Author(s): Tadesse K, Haile F, Hiruy N
Abstract Share this page
Abstract BACKGROUND: Since launching of antiretroviral (ART) treatment, the numbers of patients enrolled in to ART are increasing in many developing countries. But many studies done across Africa including Ethiopia on antiretroviral therapy programs have shown higher mortality at the first six months of treatment initiation. But the factors associated with this high mortality are poorly characterized. So this study aims to determine mortality and identify predictors of it among patients on ART. METHODS: Retrospective cohort study was employed among a total of 520 records of patients who were enrolled on antiretroviral therapy in Aksum hospital from September 2006 to August 2011. Baseline patient records were extracted from electronic and paper based medical records database and analysed using Kaplan Meier survival and Cox proportional hazard model to identify the independent predictors of mortality of patients on ART. RESULTS: A total of 46 (8.85\%) deaths was observed giving an overall mortality rate of 3.2 per 100 person-years. The independent predictor of mortality identified for this cohort were haemoglobin level <11 mg/dl (Hazard Ratio (HR) = 1.9, 95\%-CI = 1.01, 3.52), CD4 cell counts lower than 50 cells/µl (HR = 2.1, 95\%- CI = 1.13,3.89), Male gender (HR = 1.9, 95\%-CI = 1.01,3.52), Weight <40 kg (HR = 2.3,95\% CI = 1.24,4.55), primary level of education and lower (HR = 2.6, 95\%- CI = 1.29,5.55). CONCLUSIONS: The over all mortality of adults patients on ART was low but higher in the early months of ART initiation. low levels of haemoglobin <11 gm/dl, lower CD4 cell count, male gender, weight <40 Kg and individuals who have primary level of education and lower were identified as the independent predictors of mortality. For this reason, early initiation of ART despite the CD4 count and method of HIV diagnosis, nutritional support and close monitoring of patients in the early periods of ART treatment initiation is very crucial to improve patient survival.
This article was published in PLoS One
and referenced in Journal of Proteomics & Bioinformatics