DOI: https://doi.org/10.36347/sjams.2026.v14i05.039

Pages: 878-884

Introduction: Hemorrhagic shock carries high case-fatality rates in trauma and surgical populations, yet data on factors associated with mortality in North African intensive care settings remain limited. This study aimed to explore which clinical, biological, and therapeutic variables were associated with in-hospital mortality in a mixed-etiology hemorrhagic shock cohort admitted to a Moroccan surgical ICU. Methods: We performed a retrospective analysis of ICU admissions at Ibn Rochd University Hospital Center, Casablanca, from January 2022 to December 2024. Among 223 screened patients, 66 fulfilled the inclusion criterion of systolic blood pressure < 90 mmHg in the setting of polytrauma, penetrating trauma, or gastrointestinal hemorrhage. Twelve candidate variables were tested in univariate analysis; eight with p < 0.20 were considered for multivariate logistic regression. To avoid overfitting in the context of 21 events, a parsimonious three-variable model was constructed based on clinical relevance and collinearity assessment. Results: The cohort was predominantly young (67% aged < 40 years) and male (sex ratio 4:1). Polytrauma accounted for 56% of cases. Twenty-one patients died (32%), with 60% of deaths occurring within the first 24 hours. In multivariate analysis, three variables were independently associated with in-hospital mortality: Glasgow Coma Scale < 8 (OR 3.42; 95% CI 1.28-9.11; p = 0.014), Shock Index > 0.9 (OR 2.87; 95% CI 1.10-7.46; p = 0.031), and hemoglobin < 5 g/dL (OR 3.95; 95% CI 1.36-11.42; p = 0.011). The mean time to hemorrhage control was longer among non-survivors (112 ± 35 vs. 72 ± 25 minutes; p < 0.001). These findings should be considered exploratory given the limited sample size. Conclusion: In this mixed-etiology surgical ICU cohort, the Shock Index, Glasgow Coma Scale, and admission hemoglobin were independently associated with mortality and may serve as early bedside stratification tools. These findings should be interpreted with ca