Burden of diabetic ketoacidosis and its predictors among diabetic patients in Ethiopia: Systematic review and meta-analysis

Background: Multiple studies across Ethiopia have investigated the occurrence of DKA, showing significant variations and conflicting findings. This systematic review and meta-analysis seek to consolidate the overall prevalence of diabetic ketoacidosis and its associated factors in the Ethiopian context. Methods: The study adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA-P) guidelines. Data was collected from PubMed/MEDLINE, Science direct, Google Scholar, and gray literature sources. Microsoft Excel was used for data extraction and summary, while the analysis was performed with R software version 4.3.2. The overall pooled prevalence of diabetic ketoacidosis and its components was estimated using a random effects model. Publication bias was assessed both graphically, using funnel plots, and statistically, with tests such as Egger’s regression test. Subgroup analysis were carried out to minimize random variations in the estimates from the primary studies. Result: The pooled estimated prevalence of diabetic ketoacidosis among diabetic patients in Ethiopia was 46% (95% CI; 36, 57; I2 = 100%, P≤0.001). Medication discontinuations (AOR = 1.30, 95 CI 1.20, 1.64), presence of comorbidity (AOR = 1.53, 95 CI 1.10, 2.20) and presence of infection (AOR = 1.62, 95 CI 1.31, 1.98) had an association with diabetic ketoacidosis among diabetic patients. Conclusions: Medication discontinuations, comorbidity, and infection are individual contributors to diabetic ketoacidosis in diabetic patients. Implementing initiatives to enhance medication adherence and establish comprehensive diabetes management programs covering glycemic control, comorbidities, and infection management can effectively address these factors.