Methods: For this systematic review and meta-analysis, we searched electronic databases including MEDLINE and Embase (from Jan 1, 1990, to Nov 20, 2017), without language restrictions, for studies on maternal or perinatal outcomes following caesarean sections in LMICs. We excluded studies in high-income countries, those involving non-pregnant women, case reports, and studies published before 1990. Two reviewers undertook the study selection, quality assessment, and data extraction independently. The main outcome being assessed was prevalence of maternal mortality in women undergoing caesarean sections in LMICs. We used a random effects model to synthesise the rate data, and reported the association between risk factors and outcomes using odds ratios with 95% CIs. The study protocol has been registered with PROSPERO, number CRD42015029191.
Findings: We included 196 studies from 67 LMICs. The risk of maternal death in women who had a caesarean section (116 studies, 2 933 457 caesarean sections) was 7·6 per 1000 procedures (95% CI 6·6–8·6, τ2=0·81); the highest burden was in sub-Saharan Africa (10·9 per 1000; 9·5–12·5, τ2=0·81). A quarter of all women who died in LMICs (72 studies, 27 651 deaths) had undergone a caesarean section (23·8%, 95% CI 21·0–26·7; τ2=0·62).
Interpretation: Maternal deaths and perinatal deaths following caesarean sections are disproportionately high in LMICs. The timing and urgency of caesarean section pose major risks.
Funding: Ammalife Charity and ELLY Appeal, Barts Charity, and the UK National Institute for Health Research.