A survival analysis of prognostic determinant factors of time-to-death of HIV/TB co-infected patients under HAART followed-up in a public hospital in Ethiopia

Abstract

Introduction: Human immunodeficiency virus (HIV)/acquired immunodeficiency syndrome (AIDS) and its’ damage are prevailing at a shocking level in the world, and tuberculosis (TB) also adds to this damage, which make things “Mumps on the Goiter”. In this case, highly active antiretroviral treatment (HAART) plays a great role in reducing the damage, and it is a lifetime treatment to reduce HIV-related mortality and morbidity, and prolong patients’ survival time. Material and methods: A retrospective survival study was conducted among 407 HIV-positive TB co-infected patients under HAART to observe the effects of HAART treatment and other covariates for the improvement of patient’s life expectancy. Appropriate survival model was selected using AIC, BIC, and log-likelihood values. Results: Out of the total 407 patients, 120 (29.48%) experienced the event of interest. A majority (n = 74, 61.67%) of those who died of HIV/TB co-infection were males, 108 (90%) had pulmonary TB, and 12 (10%) patients suffered from extra-pulmonary TB. For the log-normal AFT model, marital status, WHO clinical stages, functional status, antiretroviral treatment (ART) regimen, religion, sqrt CD4+, and baseline CD4+ were among significant predictors at a 5% level of significance for the change in patient's lifetime. Conclusions: From this study, AFT models presented a better fit compared with Cox regression model. Among AFT models, the log-normal AFT model was selected, and hence, the study showed that prognostic factors, such as WHO clinical stages, functional status, sqrt CD4+ counts, ART regimen, marital status, baseline CD4+ counts, and their interactions with time, were among the significant predictors for the selected model at 5% significance level.