• high dependency unit.
• low-resource settings
• MEOWS
• obstetric

Maternal mortality is still a challenge for health systems, with 287 000 women died during and following pregnancy and childbirth in 2020. Almost 95% of all maternal deaths occurred in low-income countries, and most could have been prevented(1). Timely identification of obstetric complications with appropriate management can reduce the burden of maternal mortality. For this reason, the 2003–2005 CEMACH report recommended the routine use of the modified early obstetric warning system (MEOWS), which is an early warning system adapted for the obstetric population(2). In this analysis we seek to determine whether the MEOWS, adapted from the CEMACH 2007 guidelines and externally validated in 2012(3), is effective in predicting maternal mortality and severe maternal morbidity in critically ill patients admitted to obstetric High Dependency Unit (HDU) in an urban low-resource setting. We performed an analysis of prospectively collected registry data from January 2022 to May 2024, including any pregnant or postpartum woman who is admitted, in that period, in the high-dependency unit (HDU) of Princess Christian Maternity Hospital PCMH, the largest maternity hospital in Freetown, Sierra Leone, counting about 9000 admissions and 6500 deliveries per year(4). MOEWS is calculated at admission to the HDU and at 24 h after admission. The primary aim was to evaluate the performance of admission MEOWS as a predictive tool of mortality and severe maternal morbidity, according to the definition by WHO (ref.). The secondary aim was to evaluate whether the 24h MEOWS is superior to baseline MEOWS in predicting same outcome. A total of 833 patients were included in the analysis, with a median age of 25 years, and the most common causes of admission to HDU were eclampsia (34%), followed by peripartum hemorrhage (18.4%) and puerperal sepsis (10.6%), resulting in a crude mortality rate of 12.7%. The results shown that Admission MEOWS does not appear to be an accurate tool for predicting maternal mortality (AUROC 0.71). Instead, the 24h MEOWS demonstrated better performance (AUROC 0.86). Mortality showed a linear increase corresponding to higher MEOWS levels, observed in both Admission MEOWS (MEOWS 6-8 mortality rate 8.2% - MEOWS >8 mortality rate 10.3%) and 24h MEOWS (MEOWS 6-8 mortality rate 18.5% - MEOWS >8 mortality rate 42.1%). About the ability to predict SMM, both MEOWS on admission and at 24h shown a poor performance. Despite these, all five selected criteria to define SMM are more prevalent in the subgroups of patients with high MEOWS (score ≥6). In conclusion, within the limitations of a retrospective study, MEOWS does not appear to be a useful tool for predicting maternal mortality and severe maternal morbidity, in the cohort of patients examined in our setting. Further prospective studies will be needed to confirm these findings.