||0.70 (0.42–1.17)|
|30–34||98 801||108||1.09||1.34 (0.98–1.84)|
|35+||33 011||47||1.42||1.78 (1.21–2.62)|
|BMI at second delivery (kg/m2)|
|25.0–29.9||57 818||63||1.09||0.93 (0.68–1.27)|
|30+||19 751||36||1.82||1.30 (0.89–1.91)|
|Missing||46 939||36||0.77|| |
|−159||30 522||51||1.67||2.09 (1.38–3.17)|
|160–164||70 394||82||1.16||1.64 (1.14–2.37)|
|165–169||84 568||68||0.80||1.13 (0.77–1.66)|
|Missing||23 902||19||0.79|| |
|Highest level of education (years)|
|10–11||125 154||121||0.97||1.08 (0.76–1.53)|
|13+||107 390||92||0.86||0.84 (0.58–1.22)|
|Smoking status at second delivery|
|Nondaily smoker**||243 893||227||0.93||1.00|
|Daily smoker||40 837||29||0.71||0.75 (0.48–1.18)|
|Missing||15 470||18||1.16|| |
|Interpregnancy interval (months)|
|<12||49 857||59||1.18||1.26 (0.90–1.76)|
|12 to <36**||171 804||160||0.93||1.00|
|>36||78 182||55||0.70||0.69 (0.49–0.97)|
The rate of uterine rupture was substantially higher among women with a vaginal instrumental delivery than among women who did not have a vaginal instrumental delivery (3.46 and 0.83/1000, respectively). However, a vaginal instrumental delivery was not associated with an increased risk in the adjusted analyses (OR 0.77; 95% CI 0.50–1.20) (Table 1). Of all 33 uterine ruptures after a vaginal instrumental delivery, a history of a previous caesarean was present in 29 pregnancies. Thus, the association between vaginal instrumental delivery and risk of uterine rupture was confounded by previous caesarean delivery.
Compared with women whose infants’ birthweights ranged from 2499 to 3999 g, women with higher birthweight infants were at increased risk of uterine rupture. The risk of uterine rupture was also increased in women with postterm (≥42 weeks) compared with term (37–41 weeks) pregnancies. A high maternal age (≥35 years) at second delivery and a short maternal stature (height ≤164 cm) was also associated with increased risks of uterine rupture in the second delivery (Table 1). The adjusted percent PAR for uterine rupture due to previous caesarean was 76.7%.
Neonatal mortality and morbidity
Next, we examined uterine rupture in the second delivery and the risk of neonatal mortality. Among 274 women with uterine rupture, there were 14 neonatal deaths (rate 51.09/1000), while the corresponding neonatal death rate among women without uterine rupture was 1.4/1000. Thus, uterine rupture in the second delivery was substantially associated with an increased risk of neonatal mortality (adjusted OR 65.62; 95% CI 32.60–132.08) (Table 2).
Table 2. Risk factors for neonatal mortality
|Uterine rupture in the second delivery|
Apgar score was examined as an indicator of infant morbidity. Uterine rupture in the second delivery was associated with a low (0–6) Apgar score at 5 minutes in the second delivery (adjusted OR 13.88; 95% CI 9.82–19.63). In the subgroup of infants who survived the neonatal period, the association between uterine rupture and a low Apgar score was of a similar magnitude (adjusted OR 11.49; 7.87–16.77).
In our nationwide Swedish study of more than 300 000 women who underwent labour in their second consecutive single pregnancy, caesarean section in the first delivery was the strongest predictor of uterine rupture. We found that the rate of uterine rupture was 50 times higher among women attempting vaginal birth after a previous caesarean section (rate 9.00/1000) than among women attempting a second vaginal delivery (rate 0.18/1000). This relative risk estimate was attenuated to a 40-fold increase in risk after multivariate adjustments. Thus, this finding is consistent with the relative risk of 40 reported in a large population-based study from Switzerland11 but is considerably higher than those of three other studies that reported relative risk estimates ranging from 9.0 to 19.5.3,4,10 The adjusted PAR of 76.7% in our study was comparable with those from the two studies that also reported the PAR of uterine rupture due to previous caesarean.3,4 The increased risk of uterine rupture among women with a history of caesarean delivery may be explained, in part, by induction of labour.
It has been hypothesised that induction agents, such as prostaglandins and oxytocin, that are used to ‘ripen’ the cervix and to increase uterine contractions can lead to hyperstimulation of the uterus, which may weaken scars from previous caesarean sections and subsequently increase the chances of the scar opening during labour.2,18,20 In a study designed to test this hypothesis, it was found that women who were induced with prostaglandins ruptured at the site of the uterine scar more often than those who were induced with oxytocin.18 This is consistent with results from two studies, reporting that prostaglandins are associated with a higher risk of uterine rupture than other induction agents.8,21 Reports in the literature concerning the risk of uterine rupture after induction of labour in women with a history of caesarean section attempting vaginal birth are inconsistent.9–11,22 However, comparison of reports concerning the association between uterine rupture and induction are difficult due to the inclusion of low numbers of uterine rupture, differences in the method of induction, dosage of induction agent used, and inclusion criteria.
After adjusting for confounding, we found that induction of labour doubled the risk of uterine rupture compared with spontaneous labour among all women attempting a vaginal second labour. A similarly designed population-based study that examined the relationship of uterine rupture and induction did not find an association between induction and uterine rupture,10 although the statistical power in that study was limited by the inclusion of only 42 cases of uterine rupture. Although previous caesarean section is the most common cause of a uterine scar, other procedures, such as perforation, uteroplasty, or cornal resection of a uterine injury may result in scarring.12 Therefore, induction of labour is not only a concern for women with a previous caesarean. This is consistent with our finding that induction was associated with an increased risk of uterine rupture among women with and without a history of caesarean delivery. Also, in rare cases, oxytocin has been found to be associated with rupture of an unscarred uterus,23 offering further explanation to our finding.
The rate of uterine rupture among women with a history of caesarean previous section whose labour is induced may increase with advancing gestational age.24 The thickness of lower uterine scars has been show to decrease as the pregnancy progresses, offering a possible explanation of this increase.25 A study designed to examine the relationship between gestational age and uterine rupture in women with a history of caesarean delivery found a significant increase in the rate of uterine rupture as gestational age increased in all women undergoing vaginal birth and specifically among women whose labours were induced.24 In our study, we found a gestational age of ≥42 weeks to be associated with an increased risk of uterine rupture, regardless of caesarean history or onset of labour compared with a gestational age of 37–41 weeks. This is consistent with the findings of a case–control study, which included 134 cases of uterine rupture26 but contradicts the findings of three other studies.9,10,27 This contradiction may be due to inclusion of a low number of uterine ruptures.
Instrumental vaginal delivery, including ventouse (vacuum extraction) and forceps delivery, has been previously identified as a risk factor for maternal and fetal complications.28,29 The risk of uterine rupture associated with instrumental vaginal delivery is unclear due to an absence of previous studies. We found that the rate of uterine rupture was higher among women who underwent an instrumental vaginal delivery than among women who did not, although this difference was not significant after controlling for confounding by previous caesarean section. Instrumental vaginal delivery is often preformed in women with a history of caesarean delivery to minimise the risk of the uterine scar opening during labour.29 In a case–control study of women with one previous caesarean section who attempted vaginal birth in a subsequent delivery, uterine rupture was found to be significantly associated with instrumental vaginal delivery.13 This further emphasises the need for the prudent management of labour in women with a history of caesarean delivery.
Birthweight of more than 4500 g has also been associated with an increase in maternal and fetal morbidity and mortality.30 Consistent with a population-based cohort study of women with one previous caesarean delivery attempting a vaginal birth,31 we found a nearly doubled increase in the risk of uterine rupture among high birthweight infants (≥4000 g) compared with infants weighing less than 4000 g at birth. This finding contradicts several other reports in which high birthweight was not associated with an increase in uterine rupture.9,10,32,33 In light of a recent trend of increasing birthweight in Western countries,30 it is imperative that the association between high birthweight and uterine rupture be further explored.
In addition to the afore mentioned risk factors, we also found high maternal age (≥35 years) to nearly double the risk of uterine rupture compared with women 25–29 years. This is consistent with a report of a 3.2-fold increased risk of uterine rupture among women with a history of previous caesarean ≥30 years compared with women <30 years32 As the birth rate for women older than 30 years continues to rise,34 more women are at risk for uterine rupture, especially when attempting a trial of labour after a previous caesarean delivery.
As the rates of caesarean section and induction of labour continue to rise in developed countries, the number of women at risk for uterine rupture is also increasing17,19,35 Although the rates of uterine rupture are low in the developed world, there are serious consequences associated with uterine rupture. For example, among 274 cases of uterine rupture in our study, there were 14 (5%) cases of neonatal death, a more than 60-fold increase in the risk of neonatal mortality after the occurrence of a uterine rupture, compared with deliveries in which a rupture did not occur. This finding is comparable with several studies that reported perinatal or neonatal mortality rates of 5–6% in deliveries in which a uterine rupture occurred,5,7,8,10 and even higher rates have been reported.6,36 We also found an increased risk of low Apgar score for infants delivered after a uterine rupture compared with infants delivered without the occurrence of such an event, a finding that was also consistent with those of other studies.5,9,10,22
The strengths of our study include our ability to study successive singleton gestations due to the availability of each woman’s unique Swedish national registration number. The prospective method of data collection allowed us to control for parity, number of previous caesarean sections, and limit recall bias.37 Other advantages of our study include its population-based character and large sample size.
The limitations of our study include the use of ICD-9 and ICD-10 codes to define cases of uterine rupture. ICD-9 and ICD-10 codes do not discern between complete and partial uterine rupture or uterine scar dehiscence. The uterine rupture rate in our study population may have been higher than the reported rates of other population-based studies due to the inclusion of cases of uterine dehiscence. However, we find it unlikely that this influenced our findings appreciably due to our focus on uterine rupture after the onset of labour. Macones et al. found an identical rate of uterine rupture as in our study when using a precise definition of uterine rupture that distinguished between complete rupture and scare dehiscence.26 In addition, we were restricted to data included in the Swedish Medical Birth Registry. For example, we lacked information about indications for caesarean in first delivery, indication or methods for induction, and the exact induction agent (prostaglandins, oxytocin, or other) used to enhance contractions during labour. We chose not to include maternal complications in our analysis due to lack of precise data concerning maternal outcomes, such as excessive blood loss and associated clinical complications. There were eight hysterectomies reported in our study group: seven among women who did not have a uterine rupture and one among women who did experience a rupture. We suspect that the low number of hysterectomies is due to severe underreporting; therefore, we chose not to include this variable in the analysis.
In conclusion, our large, population-based prospective cohort study has confirmed the reports of previous studies. We not only confirmed a strong association between a previous caesarean delivery and risk of uterine rupture in subsequent pregnancy but also that other factors including high maternal age, induction of labour, and high birthweight increased the risk of uterine rupture. All these factors are becoming increasingly prevalent in the pregnant population, which highlights the importance of active management of labour aiming to detect early signs of and preventing threatening uterine rupture during labour.
The study was supported by grants from Karolinska Institutet.
Contribution to authorship
M.K. assisted in study design, implemented the study, conducted the analysis, conducted the research review, and helped prepare the Introduction, Material and methods, Results, and Discussion sections of the text. P.S. assisted in study design and oversaw study implementation. P.T. helped prepare the Introduction, Material and methods, Results, and Discussion sections of the text. S.C. had the study idea, assisted in study design and analyses, oversaw implementation, and helped drafting the paper.
Details of ethics approval
This study was approved by the research ethics committee at Karolinska Institutet (number 4863/2005).