Maternal near miss incidence ratio, its causes and risk factors in East Africa: a systematic review and meta-analysis Authors: Alehegn Bishaw Geremew, Claire T. Roberts, Shahid Ullah, Ayal Debie, Jacqueline H. Stephens Published: 2025 ABSTRACT BackgroundMaternal Near-Miss (MNM) is a woman who nearly dies but survives complications during pregnancy, childbirth, or 42 days following termination of pregnancy. MNM has emerged as a proxy indicator for the quality of maternal health services. For every maternal death, there are 20–30 MNM cases. However, there is a dearth of comprehensive evidence on MNM according to the WHO’s MNM approach, as well as inconsistencies in risk factors from primary studies. Hence, we conducted a systematic review and meta-analysis on the maternal near-miss incidence ratio, causes, and risk factors in East Africa.MethodsThe review was conducted according to PRISMA 2020 checklist guidelines. Studies were identified from MEDLINE, Scopus, CNIAHL, PsycINFO, ProQuest, Web of Science, Emcare, and Google Scholar. Random-effect meta-analyses were conducted to produce pooled estimates. Further, an influential leave-one-out sensitivity analyses were conducted to investigate the impact of each study on the pooled estimates.ResultsThirty-nine primary studies on the MNM incidence ratio and/or causes and/or factors associated with MNM were included. The pooled MNM incidence ratio was 37.12 per 1,000 live births (95% CI: 20.22–54.03), ranging from 12 per 1,000 live births in Kenya to 53 per 1,000 live births in Ethiopia. The pooled proportion of direct causes of MNM were as follows: obstetric haemorrhage 29.9% (95%CI: 25.7%-34%), hypertensive disorder of pregnancy 27.8% (95%CI: 22.3%-33.3%), uterine rupture 13.8% (95%CI: 7.9%-19.6%), sepsis 9.9% (95%CI: 7.1-12.8%), abortion complications 8.1% (95%CI: 5.9%-10.3%), and ectopic pregnancy 7.3% (95%%CI: 3.1–11.5%). Several factors were significantly associated with MNM. Low-income women had higher odds of MNM (AOR 2.95, 95% CI: 1.28–4.61), while employed women had lower odds (AOR 0.53, 95% CI: 0.21–0.81). Parity ≥ 5 (AOR 1.65, 95% I: 1.14–2.16), history of abortion (AOR 1.63, 95%CI: 1.04–2.22), history of pregnancy complications (AOR 4.09, 95%CI: 1.43–6.78), history of cesarean section (CS) (AOR = 5.04, 95%CI: 3.17–6.90), lack of antenatal care (AOR 2.91, 95%CI: 2.02–3.78), lack of birth preparedness and complication readiness (AOR 4.18, 95%CI: 2.64–5.72), and long distance to health facility (AOR 3.06, 95%CI:1.09–5.03) were positively associated with MNM. Conversely, women admitted by provider referral had lower odds of MNM (AOR 0.31, 95% CI: 0.07–0.55).ConclusionsThe MNM incidence ratio in East Africa was high. Obstetric haemorrhage, hypertensive disorder of pregnancy, and uterine rupture were the three leading obstetric causes of MNM. Income, employment status, parity, ANC, history of CS, and distance to health facilities were found to be associated with MNM. Policymakers and healthcare workers need to consider these identified risk factors in interventions aimed at improving maternal health.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12978-025-02221-9. Background Despite the global reduction of maternal mortality by 44% between 1990 and 2015 [1], every day, 800 women died from preventable causes associated with pregnancy and delivery in 2020. There was an estimated total of 287,000 maternal deaths worldwide, with 94% of maternal mortality occurring in low- and middle-income countries (LMIC) [1, 2]. In East Africa, between 2000 and 2017, there was a significant reduction in the maternal mortality ratio (MMR), decreasing by 48%, from 853 deaths per 100,000 live births in 2000 to 443 deaths per 100,000 live births in 2017 [3]. This reduction suggests efforts to enhance maternal healthcare services, such as access to emergency obstetric care and maternal health interventions [4], have had a positive impact. Despite the significant reduction, the MMR in East Africa remains high compared to the rest of the world’s regions and is more than twice the global average [1, 3]. The high MMR in East Africa reflects the challenges and disparities in accessing quality maternal healthcare, particularly in low-resource settings [5, 6]. Furthermore, the progress falls short of achieving Sustainable Development Goals (SDG) maternal health targets set by the United Nations Development Programme. Globally, despite the progressive reduction in maternal mortality during the Millennium Development Goal (MDG) era [7], in the first five years of the SDG period, the annual average reduction rate of MMR has reduced [2]. Also, the persistently high maternal mortality continues to alarm the global community, as it seems unlikely the SDG target of maternal mortality of less than 70 per 100,000 live births by 2030 will be achieved [8]. As such, maternal health remains a key focus on the global agenda. Addressing maternal mortality requires sustained commitment, increased investment in the health system, and a multifaceted approach that considers social, economic, and cultural factors affecting maternal health. Measuring maternal mortality burden alone does not give a comprehensive understanding of the number of women encountering life-threatening complications [9–11]. As such, Maternal Near Miss (MNM) is more appropriate for evaluating maternal health interventions and planning the health system [10, 12]. According to the World Health Organisation (WHO), the concept of MNM is defined as “a woman who nearly dies but survives complications during pregnancy, childbirth, or 42 days following termination of pregnancy” [13, 14]. This MNM measure has emerged as a more appropriate proxy measure of maternal health service quality rather than focusing solely on maternal mortality [10, 12, 15]. Review of MNM events facilitates detailed reporting of clinical case characteristics and rapid translation of findings into clinical practice. Easier for clinical care providers to report near-miss cases because less threatening as women survived [16]. The WHO MNM criteria consist of 25 indicators grouped into three categories: 11 clinical, 8 laboratory, and 6 management-based criteria [14]. A good uptake of WHO MNM has been reported in resource-limited settings [17], often with some adaptation, particularly from laboratory-based and management-based criteria. The rate of MNM varies both between and within countries. For East African countries, the MNM prevalence has been reported to range from 5% in Ethiopia [17] to 61.3% in Uganda [18]. Within Ethiopia, the magnitude of MNM ranges from 5% [19] to 26.6% [20]. Postpartum haemorrhage (PPH) and hypertensive disorder of pregnancy (HDP) are two leading causes of MNM [21, 22]. Across East Africa, there is high variation in the prevalence of these complications, with PPH ranging from 2.8% in Somalia [23] to 50.6% in Ethiopia [24], and HDP ranging from 8.9% in Uganda [25] to 53% in Ethiopia [26]. Elsewhere, sepsis [18] and uterine rupture have been reported as the leading cause of MNM [25]. This highlights the need for determining the leading causes of a MNM to identify where targeted evidence-based interventions should be focused. Numerous primary studies have identified a range of risk factors associated with MNM, including young maternal age [23], advanced maternal age [21, 27], place of residence [18, 23], educational status [23], antenatal care (ANC) utilisation [22, 23], delay in deciding care [28], pregnancy interval [23], previous history of obstetric complications [23], history of caesarean Sects [21, 29, 30]., use of partograph for labour monitoring [31], and a delay to the final place of care [17, 24]. However, inconsistencies have been observed across studies, with some reporting no significant association between MNM and some risk factors [24, 32]. This inconsistency may pose a significant barrier to designing effective interventions. Furthermore, there is considerable variation in the reported incidence and causes of MNM, and a lack of a pooled estimate based on the WHO MNM criteria within East African countries. WHO MNM criteria is a globally recognised definition and measurement of life-threatening conditions; applying this tool ensures consistency and allows meaningful comparisons and pooled estimates [14, 33]. Pooling data helps highlight the most prevalent causes of MNM, guiding targeted interventions and resource allocation. Meta-analysis helps resolve inconsistencies and identifies which factors are consistently associated with MNM, thereby strengthening evidence-based strategies and enhancing the quality of clinical practice. Therefore, we conducted this systematic review and meta-analysis to estimate the incidence ratio of MNM, its causes, and risk factors among women living across East Africa. Methods and materials Reporting and protocol registration A systematic review and meta-analysis study was conducted of all literature reporting MNM for locations within East Africa. The entire study was designed and reported based on the Preferred Reporting Item for Systematic Review and Meta-Analysis (PRISMA) 2020 updated guideline [34]. Initially, we searched the International Prospective Register of Systematic Reviews (PROSPERO) to check for the presence of a registered, ongoing, similar study. The study is registered in the PROSPERO database [CRD42023437583]. Databases Seven databases were searched: MEDLINE, Scopus, CNIAHL, PsycINFO, ProQuest, Web of Science, and Emcare. Additionally, we searched for grey literature using Google Scholar. Search strategy With assistance from a research librarian, we developed and used a pre-specified search strategy that included search terms, subject headings, and key terms related to MNM and its causes. Boolean operators (AND, OR, NOT) were used to combine search terms and refine the search results. Furthermore, we applied truncation, proximity, and wildcards as needed based on the requirements of each database. The database search was performed on July 5, 2023, with the Google Scholar search conducted on July 20, 2023. Initially, the MEDLINE database search strategy was developed and customised for each database, as required. The detailed search strategies for the databases are available (see Additional file 1). Eligibility criteria Studies reporting pregnant women, childbirth, or women within 42 days after the birth or termination of pregnancy in East African countries were included. East African countries were defined according to the African Development Bank’s list as Ethiopia, Kenya, Burundi, Uganda, Eritrea, South Sudan, Tanzania, Sudan, Djibouti, Rwanda, Somalia, Comoros, Seychelles, and Madagascar [35]. Studies were included if they reported MNM cases defined using the WHO MNM criteria [13, 14], including modifications of the WHO MNM criteria for local contexts [10, 36]. Primary studies were included that reported at least one of the following: MNM incidence ratio or maternal near-miss cases and live births, and/or causes of MNM, or risk factors associated with MNM; and had to use an observational study design, including cohort, case-control, and cross-sectional designs. Studies published since January 2011, in the English language, were included. Studies were excluded if they reported high-risk study populations only, as including these studies in the pooled estimates could lead to overestimating the MNM incidence ratio. Studies were excluded if they did not use the WHO MNM criteria or a modified version of the WHO MNM for MNM case identification; reported MNM cases after 42 days of delivery or termination of pregnancy; or were multicenter studies that included locations outside East Africa and did not report individual country data. In addition, studies with mixed study designs, qualitative studies, case studies, case series, systematic reviews, narrative reviews, editorial letters, and conference abstracts were excluded. Study selection and screening procedures First, all the literature found in each database and grey literature source were retrieved using EndNote 20 referencing software. Second, all the studies were exported to Covidence support. After duplication removal, studies were screened by title and abstract, followed by a full-text review. We used Covidence support for duplicate removal and screening. Two reviewers (ABG, AD) screened the articles by title and abstract for relevance to the review objectives. The possible discrepancy between the two reviewers’ screening was resolved by discussion with the third author (JS). Then, studies that met the eligibility criteria for the titles and abstract screening based on the predefined criteria were further assessed through a full-text study review for eligibility of inclusion by two independent reviewers (ABG, AD). We obtained the full text for published articles with inaccessible full texts in open access through a request to Flinders University library documentation. The reviewer’s decision was categorised as either included or excluded. Cohen’s kappa statistic was calculated and used to evaluate the degree of agreement between the two bias reviewers. According to Cohen’s kappa statistic (K = 0.89) and percent agreement of 94%, the two reviewers of full-text articles have “almost perfect agreement” to include and exclude the studies [37]. Any discrepancies between the two reviewers regarding the eligibility of studies were resolved through discussion with the third author (JS). The number of articles included and excluded, along with the different reasons for exclusions, is reported in the PRISMA flowchart. Risk of bias assessment For each included study, the risk of bias assessment (critical appraisal) was conducted by using the Joanna Briggs Institute (JBI) critical appraisal tool for quantitative analytical study [38] to assess the methodological quality of the studies and the extent to which the study addressed the possible bias in the design and analysis. The risk of bias assessment carried out independently by two reviewers (ABG, AD). The possible responses for each question set in bias assessment were yes, unclear, and no. The possible disagreement between the two reviewers (ABG, AD) was solved through a discussion of the two reviewers set together. Finally, for each study, the reviewers’ yes responses per total number of questions according to the study design were calculated, with studies considered as having a high risk of bias (low quality) when scoring < 50%, moderate risk of bias 50–74%, or low risk of bias (high quality) ≥ 75%. However, no studies were excluded from this review based on the risk of bias score. Outcome measurements The maternal near-miss incidence ratio was defined as the number of MNM cases per 1000 live births [13, 14]. The proportion of each major direct obstetric complications, including obstetric haemorrhage, hypertensive disorder of pregnancy, uterine rupture, abortion complication, and ectopic pregnancy, was calculated as the number of MNM for specific cause per 100 MNM [39]. Data extraction A Microsoft Excel spreadsheet data extraction template was prepared. Data extracted included the first author, year of publication, country of study, study design, study setting, number of facilities studies conducted, sample size, sampling procedure, study population, response rate, livebirths, maternal near miss case, maternal death, and maternal mortality incidence ratio. In addition, we extracted the following data: causes of the maternal near-miss frequencies, percentages, and risk factors of MNM with their AOR and 95% confidence interval. Two independent reviewers (ABG, AD) extracted the data using the same data extraction template. After the two reviewers completed the data extraction, the extracted data were cross-checked, and the incompatibility of the extracted data was resolved by discussion of two reviewers set together. Finally, we discussed the extracted data with the third author (JS). After all available pertinent information was extracted, when the required information was not available directly from the reported results of primary studies, we recalculated data from the primary studies’ reported results. For example, to estimate the pooled proportion causes, the numerator causes specific MNM and denominator total MNM case used. When there were discrepancies between the reported maternal near miss ratio and its respective maternal near miss and livebirths, we selected the reported MNM case and livebirths, then the MNM incidence ratio and recalculated. Data synthesis and statistical analysis The Excel data was exported to Stata (version 18) software for further analysis. The reported frequency of each direct obstetric cause and total maternal near miss case were used to calculate the cause proportion of maternal near miss. Then, estimation of causes proportion was considered. In this study, to account the possible heterogeneity, random effect modelling was used for estimating the pooled effect of the outcome of interest. The overall effect size was calculated as the weighted average using inverse variance of the specific study effect size, plus the between-study variance. The restricted maximum likelihood method and DerSimonian-Larid model was used, and forest plots were used to report the pooled estimations. A first meta-analysis, using a random effect model, was applied to estimate the MNM incidence ratio, which generated the pooled MNM incidence ratio with a 95% confidence interval (CI). A subgroup analysis by country was conducted to investigate heterogeneity. The second meta-analysis was a random-effect model to estimate the pooled proportion of each direct obstetric cause of maternal near miss. The third meta-analysis employed a random-effects model to estimate the pooled effects of each risk factor associated with maternal near miss. Adjusted odds ratios (AOR) with 95% confidence intervals were generated for each factor. Heterogeneity and publication bias The heterogeneity of the studies was evaluated subjectively through observed forest plots and objectively using inverse variance (I2) test statistics and Cochran’s Q test. I2 is a statistical measure that quantifies the degree of heterogeneity among studies in meta-analysis. This was expressed as a percentage and indicates the proportion of total variation across studies due to heterogeneity rather than chance. The I2 values < 25%, 25–50%, 51–74% and ≥ 75% were considered as low, moderate, high and very high heterogeneity, respectively [40, 41]. Subgroup analysis and sensitivity analysis (leave-one-out influential analysis) were conducted. According to the leave-one-out analysis results, the study’s influence was declared if the confidence interval did not overlap with the pooled estimate’s confidence interval [42]. Publication biases and small study effects were assessed using a funnel plot and Egger regression statistical test, respectively. Egger test P-value < 0.05 considered there was a small study bias [43]. Results Selection of studies A total of 12,071 articles were retrieved (11,991 from electronic databases and 80 from Google Scholar) using the prespecified search strategies. Out of the total retrieved articles, 5,254 were identified as duplicates and were removed from consideration. During the initial screening, based on the titles and abstracts of the remaining articles, 6,662 articles were identified as irrelevant and removed. Three studies found through Google Scholar were not accessible in full text. These studies had only an abstract available in the research repository; we had requested the full text via email, but had no response from the authors, and therefore, these were not included. A total of 90 studies underwent full-text assessment for eligibility. Of the 90 studies, 51 articles were excluded for various reasons. The major reason for exclusion was related to the measurement of maternal near misses. Specifically, 24 of the excluded studies were primarily excluded because they did not use either the WHO MNM approach or the modified WHO MNM approach for measuring maternal near miss. Finally, 39 studies were included in the study based on the eligibility criteria, giving 468,563 participants. Of these, 26 studies were found relevant for maternal near miss incidence ratio, 28 studies for causes of maternal near miss, and 18 studies for maternal near miss risk factors (Fig. 1). Fig. 1Study selection procedure for systematic review and meta-analysis on MNM incidence ratio, causes and risk factors in East Africa 2011 to 2023 Study selection procedure for systematic review and meta-analysis on MNM incidence ratio, causes and risk factors in East Africa 2011 to 2023 Quality assessments for included studies Quality assessment was applied to all the included studies; the sample size was inadequate in seven cross-sectional studies [31, 44–49]. Among the case-control studies included, only two used matching for at least one variable to control potential confounders [25, 50]. Controlling for confounding during analysis was not applied for three case-control studies [17, 23, 24]. None of the included cohort studies had exposure and exposure groups. One article’s quality assessment score was found to be less than 50% and, thus, labelled as having a high risk of bias [24]. Five articles were deemed to have a moderate risk of bias [22, 23, 45, 49, 51] with all remaining included articles having a low risk of bias (See Additional file 2). Synthesis of study characteristics The included studies were from seven East African countries, with no eligible articles were found from the six East African countries (Eritrea, Djibouti, Burundi, Seychelles, Comoros, and Madagascar). Of the total included studies, Ethiopia (22), Tanzania (5), Kenya (3), Rwanda (3), Somalia (3), South Sudan (1) and Uganda (1). One included study was from a joint setting in Tanzania and Uganda, but the results were reported separately for each country. Out of the 39 included studies, the majority (n = 26, 66.7%) were cross-sectional in design, nine (23.1%) were case-control studies, and four (10.2%) were cohort or follow-up studies. Among the case-control studies, two were nested case-control studies and one was a prospective case-control study. The smallest sample size among the included studies was 183 participants from a case-control study conducted in Ethiopia [17], whereas the largest sample size was 248,112 participants from a cross-sectional study conducted jointly in Tanzania and Uganda [52]. Among the studies, 24 used the WHO MNM approach to measure maternal near misses [17, 23, 26, 30, 31, 46, 48, 50, 51, 53–65]. This approach is likely based on the criteria and definitions provided by the World Health Organisation for identifying and classifying maternal near-miss events. Nine studies used a modified version of the WHO MNM approach. Of the studies that used modified WHO MNM, four used modified WHO MNM in Sub-Saharan African countries [29, 45, 66, 67]. In four studies, authors used the WHO MNM modified by Nelissen et al. in Tanzania [19, 21, 68, 69]. Five studies used the WHO MNM and its modified criteria [27, 49, 52, 70, 71] while one study the WHO MNM and morbidity criteria [47] (See Table 1). Table 1Characteristics of included primary studies reporting maternal near miss (MNM) in East Africa between 2011 and 2023First Author (publication year, reference)CountryStudy designMeasurement usedParticipated womenLivebirthsNo of MNMMDMNM per 1000 livebirthsTeka (2022) [21]EthiopiaCross-sectionalWHO MNM modified6915,1161462428.5Litorp (2014) [53]TanzaniaCross-sectionalWHO MNM13,58413,1214677736Gebremariam (2022) [31]EthiopiaCross sectionalWHO MNM905NR129NRNREgal (2022) [70]SomaliaCross-sectionalWHO MNM, WHO MNM modified6,6586,055792813Shegaw(2021) [27]EthiopiaCross-sectionalWHO MNM, WHO MNM modified1,2441,17348340.9Dahie (2022) [23]SomaliaUnmatched Case ControlWHO MNM533NR178NRNRLiyew (2018) [50]EthiopiaNested case-controlWHO MNM864NR216NRNRGeze (2021) [29]EthiopiaUnmatched nested case ControlWHO MNM modified432NR108NRNRKiruja (2017) [54]SomaliaCross sectionalWHO MNM1,38513551201888.6Rulisa (2015 [55]RwandaCross-sectionalWHO MNM1,739173914250-Tariku (2020) [44]EthiopiaCross sectionalWHO MNM modified337NR594NRHerklots (2017) [56]TanzaniaCross-sectionalWHO MNM55514,12537289De Haan (2023) [45]EthiopiaCross-sectionalWHO MNM modified637NR700NRGeleto (2020) [57]EthiopiaCross-sectionalWHO MNM78,195323,82477,714481239.9Kumela (2020) [17]EthiopiaCase-controlWHO MNM modified183NR61NRNROwolabi (2020) [58]KenyaCross-sectionalWHO MNM69536,168360179.97Asaye (2020) [46]EthiopiaCross-sectionalWHO MNM303NR48NRNRAlemu (2019) [47]South SudanCross-sectionalWHO MNM, morbidity criteria1010993311031.3Beyene (2022) [67]EthiopiaFollow up StudyWHO MNM modified30062,880901831.3Danusa (2022) [69]EthiopiaUnmatched case-controlWHO MNM modified664NR166NRNRHabte (2021) [30]EthiopiaUnmatched case-controlWHO MNM322NR81NRNRKalisa (2016) [59]RwandaFollow up StudyWHO MNM3,9793,994861321.5Lilungulu (2020) [60]TanzaniaCross sectionalWHO MNM3,6013,48012416NRLiyew (2017) [26]EthiopiaCross-sectionalWHO MNM2,79529,697238NR8.01Nakimuli (2016) [61]UgandaCohort studyWHO MNM3,10025,84069513026.9Nelissen (2013) [68]TanzaniaCross-sectionalWHO MNM modified9,4719,1362163223.6Sayinzoga (2017) [25]RwandaProspective case-controlWHO MNM modified6425,5572011336Teshome (2022) [62]EthiopiaUnmatched case-controlWHO MNM264NR88NRNRTura (2018) [71]EthiopiaCohort studyWHO MNM, WHO MNM modified1,0547,4041282617Abdulrazak (2021) [51]EthiopiaCross-sectionalWHO MNM1667NR344NR72Pembe (2019) [52]TanzaniaCross-sectionalWHO MNM, WHO MNM modified120,553NR2,15820217.9UgandaCross-sectionalWHO MNM, WHO MNM modified163,559NR6,53524040Wakgar (2019) [63]EthiopiaCross-sectionalWHO MNM15,059NR5019033.3Woldeyes (2018) [48]EthiopiaCross-sectionalWHO MNM3642,7371382450.4Yayehyirad (2020) [64]EthiopiaCross-sectionalWHO MNM8455,53021,011737.97Ayele (2014) [65]EthiopiaCross-sectionalWHO MNM8,9898,5092062324.2Mekonnin (unpublished) [66]EthiopiaCross-sectionalWHO MNM modified2,694NR287NRNRNdingori (2022) [24]KenyaCase-controlWHO MNM225NR45NRNRMulongo (2022) [22]KenyaCross sectionalWHO MNM9,8239770114NR11.7Van der (2021) [49]TanzaniaCross-sectionalWHO MNM, WHO MNM modified97194810511 Table 1Characteristics of included primary studies reporting maternal near miss (MNM) in East Africa between 2011 and 2023First Author (publication year, reference)CountryStudy designMeasurement usedParticipated womenLivebirthsNo of MNMMDMNM per 1000 livebirthsTeka (2022) [21]EthiopiaCross-sectionalWHO MNM modified6915,1161462428.5Litorp (2014) [53]TanzaniaCross-sectionalWHO MNM13,58413,1214677736Gebremariam (2022) [31]EthiopiaCross sectionalWHO MNM905NR129NRNREgal (2022) [70]SomaliaCross-sectionalWHO MNM, WHO MNM modified6,6586,055792813Shegaw(2021) [27]EthiopiaCross-sectionalWHO MNM, WHO MNM modified1,2441,17348340.9Dahie (2022) [23]SomaliaUnmatched Case ControlWHO MNM533NR178NRNRLiyew (2018) [50]EthiopiaNested case-controlWHO MNM864NR216NRNRGeze (2021) [29]EthiopiaUnmatched nested case ControlWHO MNM modified432NR108NRNRKiruja (2017) [54]SomaliaCross sectionalWHO MNM1,38513551201888.6Rulisa (2015 [55]RwandaCross-sectionalWHO MNM1,739173914250-Tariku (2020) [44]EthiopiaCross sectionalWHO MNM modified337NR594NRHerklots (2017) [56]TanzaniaCross-sectionalWHO MNM55514,12537289De Haan (2023) [45]EthiopiaCross-sectionalWHO MNM modified637NR700NRGeleto (2020) [57]EthiopiaCross-sectionalWHO MNM78,195323,82477,714481239.9Kumela (2020) [17]EthiopiaCase-controlWHO MNM modified183NR61NRNROwolabi (2020) [58]KenyaCross-sectionalWHO MNM69536,168360179.97Asaye (2020) [46]EthiopiaCross-sectionalWHO MNM303NR48NRNRAlemu (2019) [47]South SudanCross-sectionalWHO MNM, morbidity criteria1010993311031.3Beyene (2022) [67]EthiopiaFollow up StudyWHO MNM modified30062,880901831.3Danusa (2022) [69]EthiopiaUnmatched case-controlWHO MNM modified664NR166NRNRHabte (2021) [30]EthiopiaUnmatched case-controlWHO MNM322NR81NRNRKalisa (2016) [59]RwandaFollow up StudyWHO MNM3,9793,994861321.5Lilungulu (2020) [60]TanzaniaCross sectionalWHO MNM3,6013,48012416NRLiyew (2017) [26]EthiopiaCross-sectionalWHO MNM2,79529,697238NR8.01Nakimuli (2016) [61]UgandaCohort studyWHO MNM3,10025,84069513026.9Nelissen (2013) [68]TanzaniaCross-sectionalWHO MNM modified9,4719,1362163223.6Sayinzoga (2017) [25]RwandaProspective case-controlWHO MNM modified6425,5572011336Teshome (2022) [62]EthiopiaUnmatched case-controlWHO MNM264NR88NRNRTura (2018) [71]EthiopiaCohort studyWHO MNM, WHO MNM modified1,0547,4041282617Abdulrazak (2021) [51]EthiopiaCross-sectionalWHO MNM1667NR344NR72Pembe (2019) [52]TanzaniaCross-sectionalWHO MNM, WHO MNM modified120,553NR2,15820217.9UgandaCross-sectionalWHO MNM, WHO MNM modified163,559NR6,53524040Wakgar (2019) [63]EthiopiaCross-sectionalWHO MNM15,059NR5019033.3Woldeyes (2018) [48]EthiopiaCross-sectionalWHO MNM3642,7371382450.4Yayehyirad (2020) [64]EthiopiaCross-sectionalWHO MNM8455,53021,011737.97Ayele (2014) [65]EthiopiaCross-sectionalWHO MNM8,9898,5092062324.2Mekonnin (unpublished) [66]EthiopiaCross-sectionalWHO MNM modified2,694NR287NRNRNdingori (2022) [24]KenyaCase-controlWHO MNM225NR45NRNRMulongo (2022) [22]KenyaCross sectionalWHO MNM9,8239770114NR11.7Van der (2021) [49]TanzaniaCross-sectionalWHO MNM, WHO MNM modified97194810511 Characteristics of included primary studies reporting maternal near miss (MNM) in East Africa between 2011 and 2023 Maternal near-miss incidence ratio Out of the included studies, 26 studies contained pertinent data for the pooled estimation of the MNM incidence ratio [21, 26, 27, 47, 53–61, 67, 70]. Studies included for estimating maternal near-miss ratio were primarily cross-sectional study designs and cohort/follow-up studies. However, one prospective case-control study contained all the pertinent information needed for the outcome of interest and was included for MNM incidence ratio pool estimation [25]. The maternal near miss incidence ratio in the included studies ranged from 8 per 1000 live births in Rwanda [55] to 239.9 per 1000 live births in Ethiopia [57]. The pooled weighted maternal incidence ratio was 37.12 MNM per 1000 live births (95%CI:20.22–54.03). The forest plot indicated significant heterogeneity between the studies (I2 = 99.94%, Cochran Q-statistic P < 0.001) (Fig. 2). Therefore, we conducted a subgroup analysis by country, showing that the lowest MNM incidence ratio was in Kenya, at 12 MNM per 1000 live births (95% CI 8.95–12.19) (I2 = 51.6%, p = 0.15), and the highest was in Ethiopia, at 53 MNM per 1000 live births (95% CI 15.02–91.12) (Fig. 3). Further sensitivity analysis (leave-one-out influential analysis) was conducted, and the result showed one study [57] was largely responsible for the influence of the overall effect size and when the study by Geleto was removed, the MNM incidence ratio reduced to 28.89 per 1000 livebirths. The inclusion of influential studies may affect the validity and robustness of the overall results. Otherwise, the leave-one-out influential analysis showed that no other studies substantially influence the result of pooled MNM incidence ratio (the detail depicted in Additional file 3). The funnel plot appears symmetrical indicating there was no publication bias (Additional file 4) and the Egger test shows no significant small study bias (P value = 0.52). Thus, trim-and -fill technique was not necessary. Fig. 2Forest plot shows the pooled maternal near-miss incidence ratio estimate in East Africa. Each square box indicates the individual estimate, the diamond-shaped box indicates the pooled estimate of MNM incidence ratio, and the box on the left of the vertical line indicates the effect size of the individual study lower than the pooled estimate. In contrast, the box on the right indicates higher than the pooled estimate Forest plot shows the pooled maternal near-miss incidence ratio estimate in East Africa. Each square box indicates the individual estimate, the diamond-shaped box indicates the pooled estimate of MNM incidence ratio, and the box on the left of the vertical line indicates the effect size of the individual study lower than the pooled estimate. In contrast, the box on the right indicates higher than the pooled estimate Fig. 3Forest plot shows the subgroup analyses of MNM incidence ratio by country in East Africa. Each square box indicates the individual estimate, the red diamond-shaped box indicates the pooled estimate of MNM incidence ratio in each country, and the green diamond-shaped box indicates the overall pooled estimate of MNM incidence ratio Forest plot shows the subgroup analyses of MNM incidence ratio by country in East Africa. Each square box indicates the individual estimate, the red diamond-shaped box indicates the pooled estimate of MNM incidence ratio in each country, and the green diamond-shaped box indicates the overall pooled estimate of MNM incidence ratio Causes of maternal near miss Out of 39 included studies, 28 studies contained pertinent data on causes of maternal near-miss for pooling [17, 19, 21–24, 26, 29–31, 45, 48, 52–54, 56, 57, 59–61, 64, 66–71]. Each included study had reported at least two major obstetric causes of maternal near miss. Two studies reported that the causes of MNM were not included in our pool estimation for different reasons. For instance, a study by Owolabi et al. [58] reported causes of MNM, but the reported causes were not mutually exclusive, which could lead to an overestimation in the pool estimation. Due to this issue, the study was not included in the pool estimation of the cause proportion of MNM. In another study from Tanzania [49], causes of MNM for women who had shock and cardiac arrest were not specifically reported. This lack of specific cause identification might underestimate the proportion of a cause. As a result, the study was not included in the pool estimate of the cause proportion of MNM. The summary of the causes of MNM reported in primary studies is presented (Additional file 5). All studies included for the causes pool estimation had pertinent data for the obstetric haemorrhage pool analysis. In random effect meta-analysis, the pooled proportion of obstetric haemorrhage as a cause of maternal near miss was 29.9% with a 95% CI of 25.7% to 34.0%. The forest plot indicated heterogeneity among included studies (I2 = 97.92%), p < 0.001) (Table 1). Subgroup analysis by country showed that the lowest proportion of obstetric haemorrhage occurred in Kenya at 18.5% (95% CI: 12.3–24.7, I2 = 0.00%), while the highest proportion occurred in Uganda at 38.1% (95% CI: 15.4–60.9, I2 = 99.4%). The pooled proportion of hypertensive disorder of pregnancy as a cause of MNM was 27.8% (95% CI: 22.3–33.3). The forest plot indicated high heterogeneity between studies (I2 = 99.34%, p < 0.0001). For the heterogeneity, subgroup analysis performed by country shows that the cause proportion of hypertensive disorder in Rwanda was 19.4%, while in Somalia, it was 43.4% with a 95% CI of 25.2–61.6%. In the random effect pooled analysis, the proportion of uterine rupture as a cause of MNM was 13.8% (95% CI: 7.9–19.6). Forest plot results revealed high heterogeneity between studies (I2 = 99.66%, p < 0.0001). Subgroup analysis by country shows the highest in Ethiopia, 19.0% (95% CI 9.4–28.8%, I2 = 98.26%) and the lowest in Somalia, 2.6% (95%CI:0.8–4.4%, I2 = 0.00%). Sepsis/systemic infection causes of maternal near miss had been reported in 27 primary studies, and the pooled proportion of sepsis was 9.9% (95%CI: 7.1–12.7). High heterogeneity was identified between studies (I2 = 98.3%, p < 0.0001). Subgroup analysis by country revealed that in Somalia, 2.9% (95% CI: 1.2–4.5, I2 = 0.00%) and in Uganda, 14.9% (95% CI: 8.9–21.0, I2 = 96.45). In the pooled result, abortion accounted for 8.1% (95% CI: 5.9–10.3, I2 = 96.54%). Subgroup analysis by country shows in Ethiopia it was 6.1% (95% CI:4.2–8.0.2.0, I2 = 74.4%) and Rwanda 12.8% (95% CI: 3.7–21.9, I2 = 81.57%). Ectopic pregnancy also accounted for 7.3% (95% CI: 3.8%−11.5%, I2 = 99.3%). Subgroup analysis by country shows that in Ethiopia, 3.7% (95% CI: 1.0–6.4), and in Tanzania, 13.9% (Table 2).Table 2The overall pooled proportion of direct obstetric causes of MNM and subgroup random effect model analysis in East Africa between 2011 and 2023CausesNumber of studiesVariablesCharacteristicsProportion per 100 with 95%CII2Obstetric haemorrhage [17, 19, 21–24, 26, 29–31, 45, 48, 52–54, 56, 57, 59–61, 64, 66–71]28East Africa29.9% (25.7–34.0%)97.92%CountryEthiopia29.5% (23.0–36.0%)96.0%Kenya18.5% (12.3–24.7%)0.00%Rwanda34.0% (13.2–55.0%)91.78%Somalia29.0% (17.0–41.0%)85.3%Tanzania31.0% (25.1–36.3%)86.5%Uganda38.1% (15.4–60.9%)99.4%Hypertensive disorder [17, 19, 21–24, 26, 27, 30, 31, 48, 52–54, 56, 57, 59–61, 64–71]27East Africa27.8% (22.3–33.3%)99.34%CountryEthiopia28.6% (20.8–36.2%)97.4%Kenya29.5% (22.3–36.8%)0.00%Rwanda19.4% (2.9–41.9%)95.36%)Somalia43.4% (25.2–61.6%)92.5%Tanzania21.4% (7.6–35.2%)98.2%Uganda21.0% (3.6–41.0%)99.1%Uterine rupture [17, 23, 29–31, 48, 52–54, 56, 57, 60, 61, 64–68, 71]20East Africa13.8% (7.9–19.6%)99.66%CountryEthiopia19.0% (9.0–28%)97.78%Somalia2.6% (0.8–4.3%)0.00%Tanzania8.3% (7.2–9.5%)0.23%Uganda11.6% (2.2–21.0%)98.3%Sepsis [17, 19, 21–23, 26, 27, 30, 31, 48, 52–54, 56, 57, 59–61, 64, 65, 67–71]26East Africa9.9% (7.1–12.8%)98.3%CountryEthiopia12.5% (7.3–17.7%)96.5%Kenya9.4% (3.8–14.9%)0.00%Rwanda7.0% (4–11%)0.00%Somalia2.9% (1.2–4.5%)0.00%Tanzania7.3% (3.4–11%)88.89%Uganda14.9% (8.9–21.0%)9.63%Abortion [17, 19, 21, 23, 26, 29, 48, 52, 53, 57, 59, 61, 65–68, 70, 71]18East Africa8.1% (5.9–10.3%)96.54%CountryEthiopia6.1% (4.2–8.0.%)74.4%Rwanda12.8 (3.8–21.9%)81.57%Somalia6.1% (1.6–10.6%)58.2%Tanzania10.2% (1.9–16.0%)98.0%Uganda11.3% (0.2–22%)99.24%Ectopic pregnancy [17, 27, 52, 53, 57, 59, 65, 66, 71]9East Africa7.3% (3.1–11.5%)99.38%)CountryEthiopia3.7% (1.0–6.4%)95.15%Tanzania13.9% (1.6–29.0%)99% Table 2The overall pooled proportion of direct obstetric causes of MNM and subgroup random effect model analysis in East Africa between 2011 and 2023CausesNumber of studiesVariablesCharacteristicsProportion per 100 with 95%CII2Obstetric haemorrhage [17, 19, 21–24, 26, 29–31, 45, 48, 52–54, 56, 57, 59–61, 64, 66–71]28East Africa29.9% (25.7–34.0%)97.92%CountryEthiopia29.5% (23.0–36.0%)96.0%Kenya18.5% (12.3–24.7%)0.00%Rwanda34.0% (13.2–55.0%)91.78%Somalia29.0% (17.0–41.0%)85.3%Tanzania31.0% (25.1–36.3%)86.5%Uganda38.1% (15.4–60.9%)99.4%Hypertensive disorder [17, 19, 21–24, 26, 27, 30, 31, 48, 52–54, 56, 57, 59–61, 64–71]27East Africa27.8% (22.3–33.3%)99.34%CountryEthiopia28.6% (20.8–36.2%)97.4%Kenya29.5% (22.3–36.8%)0.00%Rwanda19.4% (2.9–41.9%)95.36%)Somalia43.4% (25.2–61.6%)92.5%Tanzania21.4% (7.6–35.2%)98.2%Uganda21.0% (3.6–41.0%)99.1%Uterine rupture [17, 23, 29–31, 48, 52–54, 56, 57, 60, 61, 64–68, 71]20East Africa13.8% (7.9–19.6%)99.66%CountryEthiopia19.0% (9.0–28%)97.78%Somalia2.6% (0.8–4.3%)0.00%Tanzania8.3% (7.2–9.5%)0.23%Uganda11.6% (2.2–21.0%)98.3%Sepsis [17, 19, 21–23, 26, 27, 30, 31, 48, 52–54, 56, 57, 59–61, 64, 65, 67–71]26East Africa9.9% (7.1–12.8%)98.3%CountryEthiopia12.5% (7.3–17.7%)96.5%Kenya9.4% (3.8–14.9%)0.00%Rwanda7.0% (4–11%)0.00%Somalia2.9% (1.2–4.5%)0.00%Tanzania7.3% (3.4–11%)88.89%Uganda14.9% (8.9–21.0%)9.63%Abortion [17, 19, 21, 23, 26, 29, 48, 52, 53, 57, 59, 61, 65–68, 70, 71]18East Africa8.1% (5.9–10.3%)96.54%CountryEthiopia6.1% (4.2–8.0.%)74.4%Rwanda12.8 (3.8–21.9%)81.57%Somalia6.1% (1.6–10.6%)58.2%Tanzania10.2% (1.9–16.0%)98.0%Uganda11.3% (0.2–22%)99.24%Ectopic pregnancy [17, 27, 52, 53, 57, 59, 65, 66, 71]9East Africa7.3% (3.1–11.5%)99.38%)CountryEthiopia3.7% (1.0–6.4%)95.15%Tanzania13.9% (1.6–29.0%)99% The overall pooled proportion of direct obstetric causes of MNM and subgroup random effect model analysis in East Africa between 2011 and 2023 For each direct major obstetric cause of MNM, leave-one-out influential analysis was performed to check whether any of the individual studies influenced the pooled result. The result shows in proportion of uterine rupture, omitted study by Ayele et al. [65] reduced proportion of uterine rupture by 2.8% (Additional file 3). Risk factors of MNM Sociodemographic factors In this meta-analysis, data pooled from eight studies [23, 27, 29–31, 44, 62, 69] being aged over 35 years (AOR = 1.28, 95%CI: 0.75–1.81), thirteen studies [21, 23, 27, 29, 30, 44–47, 50, 51, 62, 69] rural residency (AOR = 1.38, 95%CI: 0.94–1.83) and five studies [23, 30, 44, 62, 69] educational status (AOR = 1.9, 95%CI: 0.92–2.88) was not found to be statistically associated with MNM (Fig. 4). However, there was significant heterogeneity between studies. In this study pooled result from seven studies [22–24, 44, 45, 62, 69], women’s employment status was statistically associated with MNM (AOR = 0.53, 95%CI: 0.21–0.84) (Fig. 5). This indicates women who are employed had a 47% reduced likelihood of MNM compared to unemployed women. The pooled estimate shows women from households with a low monthly income were at 2.9 times increased risk of MNM compared with households having a high monthly income (AOR = 2.95, 95%CI: 1.28–4.61). Fig. 4Forest plot of the pooled effect sizes of age and residency on MNM in East Africa Forest plot of the pooled effect sizes of age and residency on MNM in East Africa Fig. 5Forest plot showing the pooled effect size of education, employment, and family income on MNM in East Africa Forest plot showing the pooled effect size of education, employment, and family income on MNM in East Africa Obstetric history factors In the pooled analysis from five studies [21, 23, 27, 31, 46], women with a parity greater than five had an increased risk of MNM; that is, they were 1.65 times more likely to experience MNM compared with low parity women (AOR = 1.65, 95%CI: 1.14–2.16) (Fig. 6). Pooled result from four studies [27, 29, 30, 53] revealed that women with a history of caesarean section (CS) were 5 times more likely to have MNM than those without a history of CS (AOR = 5.04, 95%CI: 3.1–6.90). The pooled effect-sized result from three studies [27, 29, 31] shows women with a history of abortion were 1.63 times more likely to experience MNM than those without an abortion history (AOR = 1.63, 95%CI: 1.04–2.22). A history of stillbirth was statistically insignificant (AOR = 1.53, 95%CI: 0.81–2.24), as was a birth interval between children of less than 24 months (AOR = 3.59, 95%CI: 0.65–7.83). However, women who had a history of pregnancy complications were four times more likely to experience MNM compared to women without a history of complications (AOR = 4.09, 95%CI: 1.39–6.78); although there was high heterogeneity between studies (I2 = 78.5%). Fig. 6Forest plot showing the pooled effect of obstetric variables on MNM in East Africa Forest plot showing the pooled effect of obstetric variables on MNM in East Africa Maternal health service uses related factors In this pooled analysis of ten studies [17, 21, 23, 27, 29, 30, 50, 51, 62, 69], women lacking ANC utilisation were 2.91 times more likely to experience MNM than those had ANC (AOR = 2.91, 95%CI:2.02–3.78) (Fig. 7). Furthermore, data from two studies [23, 30] women who lacked BPCR were four times more likely to experience MNM compared to women who had BPCR (AOR = 4.18, 95%CI: 2.64–5.72). However, women’s autonomous decision-making (AOR = 1.82, 95%CI: 0.65–2.99, I2 = 2.34%), delays deciding care (AOR = 1.36, 95%CI: 0.14–2.58), delays in reaching the health facility(AOR = 1.83, 95%CI: 0.15–3.52), and delays [62]in receiving appropriate health care (AOR = 1.37, 95%CI: 0.97–1.77) were a statistically insignificant. However, women living a long distance from health facility were 3 times more likely to experience MNM compared with those living closer (AOR = 3.06, 95%CI:1.09–5.03, I2 = 18.6%) (Fig. 8). Data pooled from four studies [23, 44, 46, 62] revealed that women admitted via provider referral were 69% less likely to experience MNM compared to women who were self-referred (AOR = 0.31, 95%CI: 0.07–0.55, I2 = 68.3%). Fig. 7Forest plot showing the pooled effect size of maternal health service-related variables on MNM in East Africa Forest plot showing the pooled effect size of maternal health service-related variables on MNM in East Africa Fig. 8Forest plot showing the effect of delays, admistion type and distance to health facility on MNM in East Africa Forest plot showing the effect of delays, admistion type and distance to health facility on MNM in East Africa Discussion Herein, we report the first pooled estimates of Maternal Near Miss (MNM) based on WHO MNM criteria in East Africa. The findings show the weighted pooled MNM incidence ratio was 37.12 MNM (95% 20.22–54.03) per 1000 live births. In addition, our results indicate that obstetric haemorrhage, hypertensive disorder of pregnancy, and uterine rupture are the three leading causes of MNM, accounting for approximately 70% of cases. Our findings also demonstrate that the women’s occupation, household income, parity, history of abortion, history of CS, history of pregnancy-related complications, BPCR, distance and admission type to the health facility were significantly associated with MNM. The findings of the MNM incidence ratio in this study is in line with a systematic review study using WHO MNM criteria in Africa and India [33]. However, the rates herein were higher than reported across 27 Asia Pacific countries [72], and low and middle-income countries from 2009 to 2020 [73]. A review study shows global MNM incidence ratio is similarly lower than our findings [33]. It is obvious that maternal complication and MNM is expected to be high as maternal mortality in resource-limited settings. This may be due to differences in access and quality of maternal health services, including emergency obstetric care services, as well as low utilisation of maternal health services. In low-income countries, including those in East Africa, only 21% women have access to emergency obstetric care [74]. Evidence-based interventions to prevent MNM and maternal death have existed for a long time; however, in low-resource settings, these interventions are unavailable and inaccessible [2]. The other possible reasons for high MNM ratio in the current study might be due to the low quality of maternal health care service [47], which can occur since women can have potential life-threatening complications at admission to hospitals and receive inadequate interventions such as blood transfusion, intensive care unit admission, and dialysis [26, 54]. However, previous findings reported in Ethiopia is higher than our findings [75]. The plausible reasons for this might be in the previous review study, various tools were used to measure MNM, for example complications specific criteria is one of the MNM measurement used, this criteria might have been result in high magnitude of MNM and also the denominator used in the previous study was total participants, this also likely to result in high maternal near miss incidence ratio. In contrast, our study used total live births as the denominator, in accordance with the WHO’s recommendations [14]. In this study, we also investigated the MNM incidence ratio by countries, with the highest burden identified in Ethiopia and the lowest in Kenya. The MNM incidence ratio in Ethiopia was 53 MNM per 1000 live births and in Somalia, 50 MNM per 1000 live births. This result is lower than findings from other meta-analyses in Ethiopia [75] and Indonesia [72]. The possible reasons might be that the former studies used both WHO MNM and disease-specific criteria, whereas disease-specific criteria would inflate the magnitude of MNM, because not all women with preeclampsia/eclampsia and severe obstetric haemorrhage may not fulfil MNM criteria [72]. However, the current findings of both countries is higher than Brazil [33]. In Kenya, the MNM IR was found to be 12 per 1,000 live births. These results imply that the maternal near-miss incidence ratio varies between East African countries, ranging from 53 per 1,000 live births in Ethiopia to 12 per 1,000 live births in Kenya. This might be explained by inequalities in maternal health service utilisation and a high unmet need for emergency obstetric and newborn care services in Ethiopia [76]. Additionally, delay in seeking care for emergency obstetric complications, late admission to health facilities, and poor referral from basic to comprehensive health facilities contribute to high MNM. The first leading obstetric cause of MNM was found to be obstetric haemorrhage, causing 29.9% of the cases. This supports other literature, which has found postpartum haemorrhage to be the leading cause of MNM in the Pacific Asia [72] and in middle-income countries [73], and in low-resource countries [77]. This occurs despite the presence of evidence-based interventions, such as blood transfusion, CS, uterotonic drug, active third stage management, manual removal of placenta and retained product, and hysterectomy [78]. The possible reasons for the high proportion of haemorrhage-related MNM might be inadequate treatment and factors such as previous CS, abnormal placentation, and prolonged labour, which increase the risk of postpartum haemorrhage [79]. For example, a significant number of women with antepartum haemorrhage who needed blood transfusion did not receive blood transfusion [58]. In several studies, blood transfusion was the least provided emergency obstetric care signal function in hospitals [57]. Inadequate monitoring, failure to get better care for the abnormal vital sign, insufficient communication between providers, and lack of timely involvement of a senior physician could increase severe obstetric haemorrhage [16]. Other reasons might be that many women in resource-limited settings experience anaemia during pregnancy [26], which results in an increased risk of MNM due to obstetric haemorrhage [79]. Since women with low pre-labour haemoglobin levels are more likely to experience deterioration in vital signs. In case of severe obstetric haemorrhage, there can be consumption of clotting factors and platelets, leading to a coagulopathy characterised by low fibrinogen and thrombin [80]. This can result in uncontrolled bleeding, leading to increased MNM. In the current review, hypertensive disorders of pregnancy were the second leading cause of MNM, accounting for 27.8% cases. Similar findings reported in middle-income countries and Pacific Asia found hypertensive disorder as the second leading cause of MNM [72, 73], with eclampsia found to be the leading cause in South Asia [72]. This complication increases a significant risk to both mother and fetus; timely intervention can make a substantial difference in the outcome. ANC is the ideal contact point for pregnant women to ensure early identification of hypertensive disorder of pregnancy and prevent the progression of complications. However, maternal health service utilisation is low in many resource-limited countries and also for severe preeclampsia/eclampsia, the provision of magnesium sulphate is inadequate in many settings in East African countries [58]. Inadequate provision of Aspirin for prevention of hypertensive disorder of pregnancy for high-risk women during pregnancy in low-income countries [81]. In this study, uterine rupture was the third leading cause of MNM, accounting for 13.8% cases. This might be due to the magnitude of obstructed labour being high, and access to emergency caesarean section is limited in resource-limited settings. Additionally, the late presentation of women in health facilities. Thus, obstructed labour increases the risk of uterine rupture [82]. For example, a review study in Ethiopia 41% of obstructed labour ends up with uterine rupture. This could be due to delay or lack of emergency obstetric interventions, inadequate labour monitoring and high home childbirth [83]. The appropriate use of a partograph for labour monitoring may reduce the risk of uterine rupture through early detection of obstructed labour. The findings of this study revealed that the pooled proportion of sepsis/systemic infection contributing to MNM was 9.9%. This aligns with the previous review studies that reported sepsis contributing 7.5% of MNM cases [77]. The contributions of sepsis to MNM have been reducing. This could be because in many studies sepsis had a high mortality index, or might be due to improvement of clean delivery service and parental antibiotics for prolonged rupture of membrane, and prophylaxis antibiotics for caesarean Sect [58]. As the study reported, parental antibiotic provision was the most frequently applied emergency obstetric signal function [57]. However, our subgroup analysis by country shows that sepsis was the third most common obstetric cause of MNM in Rwanda (15%). This may be due to inequalities in the implementation of obstetric signal functional services. In the pool effect-sized estimate, several factors were found to be statistically associated with MNM; women who are employed had a 44% reduction in MNM compared to the unemployed. Women from low-income households were 2.95 times more at risk of MNM. This might be a barrier related to transportation and the cost of treatment, which affects women’s timely access to maternal health services, including emergency obstetric care services [84]. The inability of poor women to access quality maternal health services because of non-existent or weak financing strategies that require out-of-pocket payments is a complex and multifaceted problem, which makes women fail to get timely obstetric care [6]. Evidence suggested that health insurance coverage in East Africa remains low [85], which has significant implications for healthcare access and affordability. In the pooled analysis, grand multiparity was associated with an increased risk of MNM and 1.65 times more likely to experience MNM compared with low parity. This might be the risk that the obstetric cause of MNM can be increased with parity, for example, grand parity increases the risk of obstetric haemorrhage due to uterine atony because the uterine muscle cannot contract enough to close placental bleeding vessels after childbirth. This made uterine atony the leading cause of postpartum haemorrhage [86]. We found women with a history of abortion were 1.63 times more likely to experience MNM than those without an abortion history. This is supported by previous evidence, which shows that a history of abortion increases the risk of severe maternal morbidity [87]. An explanation for this is that women with a history of abortion might be more likely to have abnormal placentation, such as placenta accreta, previa, and this abnormal placentation increases the risk of antepartum haemorrhage [88]. In this study, women with a history of caesarean section were 5 times more likely to have MNM than those without a history of CS. Meta-analysis studies support this result [75, 87]. The increased risk of uterine rupture and risk of abnormal placentation, such as placenta previa, and accerta might explain this [16, 89]. Additionally, after a previous caesarean section, induction of labour using oxytocin during the trial of labour increased the risk of uterine rupture [90], increases the MNM. To date, the rate of repeated CS is increasing, which is more likely to increase the risk of uterine rupture. This implies a special focus is needed during pregnancy and labour for women who have a history of CS. Additionally, as the CS rate has been increasing [91]. This may imply that primary CS should be performed only when clearly indicated, as overuse and unmet need for CS increase complications. Implementation of the WHO Labour Care Guide can reduce the primary CS rate without increasing fetal and maternal complications [92]. Women with a history of pregnancy complications were four times more likely to experience MNM compared to those without a history of complications. This might be some obstetric complications, for example, hypertensive disorder of pregnancy, which is more likely to recur in future pregnancies. The implication of this finding highlights a critical health system failure in effectively screening for a history of obstetric complications during pregnancy and in implementing targeted, evidence-based interventions [93]. Thus, women with previous pregnancy complications need special attention during ANC, childbirth, and the postpartum period. In our pooled analysis, women who lacked ANC utilisation were 2.91 times more likely to experience MNM than those with ANC. Previous studies support this finding; ANC use decreases the risk of MNM by two-thirds to three-fourths [75, 94, 95]. ANC provides an opportunity to identify women with a history of obstetric complications or danger signs during the index pregnancy and childbirth, enabling timely referral to appropriate health services for specialised care. According to WHO suggestions, 15% of pregnant women attending ANC are expected to be put under a specialised category that demands high-level care; these women include women who have had previous CS, previous pregnancy-related complications, and chronic medical conditions. The pooled estimate indicated that women who lacked BPCR practice were 4 times more likely to experience MNM compared to women who had BPCR practice. This increased risk may be attributed by women’s inadequate preparation for childbirth and complications, limited ability to recognise danger signs, and delays in seeking emergency obstetric care following obstetric complications and labour. Such gaps allow obstetric complications to progress into life-threatening conditions. But with improved comprehensive emergency obstetric care services in place, more women can survive life-threatening complications. This leads to high MNM by reducing maternal deaths. Likewise, women with inadequate BPCR practices and poor quality of emergency obstetric care resulted in high life-threatening complications (both MNM and maternal deaths). WHO suggests BPCR is one of the first health promotion and prevention interventions in the ANC package that is effective in reducing maternal morbidity and mortality [96]. Improving birth preparedness and complication readiness practices, alongside the provision of high-quality emergency obstetric care services, plays a critical role in reducing life-threatening maternal complications, including MNM. Women living a long distance from a health facility were 3 times more likely to experience MNM compared with those living closer. In particular, long distances to comprehensive health facilities can significantly influence the progression of potential life-threatening complications. A study by Ebener et al. highlights the importance of reducing travel times for emergency obstetric and newborn care (EmONC), optimising the referral process from Basic EmONC (BEmONC) to Comprehensive EmONC (CEmONC) health facilities [97]. These are essential steps towards universal health coverage and improving maternal and newborn health outcomes by making healthcare services more accessible and equitable. Healthcare can better meet the healthcare needs of pregnant and labouring women. A study in Mozambique discussed that the combination of upgrading geographical strategic health facilities and providing transportation across the midlevel health facility can increase women’s access to high-level care within 2 h and reduce transfer time between facilities by 39% [98]. These findings highlight the importance of expanding maternity waiting homes and utilisation by women for those who have difficulty accessing health facilities due to geographical access. In this meta-analysis, we found that women admitted by provider referral were 69% less likely to experience MNM compared to those who self-referred. This may be explained by the fact that women’s first visit to the nearest health facility may receive pre-referral treatment and help them reach the referral facility before obstetric complications worsen. While women bypass the nearest low-level health facility to go to a high-level health facility, this may lead to their complication status worsening and result in them having some organ dysfunction. Additionally, women unable to recognise emergency conditions early are often compounded by a delay in seeking care from health facilities, with a significant proportion of women seeking care from informal sources. The high MNM incidence ratio reflects underlying challenges in the quality and utilisation of general maternal healthcare services. Limited access to routine maternal health care services, coupled with poor service quality, hinders women’s awareness of obstetric complications, increasing the risk of MNM. Women’s failure to recognise complications and delays in accessing emergency obstetric care further exacerbates the risk, contributing to life-threatening complications such as MNM and maternal deaths. Additionally, healthcare providers’ failure in early detection of complications and provision of appropriate management during pregnancy, childbirth, and the postpartum period contributes to women experiencing life-threatening complications. Conversely, improvements in the quality of emergency obstetric care, even in the context of poor routine maternal healthcare services, can increase the likelihood of survival even when women present late with severe complications. These can contribute to high MNM incidence by reducing maternal deaths. In contrast, the presence of inadequate routine maternal health services combined with poor-quality emergency obstetric care services significantly contributes to both high MNM and maternal deaths. Thus, the implementation of the new WHO Labour Care Guide can facilitate early detection of labour abnormalities and prompt clinical decision-making [99], which may significantly reduce the incidence of maternal near miss (MNM) events. Limitations of the study The absence of primary studies on MNM in six East African countries, as per the WHO MNM criteria or its modification in the local context, is a limitation that can impact the generalizability of the findings to the whole of East African region. The merging of sub causes and calculating some data from reported results may introduce some imprecision. When possible, we used the frequencies reported in the main causes, but when both the frequencies of main causes and sub-cause frequencies were reported in the primary studies, the frequencies of each sub-cause under specific main causes were summed up to get the frequencies of main causes. We found considerable heterogeneity and employed subgroup analysis to identify the source of heterogeneity. Additionally, an influential analysis was conducted to investigate the impact of each study on the overall effect and identify the influential studies. Thus, the interpretation of the estimates in this study needs to consider the possible limitations. Conclusion and implications In this study, the MNM incidence ratio was found to be high in East Africa. Obstetric haemorrhage, hypertensive disorder of pregnancy and uterine rupture are the three leading obstetric causes of MNM in descending order. Low household income, lack of ANC, grand multiparous, abortion history, history of CS, history of obstetric complication, lack of BPCR, and long distance to health facility are factors that increase the risk of MNM, while employed status and being referred to by health care providers are the protective factors of MNM. These findings indicate that special attention is required for women with a history of caesarean section, history of abortion and previous pregnancy-related complications. The high MNM incidence ratio can have significant implications for maternal health and the healthcare system in East Africa. Risk screening during pregnancy, followed by providing targeted evidence-based interventions and the WHO Labour Care Guide implementation, may reduce MNM. Priority setting on interventions that target obstetric haemorrhage, preeclampsia/eclampsia, and uterine rupture may extensively reduce MNM and improve maternal health. Improving women’s employment status and enhancing ANC use would help significantly to reduce the MNM incidence ratio. The findings may give insight for advocacy purposes for donor organisations interested in maternal health in resource-limited settings. Additionally, future research on the trend of the MNM incidence ratio may help to understand the impact of interventions and inform the development of appropriate policies and strategies. Supplementary Information Additional file 1: Search strategy used in database search for MNM incidence ratio, causes and risk factors in East Africa. Additional file 1: Search strategy used in database search for MNM incidence ratio, causes and risk factors in East Africa. Additional file 2: Quality assessment of the primary studies using JBI critical appraisal tools. Additional file 2: Quality assessment of the primary studies using JBI critical appraisal tools. Additional file 3: Forest plot showing the influential leave-one-out meta-analysis of MNM incidence ratio and causes of MNM in East Africa. Additional file 3: Forest plot showing the influential leave-one-out meta-analysis of MNM incidence ratio and causes of MNM in East Africa. Additional file 4: Funnel plot showing the publication bias for MNM incidence ratio in East Africa. Additional file 4: Funnel plot showing the publication bias for MNM incidence ratio in East Africa. Additional file 5: Summary on causes of maternal near miss (MNM) and risk factors reported in primary studies from East Africa between 2011 to 2023. Additional file 5: Summary on causes of maternal near miss (MNM) and risk factors reported in primary studies from East Africa between 2011 to 2023.