Introduction: Anti-retroviral therapy (ART) adverse reactions are among the most common reasons for poor adherence to treatment. There is little information about predictors of antiretroviral therapy adverse reaction and its association with weight change. This study aimed to determine time to antiretroviral therapy adverse reaction and mean change in weight among adult HIV/AIDS patients on antiretroviral therapy at public hospitals in West Hararghe zone, Eastern Ethiopia.

Methods: Institutional based retrospective follow up study was conducted on 558 HIV/AIDS patients on ART from September 2013 to January 2019 at public hospitals in West Hararghe zone, Eastern Ethiopia. Data were collected using checklist and document review. Data were entered using Epi-info and analyzed in R software. Cox proportional hazard model with linear mixed effect model was jointly modeled with current and slope parameterizations. Model adequacy was checked using cox Snell residuals and marginal residuals. Adjusted hazard ratio with its confidence interval was used to show presence and strength of association at 95% confidence level.

Result: Overall, 51 patients developed ART adverse reaction with incidence density of 3.32/100 person-year of observations (95% CI: 2.53, 4.38). Both current value of weight and rate of weight change were not significantly associated with time to ART adverse reaction. Duration on therapy (adjusted hazard ratio [AHR=0.96, 95%CI, 0.93, 0.99]), fair baseline adherence [AHR=4.01, 95%CI 1.26,12.72], poor adherence [AHR=4.23, 95% CI 1.60, 11.17], moderate body mass index at baseline [AHR=2.81, 95%CI 1.07, 7.37], world health organization (WHO) stage II [AHR=14.54, 95%CI 1.78,118.75] and initial ART regimen (TDF, 3TC, EFZ) [AHR=0.22, 95%CI 0.08,0.59]  were significant predictors of time to ART adverse reaction.

Conclusion: Incidence of antiretroviral therapy adverse reaction was relatively low with early onset. Duration on therapy, baseline adherence, HIV/AIDS stage, initial regimen and body mass index were significant predictors of time to ART adverse reaction. Weight change was not significant biomarker of ART adverse reaction. Continuous counseling for clients in clinical stage II and above and patients with poor adherence need to get close follow up to prevent the associated ADRs.

Keywords: Antiretroviral therapy adverse reaction, Adults, Ethiopia, HIV, Joint modeling