Risk factors for development of acute kidney injury in hospitalised adults in Zimbabwe

BACKGROUND: Acute kidney injury (AKI) is predominantly a disease of low and middle-income countries. Despite this, there is a particular paucity of data regarding AKI in Africa. Most published studies were conducted prior to the most recent Kidney Disease: Improving Global Outcomes (KDIGO) definition of AKI. This prospective, observational, cohort study examines AKI amongst newly admitted acute medical inpatients in a large, urban, tertiary hospital in Harare, Zimbabwe. METHODS: All newly admitted, adult, medical patients in separate, randomly selected, 24-hour periods were included. Baseline demographic information, comorbidities, nephrotoxic medication use, and reason for admission were recorded on a standardised data capture record. A serum creatinine measurement was performed on all patients at the time of admission and again after 48 hours. Estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation and AKI was defined using the most recent KDIGO definition as an increase in the serum creatinine of greater than 26.5μmol/L within 48 hours, with admission creatinine used as a baseline measurement. RESULTS: 253 patients were included in the analysis; 137 patients (54.2%) were female; 100 patients (39.5%) had HIV infection. 36 patients (14.2%) met the KDIGO criteria for AKI during the 48-hour follow-up period. AKI was more common among males (19.8% vs 9.5%; p = 0.019). The AKI group had a higher serum creatinine at presentation than those without AKI (296.5μmol/L vs 91.0μmol/L; p<0.001) and at 48 hours (447.7μmol/L vs 77.1μmol/L; p<0.001). In logistic regression, AKI was related negatively to female sex (OR 0.461, 95% CI 0.211, 1.003; p = 0.051) and positively predicted by the presence of comorbid hypertension (OR 3.292, 95% CI 1.52, 7.128; p = 0.003) and chronic kidney disease (OR 6.034, 95% 1.792, 20.313; p = 0.004). CONCLUSIONS: KDIGO-defined AKI was common in hospitalised patients in Sub-Saharan Africa and was predicted by male sex, a history of comorbid hypertension and a history of comorbid chronic kidney disease.