Incidence of Recurrent Uterine Rupture

Women with a history of uterine rupture have about a 10% likelihood of developing recurrent uterine rupture in a subsequent pregnancy.

Abstract

OBJECTIVE:

We aimed to quantify the incidence of recurrent uterine rupture in pregnant women.

DATA SOURCES:

A literature search of PubMed, Web of Science, Cochrane Central, and ClinicalTrials.gov for observational studies was performed from 2000 to 2023.

METHODS OF STUDY SELECTION:

Of the 7,440 articles screened, 13 studies were included in the final review. We included studies of previous uterine ruptures that were complete uterine ruptures, defined as destruction of all uterine layers, including the serosa. The primary outcome was the pooled incidence of recurrent uterine rupture. Between-study heterogeneity was assessed with the I² value. Subgroup analyses were conducted in terms of the country development status, year of publication, and study size (single center vs national study). The secondary outcomes comprised the following: 1) mean gestational age at which recurrent rupture occurred, 2) mean gestational age at which delivery occurred without recurrent rupture, and 3) perinatal complications (blood loss, transfusion, maternal mortality, and neonatal mortality).

TABULATION, INTEGRATION, AND RESULTS:

A random-effects model was used to pool the incidence or mean value and the corresponding 95% CI with R software. The pooled incidence of recurrent uterine rupture was 10% (95% CI 6–17%). Developed countries had a significantly lower uterine rupture recurrence rate than less developed countries (6% vs 15%, P=.04). Year of publication and study size were not significantly associated with recurrent uterine rupture. The mean number of gestational weeks at the time of recurrent uterine rupture was 32.49 (95% CI 29.90–35.08). The mean number of gestational weeks at the time of delivery without recurrent uterine rupture was 35.77 (95% CI 34.95–36.60). The maternal mortality rate was 5% (95% CI 2–11%), and the neonatal mortality rate was 5% (95% CI 3–10%). Morbidity from hemorrhage, such as bleeding and transfusion, was not reported in any study and could not be evaluated.

CONCLUSION:

This systematic review estimated a 10% incidence of recurrent uterine rupture. This finding will enable appropriate risk counseling in patients with prior uterine rupture.

SYSTEMATIC REVIEW REGISTRATION:

PROSPERO, CRD42023395010.

Uterine rupture is a rare but potentially life-threatening condition in the mother and fetus. Uterine rupture occurs more frequently in the setting of a scarred uterus.^1–3 Recent studies have highlighted the growing incidence of uterine rupture attributed to the rising rate of cesarean delivery and myomectomy surgery, especially in pregnancies among older gravid patients.⁴

About 10% of uterine ruptures are managed through hysterectomy, and the remaining cases undergo repair without removal of the uterus.^4,5 Despite uterine preservation, clinicians may not recommend subsequent pregnancies. However, patients with prior uterine rupture often express an interest in childbearing. Therefore, providing patients who have experienced prior uterine rupture with comprehensive counseling on the risks of future pregnancies is important. Brief reviews have crudely estimated the incidence of recurrent uterine rupture, and they did not take into account the weights of each study; moreover, they were not systematic reviews with rigorous methodology.^6,7 Therefore, the overall pooled incidence of recurrent uterine rupture remains unknown.

The purpose of this systematic review was to estimate the pooled incidence of uterine rupture in subsequent pregnancies among patients with previous uterine rupture. The results will provide an appropriate point of counseling for patients with prior uterine rupture and help guide the management of subsequent pregnancies.

SOURCES

This systematic review and meta-analysis was performed in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines and MOOSE (Meta-analysis of Observational Studies in Epidemiology) statement.^8,9 In consultation with a librarian, we systematically searched PubMed, Web of Science, Cochrane Central, and ClinicalTrials.gov for studies published between 2000 and 2023 using the keyword “uterine rupture.” This study period was established to include studies with pregnancy outcomes applicable to contemporary practice. Full details of the search strategy are available in Appendix 1, available online at http://links.lww.com/AOG/D444. The search was limited to English-language journals. The search was not limited to the country or region of publication. We did not conduct a citation search of the selected articles.

This systematic review was registered and approved for inclusion in PROSPERO, which is an international prospective registry of systematic reviews (https://www.crd.york.ac.uk/prospero/, registration number CRD42023395010, registration date February 2, 2023). This study was exempt from ethics approval and informed consent to participate because of its use of publicly available data.

Two independent reviewers (S.S. and K. Yamawaki) evaluated each abstract and full text for eligibility. Inconsistencies after the full-text screening were resolved by discussion. The relevant data were then extracted from the included studies. We did not plan to collect additional data by contacting the authors of the included studies; thus, no attempts were made to contact authors, and only published data were used in analyses.

STUDY SELECTION

Full-text articles in which a previous complete uterine rupture occurred were included. Complete uterine rupture was defined as destruction of all layers of the uterus, including the serosa.¹⁰ Studies with prior incomplete uterine rupture, in which the serosa remained intact, were not included because aspects of clinical care and anticipated outcomes differ from those with complete uterine rupture.¹¹ The included patients were women who had previously developed a uterine rupture, subsequently become pregnant, and then given birth. The diagnostic approach to uterine rupture was determined by the methodology used in each respective article. We excluded pregnancy resulting in miscarriage in patients with prior uterine rupture, articles from which the required data could not be extracted from the original work, studies with fewer than five patients, studies not published in English, and studies published before the year 2000.

All retrieved studies were exported to EndNote 20.4. Study screening was conducted after duplicates were identified and removed with this system.

The primary outcome was the pooled incidence of recurrent uterine rupture in patients with previous uterine rupture. The secondary outcomes comprised the following: 1) mean gestational age at which recurrent uterine rupture occurred, 2) mean gestational age at which delivery occurred without recurrent uterine rupture, and 3) perinatal complications (blood loss, transfusion, maternal mortality, and neonatal mortality).

The quality of each included study was assessed with the Risk-of-Bias Assessment Tool for Non-randomized Studies.¹² This tool was developed to assess the risk of bias in nonrandomized studies and consists of the following six domains: 1) selection of participants (selection bias), 2) confounding variables (selection bias), 3) measurement of exposure (performance bias), 4) blinding of outcome assessments (detection bias), 5) incomplete outcome data (attrition bias), and 6) selective outcome reporting (reporting bias). The risk of bias in each area was categorized as a low risk, high risk, or unclear risk. The details and evaluation criteria of these domains are shown in Appendix 2, available online at http://links.lww.com/AOG/D444. A funnel plot was created to examine the publication bias. When the number of studies was sufficient, the Egger test was used to examine the asymmetry of the funnel plot.¹³

All data analysis was performed with R 4.3.0. To calculate the pooled mean value, the mean value and SD of gestational age were calculated for studies that reported individual patient data. In studies that reported the median and range, the mean and SD were approximated with conversion formulas.^14,15 A random-effects meta-analysis was performed with the meta package in R.

We calculated the I² value for heterogeneity among studies. Subgroup analyses were conducted in terms of the country development status (developed vs developing countries), publication period (2000–2012 vs 2013–2023), and study size (single-center vs national studies). The country development status was classified with the United Nations definition.¹⁶

Statistical analyses were based on two-sided hypotheses, and a value of P<.05 was considered statistically significant. The Egger test was used to assess the publication bias using R, and a value of P<.1 was considered statistically significant.

RESULTS

The study selection process is shown in Figure 1. The literature search identified 10,706 articles. The total number of articles screened was 7,440. We excluded 7,404 articles after screening titles and abstracts according to the inclusion and exclusion criteria. The remaining 36 articles were evaluated by full-text screening. Ultimately, 13 studies were considered eligible and included in this meta-analysis.

Fig. 1. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-analyses) diagram for study selection.

Sugai. Pregnancy After Uterine Rupture. Obstet Gynecol 2023.

The 13 studies are shown in Table 1.^7,17–28 All studies were retrospective. Nine studies (69.2%) were conducted at a single site^{7,17,18,20,22–26}; two (15.4%) were conducted at multiple sites^19,21; and two (15.4%) were national studies.^27,28 Six studies (46.2%) reported individual patient data.^{7,17,18,20,21,28} The minimum sample size was five pregnancies in one study,¹⁷ and the maximum was 72 pregnancies in another study.²⁸

Table 1.

Overview of the Included Studies

An evaluation of the risk of bias of the studies is shown in Appendix 3, available online at http://links.lww.com/AOG/D444. In the categories “selection of participants (selection bias)” and “measurement of exposure (selection bias),” the majority of the studies had a low risk of bias. However, in the category “confounding variables (selection bias),” the majority of the studies had a high risk of bias. “Incomplete outcome data (attrition bias)” and “selective outcome reporting (reporting bias)” were mostly rated as an unclear risk of bias. Because this study was not a comparative study and there was no obligation to implement “blinding of outcome assessments (detection bias),” all studies were assumed to have an unclear risk of detection bias.

The pooled incidence of the primary outcome, recurrent uterine rupture, was 10% (95% CI 6–17%) (Fig. 2). Heterogeneity among studies was moderate (I²=44%, P=.05). A funnel plot is shown in Figure 3. We found that studies with small sample sizes, less precise estimates, and a higher incidence of recurrence may not have been published. The Egger test also showed asymmetry (P=.097).

Fig. 2. Forest plot of the incidence of recurrent uterine rupture as the primary outcome.

Sugai. Pregnancy After Uterine Rupture. Obstet Gynecol 2023.

Fig. 3. Funnel plot to assess publication bias of the primary outcome.

Sugai. Pregnancy After Uterine Rupture. Obstet Gynecol 2023.

The results of subgroup analyses are shown in Table 2, and forest plots are shown in Appendix 4, available online at http://links.lww.com/AOG/D444. Five studies (38.5%) were from developed countries,^{7,17,22,27,28} and eight (61.5%) were from developing countries.^{18–21,23–26} The pooled incidence of recurrent uterine rupture in developed countries was 6% (95% CI 3–11%). Heterogeneity was not observed (I²=0%, P=.77). In contrast, the pooled incidence of recurrent uterine rupture in developing countries was 15% (95% CI 8–25%), with moderate heterogeneity (I²=40%, P=.11). The country's economic situation was significantly associated with the incidence of recurrent uterine rupture (P=.04). Other subgroup analyses for year of publication (2000–2012 vs 2013–2023) and study size (single-center vs national studies) showed no statistically significant differences (P=.12 and P=.27, respectively).

Table 2.

Subgroup Analysis of the Incidence of Recurrent Uterine Rupture

The results of the secondary outcomes are shown in Appendix 5, available online at http://links.lww.com/AOG/D444. The pooled mean number of gestational weeks at recurrent uterine rupture was 32.49 (95% CI 29.90–35.08, I²=70%, 4 studies). The pooled mean number of gestational weeks at delivery without recurrent uterine rupture was 35.77 (95% CI 34.95–36.60, I²=80%, 7 studies). Overall, the maternal mortality rate in the subsequent pregnancy of women with a history of uterine rupture was 5% (95% CI 2–11%, I²=0%, 8 studies). In addition, the neonatal mortality rate was 5% (95% CI 3–10%, I²=9%, 11 studies). The funnel plots are shown in Appendix 6, available online at http://links.lww.com/AOG/D444. Because blood loss and transfusion volumes were not reported in any studies, we could not integrate these data.

DISCUSSION

Our meta-analysis estimated a pooled incidence of recurrent uterine rupture of 10% in patients with a history of uterine rupture. In a subgroup analysis, we found that residence in a developed country was associated with a significantly lower incidence of recurrent uterine rupture compared with residence in developing countries. We believe that these results will enable provision of more nuanced counseling regarding the risk of rupture in subsequent pregnancies of a patient with prior uterine rupture.

The incidence of uterine rupture is higher in patients with than without a scarred uterus. In previous reviews, the pooled incidence of uterine rupture in vaginal deliveries after cesarean delivery ranged from 0.38% to 0.62%,^29,30 and a prospective study revealed an incidence of 0.69%.³¹ Furthermore, the incidence of uterine rupture in vaginal delivery after two cesarean deliveries was reported to be 1.36%.³² The overall incidence of uterine rupture in the subsequent pregnancy after classic cesarean delivery was estimated to range from 0.88% to 1.10%.^33–35 One study showed that up to 10.6% of classic cesarean deliveries resulted in uterine rupture when accompanied by labor in the next delivery.³⁴ The incidence of uterine rupture after myomectomy has been reported to be 0.25–0.42% after laparotomy and about 1.0% after laparoscopy.^36,37 In our study, the pooled incidence of recurrent uterine rupture was 10%. The incidence of recurrent uterine rupture may be similar to that of uterine rupture in the setting of vaginal delivery after classic cesarean delivery, although it is not directly comparable.

Uterine rupture is more prevalent in developing countries than in developed countries, and it is one of the leading causes of maternal mortality.^29,38,39 We performed a subgroup analysis according to the country's economic status (developed vs developing countries). The incidence of recurrent uterine rupture was 6% in developed countries and 15% in developing countries. This difference may be attributed to a number of factors, including a delayed response to obstructed labor; differences in use of medications; lack of dissemination of health information, including warning signs, to patients; and limited access to health care in economically poor countries.^40,41 It is notable that the observed differences were driven primarily by a single study.¹⁸ Nonetheless, it is possible that increased access and improved obstetric care in developing countries could reduce the number of recurrent uterine ruptures in pregnancy.

Some prior reports concluded that allowing a subsequent pregnancy after uterine rupture is appropriate.^{7,22,25,26,28} Our pooled incidence of recurrent uterine rupture, maternal mortality, and neonatal mortality may be viewed differently by different clinicians and patients. If a patient with a history of uterine rupture wishes to become pregnant, the risk of recurrent uterine rupture and the possibility of serious conditions including death of the mother and newborn should be discussed as part of shared decision making.

We also found that the mean gestational age at which recurrent uterine rupture occurred was 32.49 weeks, which is before the time when delivery would typically be recommended. Patients should be instructed to visit the hospital as soon as signs of labor appear, and they need to be evaluated promptly. We also found that the pooled mean gestational age at delivery in patients with a history of uterine rupture was 35.77 weeks. However, it was difficult to identify the indications for delivery in each included study. Frank et al⁴² used a decision analysis model with a theoretical cohort and proposed that delivery should be planned between 34 and 36 weeks of gestation. Our result is consistent with this recommendation and may simply represent standard practice patterns. Delivery timing will ultimately be based on an assessment of the balance between complications of prematurity and any evidence of an increased risk of recurrent uterine rupture.

This study has several limitations. The Egger test showed possible publication bias in the pooled incidence of recurrent uterine rupture. Studies showing a high incidence of recurrent uterine rupture may not have been published. We excluded non–English-language publications and abstracts that did not contain the full text. The lack of uniformity of management in pregnancies after uterine rupture and missing information on the details of the cases of uterine rupture may limit clinical applicability. Examples of missing information are whether the rupture site was in the lower part or fundus of the uterus, whether there had been a previous cesarean delivery or myomectomy, and whether the patient was in labor. No prospective studies were found, and the number of cases in each study was small. However, given the relative rarity of uterine rupture, prospective studies are difficult to conduct, and future high-quality retrospective studies with sufficient sample sizes are needed.

A strength of this study is the high quality of our systematic review and meta-analysis methodology, which accounted for the weight of each study. The subgroup analysis also showed that the recurrence rate varied greatly, depending on the economic situation of the country, which may allow more accurate counseling based on maternity care setting.

In conclusion, we found a 10% incidence of recurrent uterine rupture. This information can be integrated into counseling about subsequent pregnancies in women with a history of uterine rupture.

Figure.

No available caption