The impact of armed conflict on maternal morbidity and mortality at a teaching hospital in the Tigray region of Ethiopia: a pre-war and wartime comparative analysis

Abstract

Background

Armed conflict profoundly impacts health systems, access to care, and patient outcomes. In war zones, maternal and neonatal health care is often interrupted, resulting in delayed care, shortages of essential supplies, and compromised quality of care. The war between November 2020 and 2022, the siege on Tigray, and the destruction of infrastructure, including health facilities, have caused massive deaths, morbidity, and the displacement of millions of people. This study examines changes in severe maternal outcomes before and during an armed conflict at the Ayder Comprehensive Specialized Hospital in the Tigray region of Ethiopia.

Methods

We conducted a retrospective comparative analysis of pre-war vs. wartime analysis of maternal morbidity and mortality among parturient mothers at Ayder Comprehensive Specialized Hospital, Tigray, Ethiopia. Pregnant women or those who are within 42 days postpartum/any form of pregnancy termination that satisfies the WHO Near-miss criteria were enrolled. We compared outcome of women with potentially life-threatening complications from a pre-war period (N = 691) and wartime (N = 428). EpiData 4.6 was used for data entry, and STATA 16.1 and R 4.2.0 were used for analysis. Categorical variables were described using frequencies and percentages, and differences in proportions were tested using chi-square or Fisher’s exact tests. Key indicators of severe maternal outcomes and patterns of organ dysfunction and critical interventions were examined.

Results

More than half (55%) of the study participants with potentially life-threatening obstetric condition were referred from rural setting during the prewar period as compared to just 9% during the wartime (p<0.001). During the war, the incidences of severe preeclampsia significantly increased from 36.6 to 47.9% (p < 0.001), obstructed labour from 2.0 to 4.4% (p = 0.020), and ICU admissions among mothers with potentially life-threatening complications from 3.5 to 6.1% (p = 0.041). Among women with severe maternal outcomes (i.e., women with maternal near-miss and maternal death), coagulopathy increased significantly (8.8% vs. 63.9%, p < 0.001) and multiple organ dysfunction increased (52.4% vs. 68.3%, p = 0.002). The severe maternal outcome ratio (i.e., maternal near miss + maternal death per 1000 livebirths) increased from 33.2 to 42.7 per 1000 live births (p = 0.013), while the maternal near-miss ratio increased from 28.5 to 39.0 per 1000 live births (p = 0.003). The use of essential interventions, such as oxytocin to prevent postpartum haemorrhage and magnesium sulphate for eclampsia, decreased significantly. Although there was no statistically significant change in the maternal mortality index related to hypertensive disorders (from 37.5% prewar to 61.1% wartime, p = 0.130), mortality associated with haemorrhage (from 12.5% prewar to 61.1% wartime, p = < 0.001) and infectious complications ( from 20.8% prewar, to 55.6% wartime, p = 0.020) worsened.

Conclusion

The armed conflict in Tigray resulted in a marked deterioration in maternal outcomes and reduced uptake of key obstetric interventions in Ayder Comprehensive Specialized Hospital. As can be indirectly inferred from WHO’s severe maternal outcome indicators, delayed access to care, limited resources, and disrupted referral pathways contributed to an increase in severe maternal and maternal near-miss complications. The dramatic decrease in patients from rural areas suggests that many seriously ill women never reached ACSH — likely contributing to an underestimation of the true impact on maternal mortality. Strengthening referral systems, and ensuring supply chains for essential medicines with a policy of preventing and mitigating the impact of conflict on health and specifically on maternal health.

Introduction

According to WHO, in 2020 globally a total of 287,000 women (800 per day) died from preventable causes related to pregnancy and childbirth [1]. In 2017, the number of maternal deaths worldwide was approximately 295,000; about two-thirds of these occurred in sub-Saharan Africa [2]. These figures reflect profound inequalities in maternal health and access to quality healthcare. While the maternal mortality ratio (MMR) in high-income countries is around 11 per 100,000 live births, it reaches 462 per 100,000 in low-income countries, highlighting the wide disparities between rich and poor regions [2].

These inequalities are exacerbated in humanitarian and conflict situations [3], where maternal mortality is estimated to be 1.9 times the global average [4]. Fragile states, often characterized by recurrent instability, weak governance, and underfunded health systems, bear a disproportionate share of the global maternal mortality burden. In 2015, countries classified as fragile states accounted for 61% of global maternal deaths [5]. The Fragile States Index classifies 15 countries, including Ethiopia, as “very high alert” or “high alert” states, with maternal mortality ratios ranging from 31 (Syria) to 1,150 (South Sudan) in 2017 [2]. Although Ethiopia has made remarkable progress in reducing maternal mortality, the MMR was still estimated at 401 per 100,000 live births in 2017 [6], and achievements remain fragile. In 2023, Ethiopia was ranked number 11 with the Fragile State Index score of 100.4, putting the country in the “high alert” category [7].

On November 3, 2020, a civil war broke out in the Tigray region of Ethiopia. The conflict severely disrupted health services, supply chains, and referral networks, which are crucial elements for reducing maternal morbidity and mortality. Vulnerable groups, including pregnant women and girls, were disproportionately affected, and there were numerous reports of sexual violence being used as a weapon of war [8, 9]. In addition, health infrastructure was systematically attacked. Studies reported that more than 80% of health facilities in Tigray were not functional during the war [10, 11].

Ayder Comprehensive Specialized Hospital (ACSH), a large tertiary referral and teaching hospital in Mekelle, Tigray, remained operational despite these challenges. However, it was forced to operate under siege-like conditions and with extremely limited resources [12]. Due to blockades and transportation restrictions, vital medicines, blood supplies and equipment have become scarce [13]. The collapse of financial systems, communication failures and the harassment of external aid workers have further exacerbated the crisis. Under these conditions, previously preventable maternal complications can now often develop into life-threatening situations, increasing the frequency of maternal near misses and the risk of maternal death [14]. Prior to the conflict, ACSH data on maternal near misses and deaths had been documented [15]. These previously published pre-war data provide an important reference point for assessing the extent to which the armed conflict impacted on maternal health and quality of care. Community-based data collected during the war showed a maternal mortality ratio of 840 per 100,000 live births [16], compared with a pre-war estimate of 186 per 100,000, suggesting that many women never reached the referral center.

The aim of this study was to compare maternal health indicators before (July 1, 2018– June 30, 2019) and during the war (July 1, 2021– June 30, 2022) at ACSH, examining changes in the incidence and severity of maternal complications, indicators of SMO and MNM, patterns of organ dysfunction, and the use of critical interventions. By using baseline data from before the war [15], we can demonstrate the extent of deterioration and identify urgent priorities for protecting maternal health in conflict settings.

Methods

Study design and setting

A retrospective comparative study among women who fulfilled WHO Near-miss criteria was conducted at ACSH, a tertiary referral center in Tigray, Ethiopia. The hospital provides comprehensive specialized care across various departments and units, including maternal and neonatal services, for a catchment population of approximately nine million people [17]. Prewar, ACSH used to host referrals from Tigray, neighbouring districts of Amhara and Afar regions. During the wartime, this was limited due to security concerns. Although Aksum Comprehensive Specialized Hospital has recently begun offering residency training, Ayder remained the primary referral and teaching hospital during the war. Throughout the conflict, physicians and nurses worked without salary. In addition, supply chain disruptions, security concerns, and extremely challenging working conditions led to an exodus of healthcare professionals.

Before the war, ACSH registered an annual patient turnover of more than 300,000, and employed 264 specialists and subspecialists. However, due to the conflict, 142 of these specialists were displaced or did not return from further studies. The hospital currently manages approximately 6,000 deliveries per year. According to the Department of Obstetrics and Gynecology’s Human Resource Plan for 2025, the number of obstetricians and gynecologists should have reached 29, but remains at 11.

Study population and periods

Participants in this study were pregnant women or women within 42 days of delivery/any form of pregnancy termination who visited ACSH and met the modified WHO inclusion criteria for near misses (Table 1). We included women with potentially life-threatening conditions (PLTCs), maternal near misses (MNM), and maternal deaths in two time periods: a pre-war period (July 1, 2018 to June 30, 2019) and a wartime period (July 1, 2021 to June 30, 2022). The pre-war sample comprised 691 women [15], and the wartime sample comprised 428 women.

Data sources and variables

The MNM registry was supervised for completeness by MNM focal person at the labor and delivery unit. We trained resident physicians to track and record women with PLTC upon discharge and to triangulate data from different sources such as MNM registries, maternal death reviews, and medical records, including labor ward registers and ICU logbooks. Variables included maternal demographics, obstetric history, severe maternal complications (e.g., severe preeclampsia/eclampsia, postpartum haemorrhage, sepsis, obstructed labor, uterine rupture), organ dysfunction, and critical interventions (use of oxytocin, magnesium sulphate, blood transfusions, laparotomy, ICU admission).

Five potentially life-threatening conditions (PLTC) were used as inclusion criteria for this study: severe pregnancy-related haemorrhage, severe preeclampsia, eclampsia, sepsis/severe systemic infection, and uterine rupture. Those who did not meet any of these inclusion criteria were excluded from the analysis. MNM and SMO were identified according to the WHO criteria [18, 19]. Organ dysfunctions were categorized according to the WHO definitions for near misses. Table 1 was adapted from Nelissen et al. [20].

Table 1 WHO near miss criteria adapted to the L.cal context of ACSH (reproduced from Nelissen et a L. with modification) [20]

The primary outcome measure was the MNM ratio in ACSH during the study period. Secondary outcome indicators such as the MMR, Severe Maternal Outcomes (SMO) ratio and the Mortality Index (MI) were calculated. It should be noted that Maternal near-miss, rather than focusing solely on maternal mortality, is used as a broader indicator of the quality of obstetric care because it allows for a more comprehensive assessment of severe maternal morbidity, which is more common than death, and can reveal more specific insights into preventable factors [18].

Per The WHO generic guide “near-miss approach for maternal health”;-

Severe maternal complications (SMC) i.e. potentially life-threatening conditions (PLTC) are defined as “potentially life-threatening conditions”. For this study, five potentially life-threatening conditions were used as part of the inclusion criteria set: severe pregnancy related hemorrhage, severe pre-eclampsia, eclampsia, sepsis/severe systemic infection, and ruptured uterus.

Severe maternal outcome (SMO) (or Life-threatening conditions (LTC) refers to “a life-threatening condition, i.e. organ dysfunction or death which includes MNM and maternal deaths (SMO = MNM + MD). SMO ratio (SMOR) is the number of MNM plus MD per 1000 live births.”

MNM ratio refers to “the number of maternal near-miss cases per 1000 live births (MNMR = MNM/LB). These two indicators, SMO ratio, and MNM ratio give an estimate of the amount of care and resources that would be needed in an area or facility.”

MNM ratio (MNMR) refers to “the number of maternal near-miss cases per 1000 live births (MNMR = MNM/LB). Similar to the SMOR, this indicator gives an estimation of the amount of care and resources that would be needed in an area or facility.”

Maternal near-miss mortality ratio (MNM: 1 MD) refers to “the ratio between maternal near-miss cases and maternal deaths. Higher ratios indicate better care.”

MI “refers to the number of maternal deaths divided by the sum of women with life-threatening conditions and maternal deaths expressed as a percentage [MI = MD/ (MNM + MD)].” The higher the index the more women with life-threatening conditions die (low quality of care), whereas the lower the index the fewer women with life-threatening conditions die (better quality of care).

Data collection process and data quality

Data were collected using a structured checklist by trained resident physicians. Before actual data collection, a one-day training session was conducted for the data collectors on the purpose of the study, operational definitions, and how to use EpiData for data entry. The collected data were supervised daily for completeness. EpiData version 4.6 was used for data entry. The use of EpiData helped avoid typing errors, coding problems, invalid responses, and inconsistencies. In addition, the use of EpiData eased the process of exporting data to Stata and R statistical software.

Data management and analysis

Statistical analysis was conducted using STATA version 16.1 and R version 4.2.0. Stata was used for descriptive statistics (to generate frequency, percent, measures of central tendency, and measures of dispersion). We used R for the chi-square test of association due to its flexibility in accommodating different datasets simultaneously and the ease of programming matrix functions. We used the Fisher exact test for contingency tables with more than 25% of cells having an expected frequency of less than 5.

Frequency tables and cross-tabulations were produced for the demographic and clinical variables as well as for the underlying causes per the “WHO near-miss approach for maternal health dummy tables.” Continuous variables were also described using an appropriate combination of measure of central tendency and measure of dispersion. The difference in proportions between pre-war and wartime periods was evaluated using the chi-square test or Fisher’s exact test as appropriate. Results are presented in the form of text and table. A p-value < 0.05 was considered statistically significant.

Results

Of the 691 women with PLTC in the pre-war period, 55% were from outside of Mekelle city (i.e. from rural settings) whereas only 9% of the 428 women during the wartime were from rural settings (p < 0.0001). During the war, there was a significant increase in the proportion of women presenting with severe preeclampsia (36.6–47.9%, p < 0.001) and obstructed labor (2.0–4.4%, p = 0.020), while other severe conditions remained relatively stable. Critical interventions shifted, with a notable rise in ICU admissions (3.5–6.1%, p = 0.041), even though blood product usage (p = 0.254) and laparotomy rates (p = 0.242) did not show statistically significant changes. Among women with severe maternal outcomes (SMOs), organ dysfunction became more common (24.6–47.9%, p < 0.001), with a dramatic increase in coagulation/hematologic dysfunction (8.8–63.9%, p < 0.001) and neurologic dysfunction (1.2–10.7%, p < 0.001). Respiratory dysfunction, however, decreased significantly (44.7–12.2%, p < 0.001). Multiple organ dysfunction also rose substantially from 52.4 to 68.3% (p = 0.002). In maternal deaths, while total numbers were small, there were significant increases in coagulation/hematologic dysfunction (4.2–61.1%, p < 0.001) and neurologic dysfunction (4.2–55.6%, p < 0.001), indicating a more complex and severe clinical picture during the war period (Table 2).

Table 2 Morbidity conditions in the audited sample of women with PLTC and SMO, ACSH, prewar (1 July 2018–30 June 2019) vs. wartime (July 1, 2021– June 30, 2022)

During the war, there were significant increases in obstetric haemorrhage across all groups—WWPLTCs (24.0–43.7%), MNMs (55.5–67.4%), and maternal deaths (12.5–61.1%)—and in other obstetric complications among WWPLTCs (6.9–12.4%). Additionally, maternal deaths due to pregnancy-related infections rose markedly (20.8–55.6%), while those due to medical/surgical/mental conditions decreased (75.0–38.9%). Among MNMs, anemia surged from 12.3 to 60.4% during the war, and malaria emerged as a new contributory condition (0–0.9%) (Table 3).

Table 3 Underlying causes of SMC (PLTC) a and SMO (LTC) B, prewar (1 July 2018–30 June 2019) vs. wartime(July 1, 2021– June 30, 2022)

During the war, there was a significant shift in how pregnancies ended among women with potentially life-threatening conditions (PLTCs) and maternal near misses (MNMs), but not in maternal deaths. For PLTCs, the proportion of vaginal deliveries decreased and there was an increase in procedures such as curettage/vacuum aspiration and laparotomies for ectopic pregnancies. Similar patterns of change were seen in MNM, particularly an increase in procedures such as curettage/vacuum aspiration. While the Cesarean section rate remained largely unchanged (34.0% prewar vs. 32.0% wartime) in this patient population, there was a significant increase in preterm births in the PLTCs (from 19.5% before the war to 29.3% during the war), but this was not seen in the MNMs. Among maternal deaths, no significant changes were observed in mode of delivery or perinatal outcomes, with the exception of a decrease in preterm births (Table 4).

Table 4 End of pregnancy and pregnancy outcome, ACSH, prewar (1 July 2018–30 June 2019) vs. wartime (July 1, 2021– June 30, 2022)

During the war, there was a significant increase in the severity of maternal illness, reflected in an increase in the ratio of severe maternal outcome (SMO) to maternal near-miss (MNM). Although the number of maternal deaths remained relatively stable, MNM cases increased significantly, with the SMO ratio increasing from 33.2 to 42.7 per 1000 live births and the MNM ratio increasing from 28.5 to 39.0 per 1000 live births. The maternal mortality index (MI) showed a non-significant decrease from 14.1 to 8.8%. Hospital access patterns changed: fewer SMO cases were treated within 12 h of admission (93.5% before the war versus 70.7% during the war), and among these early presentations, the MI significantly increased (from 3.8 to 11.0%). Intrahospital SMO cases and rates did not change significantly, but MI decreased non-significantly from 18 to 6.1%. Overall, the wartime period was characterized by more severe maternal morbidity and more delayed referral patterns, although the number of maternal deaths remained stable (Table 5).

Table 5 SMO and MNM indicators, ACSH, prewar (1 July 2018–30 June 2019) vs. wartime (July 1, 2021– June 30, 2022)

During the war, the use of preventive measures and treatments for serious obstetric conditions declined significantly. For postpartum hemorrhage (PPH), the administration of oxytocin for prevention fell from 85.2% before the war to 68.0% during the war (p < 0.001), and the use of uterotonics overall fell from 91.8 to 69.5% (p < 0.001). For severe PPH, uterotonic use fell from 100.0 to 53.3% (p < 0.001), while mortality rose sharply from 1.8 to 9.5% (p = 0.003) and severe maternal outcome (SMO) from 50.6 to 91.4% (p < 0.001). Among women with eclampsia, the use of magnesium sulfate decreased significantly (from 97.8 to 84.9%, p = 0.032), and the proportion of women receiving an anticonvulsant decreased from 100.0 to 84.9% (p = 0.006). The use of prophylactic antibiotics in cesarean deliveries remained unchanged (90.6% vs. 90.5%, p = 0.968), but in women with sepsis, parenteral antibiotic administration decreased from 92.3 to 72.1% (p < 0.001), with a concomitant increase in mortality from 3.8 to 14.7% (p = 0.006). In the management of uterine rupture, laparotomy after three hours increased from 9.0 to 47.4% (p = 0.005). While the administration of corticosteroids in preterm labor did not change significantly (from 58.5 to 70.0%, p = 0.095), early neonatal mortality decreased significantly from 32.4 to 18.9% (p = 0.036). Overall, these results suggest that the wartime period was associated with limited access to and use of essential and critical interventions, resulting in increment of severe maternal conditions and higher mortality rates from various obstetric complications (Table 6).

Table 6 Process and outcome indicators related with specific conditions, ACSH, prewar (1 July 2018–30 June 2019) vs. wartime (July 1, 2021– June 30, 2022)

Discussion

The aim of this study was to assess the impact of the Tigray conflict on maternal morbidity and mortality outcomes at ACSH by comparing data before and during the war. During the war, cases of severe pre-eclampsia increased (36.6% vs. 47.9%, p < 0.001), as did cases of obstructed labor (2.0% vs. 4.4%, p = 0.020) and ICU admissions (3.5% vs. 6.1%, p = 0.041). We found that the SMO ratio increased from 33.2 to 42.7 per 1000 live births (p = 0.013) and the MNM ratio increased from 28.5 to 39.0 per 1000 live births (p = 0.003). The severity of complications increased significantly, including an increase in coagulopathy (from 8.8 to 63.9%) (p < 0.001) and multiple organ dysfunction (from 52.4 to 68.3%) (p = 0.002). At the same time, the use of essential and critical interventions decreased, with a sharp decline in uterotonic prophylaxis for postpartum hemorrhage and magnesium sulfate for preelclampsia and eclampsia. Although there was no significant increase in maternal mortality at the facility, the dramatic decrease in patients from rural areas (from 55 to 9%, p < 0.0001) suggests that many seriously ill women never reached ACSH — likely contributing to an underestimation of the true impact on maternal mortality.

The increase in SMO and MNM rates during the conflict is consistent with patterns observed in other war-affected regions such as Afghanistan, Yemen and Syria, where insecurity and disrupted health services have led to higher rates of severe maternal complications and near misses [4, 21]. In these contexts, armed conflict affects referral pathways, the supply of essential medicines, supplies, and skilled attendance at birth. In stable, non-conflict settings, timely access to emergency obstetric care and the availability of life-saving interventions generally limits the progression of severe complications to near misses [22]. The rise in SMO and MNM in our setting indicates overburdened health systems and delayed utilization of care. This highlights the critical need to ensure safe pathways for patient referral, continuous supply chains for essential medicines and adequate staffing during conflicts.

The dramatic increase in severe preeclampsia, obstructed labor, coagulopathy, and multiple organ dysfunction indicates that patients reached ACSH at an advanced stage of disease. There are similar reports in other conflict-affected settings in Ethiopia. For instance, Kodo et al. reported poor pregnancy, delivery, reproductive and fertility-related health outcomes after the outbreak of conflict in the Amhara region [23]. Kebede et al. also reported an increase in maternal near-miss and complications related to hypertensive disorders, obstetric haemorrhage, and uterine rupture in the Amhara region following the conflict, echoing our findings [24]. Similar findings have also been reported from conflict-affected other countries where protracted delays due to insecurity force women to present with advanced complications due to insecurity [25, 26]. In non-conflict contexts, early detection and treatment of conditions such as severe hemorrhage and preeclampsia usually prevents the progression of multiple organ dysfunction [27]. This means that the conflict has prevented timely intervention, leading to a cascade of physiological damage. To mitigate such consequences, protective measures for health facilities, ensuring unimpeded humanitarian corridors and training of frontline staff to stabilize patients prior to referral are critical.

Our results show a significant decrease in the use of uterotonics for postpartum hemorrhage and magnesium sulfate for eclampsia. Similar patterns have been observed in conflict areas such as Yemen and South Sudan, where shortages of essential medicines and disruption of supply routes severely affect clinical management [28, 29]. Even in non-conflict, resource-poor settings, supply shortages and weak supply chains can hinder the use of standard obstetric care protocols [30]. This decline in life-saving interventions during conflict means that health providers lacked the basic tools to prevent or control serious complications. Efforts need to focus on securing supply chains, stockpiling essential medicines and implementing contingency plans for procurement and distribution in times of crisis.

The lack of a significant statistical increase in maternal mortality at ACSH contrasts with community-based data indicating a dramatic increase in maternal mortality during the war [16]. A similar discrepancy has been found in other conflict areas, where referral hospitals receive fewer patients from remote and highly insecure areas [31]. The patients who do reach a facility may represent a subgroup with fewer delays or better initial health status. In stable, non-conflict areas, facility mortality trends are more in line with community outcomes due to reliable referral systems [32]. This means that stable in-hospital mortality rates mask the true increase in maternal mortality in populations without access to care. Policy interventions must therefore prioritize restoring safe and functioning referral networks and empowering rural populations to access timely care.

Before the war, more than half of the ACSH’s patients came from rural areas, compared to only 9% during the conflict. This drastic decline, also reported in other conflict-affected areas such as parts of the Democratic Republic of Congo and Somalia, underscores that insecurity and the destruction of infrastructure disproportionately prevent rural populations from accessing medical care [21, 33]. Even in non-conflict areas, rural populations face obstacles, but these are usually related to distance and transportation costs rather than active hostilities [34]. This means that conflict is an additional, formidable obstacle to existing challenges that is likely to result in unrecorded deaths at the community level. Restoring transportation security, communication networks and community-based interventions is therefore critical to restoring equitable access to health care.

Similar to the pre-war published data [15], we utilised the “modified WHO near-miss criteria” developed by Nelissen et al. [20] rather than the original criteria. This aspect should be considered when interpreting the results of this study. Furthermore, certain parameters mentioned in the original WHO near-miss criteria, such as pH, serum lactate levels, and PaO2/FiO2, were not utilised due to limitations in the setup.

Conclusion

The armed conflict in Tigray has severely affected maternal health services in ACSH, as evidenced by increased SMO and MNM ratios, more advanced organ dysfunction at admission, and a decline in essential interventions. While maternal mortality did not increase significantly at ACSH, the significantly lower referral of patients from rural areas suggests that the actual burden of maternal mortality was far greater in the region. These findings mirror patterns from other conflict areas and underscore the urgency of protecting health infrastructure, ensuring uninterrupted care, restoring referral pathways, and implementing community-level strategies to ensure maternal health during and after armed conflict. Moreover, we recommend establishing emergency supply chain mechanisms to ensure uninterrupted provision of essential medications such as oxytocin and magnesium sulfate, deploying mobile maternal health units to bridge gaps in access within conflict zones, and strengthening regional and international partnerships to reinforce health infrastructure resilience in war-affected areas. We also emphasize on the importance of post-war health system recovery, particularly through the restoration of referral pathways, enhancement of maternal health surveillance, and attention to the long-term effects of war on reproductive health. Last but not least, we strongly recommend for nations to adopt conflict prevention and mitigation policies and procedures.