Survival outcomes for first-line antiretroviral therapy in India’s ART program

Rakhi Dandona, Bharat B. Rewari, G. Anil Kumar, Sukarma Tanwar, S. G. Prem Kumar, Venkata S. Vishnumolakala, Herbert C. Duber, Emmanuela Gakidou, Lalit Dandona

Little is known about survival outcomes of HIV patients on first-line antiretroviral therapy (ART) on a large-scale in India, or facility level factors that influence patient survival to guide further improvements in the ART program in India. We examined factors at the facility level in addition to patient factors that influence survival of adult HIV patients on ART in the publicly-funded ART program in a high- and a low-HIV prevalence state. Retrospective chart review in public sector ART facilities in the combined states of Andhra Pradesh and Telangana (APT) before these were split in 2014 and in Rajasthan (RAJ), the high- and a low-HIV prevalence states, respectively. Records of adults initiating ART between 2007-12 and 2008-13 in APT and RAJ, respectively, were reviewed and facility-level information collected at all ART centres and a sample of link ART centres. Survival probability was estimated using Kaplan-Meier method, and determinants of mortality explored with facility and patient-level factors using Cox proportional hazard model. Based on data from 6581 patients, the survival probability of ART at 60 months was 76.3 % (95 % CI 73.0-79.2) in APT and 78.3 % (74.4-81.7) in RAJ. The facilities with cumulative ART patient load above the state average had lower mortality in APT (Hazard ratio [HR] 0.74, 0.57-0.95) but higher in RAJ (HR 1.37, 1.01-1.87). Facilities with higher proportion of lost to follow-up patients in APT had higher mortality (HR 1.47, 1.06-2.05), as did those with higher ART to pre-ART patient ratio in RAJ (HR 1.62, 1.14-2.29). In both states, there was higher hazard for mortality in patients with CD4 count 100 cells/mm(3) or less at ART initiation, males, and in patients with TB co-infection. These data from the majority of facilities in a high- and a low-HIV burden state of India over 5 years reveal reasonable and similar survival outcomes in the two states. The facilities with higher ART load in the longer established ART program in APT had better survival, but facilities with a higher ART load and a higher ratio of ART to pre-ART patients in the less experienced ART program in RAJ had poorer survival. These findings have important implications for India's ART program planning as it expands further.