Abstract
Acute kidney injury (AKI) is prevalent in Intensive Care Unit settings, with rates exceeding 50%. While many studies from sub-Saharan Africa focus on critically ill AKI patients, limited data exist on non-critically ill patients, hindering effective dialysis prioritization. Studies from developed countries suggest AKI is also common in non-critical settings. This study aimed to assess mortality rates among critically ill and non-critically ill hospitalized AKI patients and identify early mortality predictors at the time of AKI diagnosis. A single-center prospective cohort study was conducted at Kilimanjaro Christian Medical Center between September 2023 and February 2024. Patients admitted to the internal medicine ward were assessed, with critical illness determined using the Universal Vital Assessment (UVA) score. Cox regression identified predictors of in-hospital mortality, and Kaplan-Meier curves assessed survival time. Out of 1,211 admissions, 139 patients met inclusion criteria. Overall hospital mortality was 39.6%, higher in critically ill patients (57.1% vs. 21.7%, p < 0.001). Predictors of mortality included critical illness (aHR 3.44, p < 0.001), traditional herbal medicine (THM) intoxication (aHR 5.99, p = 0.002), volume depletion (aHR 1.95, p = 0.028), referral from regional hospitals (aHR 2.78, p = 0.002), and age >60 (aHR 2.46, p = 0.001). Critically ill patients had shorter median survival (12 vs. 20 days; p = 0.001), which declined with higher UVA risk. While critical illness predicts AKI in-hospital mortality, non-critical AKI patients-often affected by THM, volume depletion, or regional hospital referrals are also at risk. Older age (>60 years) is a non-modifiable predictor of in-hospital mortality in AKI.