Predictors of In-Hospital Mortality among Hospitalized Adults with Diabetes Mellitus in Addis Ababa, Ethiopia: A Multicenter Prospective Cohort Study

Background Diabetes mellitus (DM) significantly increases the risk of adverse hospital outcomes; however, data on mortality predictors in sub-Saharan African hospital settings remain scarce. This study aimed to identify independent predictors of in-hospital mortality among DM patients admitted to three Ethiopian tertiary hospitals. Methods A multicenter prospective cohort study was conducted from July 2023 to February 2024, enrolling 398 DM adult patients. Baseline demographic, clinical and laboratory data were collected. Patients were followed from admission until discharge or death. Kaplan-Meier survival analysis with the log-rank test estimated median survival times (MST). Multivariable Cox proportional hazards regression was used to identify independent predictors of mortality, with statistical significance set at p < 0.05. Results Among 398 patients (55% male, median age 60 years [IQR: 50–70]), 82.9% had Type 2 DM and 93.7% had a known history of diabetes. Hypertension (61.3%) and heart failure (37.7%) were the most prevalent comorbidities. The overall in-hospital mortality rate was 7.5% (n = 30). Kaplan-Meier analysis revealed significantly shorter MST for patients with > 3 comorbidities (p = 0.002), acute hemorrhagic stroke (MST: 23 vs. 55 days, p = 0.002), sepsis (MST: 35 vs. 90 days, p = 0.001), and hospital-acquired urinary tract infection (UTI) (MST: 25 vs. 55 days, p = 0.015). After adjustment in the multivariable Cox model, five factors remained independently associated with increased mortality: >3 comorbidities (AHR = 5.50; 95% CI: 1.96–15.40; p = 0.001), community-acquired sepsis (AHR = 7.98; 95% CI: 2.44–26.14; p = 0.001), acute hemorrhagic stroke (AHR = 4.54; 95% CI: 1.07–19.24; p = 0.040), hospital-acquired UTI (AHR = 6.87; 95% CI: 1.63–29.02; p = 0.009), and failure to achieve target blood glucose levels (AHR = 3.63; 95% CI: 1.12–11.76; p = 0.032). Conclusion In-hospital mortality in Ethiopian diabetic patients is profoundly driven by acute severe infections, cerebrovascular events, high comorbidity burden, and poor glycemic control. Early risk stratification, aggressive infection management, and strict glycemic monitoring are essential to reduce the high mortality rate in this vulnerable population.