Studies included in the meta-analysis
The search strategy identified 250 articles from PubMed, 200 articles from Google Scholar, 110 articles from Cochrane Library, 102 articles from African Journals Online, 30 articles from Ethiopian’s University research repository online library, and 10 articles through manual search. Of which, 242 were excluded due to duplication, 424 through review of titles and abstracts. Additionally 40 full-text articles were excluded for not reporting the outcome variable and other reasons. Finally, 16 studies were included (Fig. 1).
Characteristics of included studies
In this review, 16 relevant studies were included with a sample size of 10, 5530. Among sixteen studies, eleven cross sectional(2,8,22-24,28-33 ), four case-control (37,50,51,52)and one cohort ([26]) studies were found. Regarding the geographical area, seven studies were conducted in Amhara region(2,8 ,24,28,29,31,50) ,four in Oromia(23,27,33,32), three in Southern Nation Nationalities and People (SNNPR) (29,31,52)and two in Tigray(22,30)(Table 1).
Quality of the included studies
Four study was assessed using JBI checklist for case-control studies [37, 50, 51, 52], eleven studies [2,8,22-24,28-33 ] using JBI checklist for cross sectional studies, and one study [27] using the JBI checklist for case control studies. None of the studies were excluded based on the quality assessment criteria (Table 1).
Prevalence of uterine rupture
Primarily, among sixteen studies, four case-control studies were not considered in the prevalence estimation. Consequently, twelve studies [2,8,22-24,27-33] were included in the final meta-analysis to estimate the prevalence of uterine rupture..
The overall pooled of the prevalence of was 14.74 %( 95%CI:10.64, 18.83). The heterogeneity among the studies used to estimate the pooled Prevalence of uterine rupture was marked (I2 = 96.1% and P≤ 0.001)(Fig. 2). Subgroup analysis was done by study area and year of study. Thus, the pooled prevalence of uterine rupture was high in the Tgiray region (21% (95%CI:18.99, 23)), and the least occurrence was in Amhara (14 %( 95% CI:7.22, 20.08)). Similarly, the subgroup analysis by year of study showed that uterine rupture was 13.19% (95%CI:9.01, 17.37) in studies conducted after 2017, Studies conducted from 2014-2015 showed a prevalence of uterine rupture was 17.16 % (95%CI:5.37, 28.94), and studies conducted before 2014 showed the prevalence of uterine rupture was 25% (95%CI: 22.17, 27.83).
Publication bias
Funnel plot was assessed for asymmetry distribution of prevalence of uterine rupture among pregnant women by visual inspection (Fig.3). Egger's regression test showed with a p-value of 0.87 indicated the evidence for no publication bias.
Sensitivity analysis
This systematic review and meta-analysis showed that the point estimate of its omitted analysis lies within the confidence interval of the combined analysis (Fig.4). Therefore, trim and fill Analysis is no further computed.
Risk factors for uterine rupture
In this systematic review and Meta-analysis; prolonged labour(>24hours), not having antenatal care during pregnancy, grand multi parity, rural residents, having cesarean section scar ,not using partograph during childbirth , and obstructed labour were the factors for uterine rupture.
Rural residency and uterine rupture
A total of seven articles were included to identify the association between rural residency and uterine rupture. Women who lived in rural area were 4 times more likely to have uterine rupture urban residents (OR: 4.17; 95% CI: 1.72, 10.12)(Fig. 5).
Grand multiparity and uterine rupture
Six studies [2, 23,25,28,30,31] showed a significant association between grand multiparity and uterine rupture. Women who were grand multipara were 2 times more likely to develop uterine rupture (OR: 1.99: 95% CI: 1.04, 3.81). Egger’s regression test was showed a p-value of 0.08(fig.6).
Not having antenatal care during pregnancy and uterine rupture
Eight articles were included in this analysis [2, 23, 27,28,31,30,50,52]. There was a higher prevalence of uterine rupture among participants who did not have antenatal care during pregnancy. Women who don’t have antenatal care during pregnancy were four times more likely to develop uterine rupture (OR: 3.7: 95% CI: 2.26, 6.08)( Fig.7)
Previous cesarean section scar and uterine rupture
Three studies [23, 25, 32] showed a significant association between Previous cesarean section scar and uterine rupture. Women who had previous cesarean section scar were 9 times more likely to develop uterine rupture (OR: 8.5: 95% CI: 4.98, 14.59). Egger’s regression test was showed a p-value of 0.26(Fig.8).
Prolonged labour and uterine rupture
Six articles with were included in this analysis [25,27,31,32,50,52]. There was a higher prevalence of uterine rupture among women who had prolonged labour. Women who had prolonged labour were four times more likely to develop uterine rupture (OR=4.3, 95% CI:2.57, 7.19)(Fig.9). Egger’s regression test was showed a p-value of 0.02.
Obstructed labour and uterine rupture
Four studies [23, 3, 32,52] showed a significant association between obstructed labour and uterine rupture. Women who had obstructed labour were 9 times more likely to develop uterine rupture (OR:8.8: 95% CI: 3.38, 22.80)(Fig.10).
No parthograh utilization during labour and uterine rupture
Three studies [2,23,52] showed a significant association between Not utilizing parthograh during labour and uterine rupture. Not using parthograh during active phase of labour increase the probability of uterine rupture by 3 times (OR:3.4:95% CI:1.62,7.29)(Fig.11). Egger’s regression test was showed a p-value of 0.05.