Sustainability of first-line antiretroviral regimens: findings from a large HIV treatment program in western Kenya.

OBJECTIVE: To describe first change or discontinuation in combination antiretroviral treatment (cART) among previously treatment naive, HIV-infected adults in a resource-constrained setting.
METHODS: The United States Agency for International Development-Academic Model Providing Access to Healthcare Partnership has enrolled >90,000 HIV-infected patients at 18 clinics throughout western Kenya. Patients in this analysis were aged > or =18 years, previously antiretroviral treatment naive, and initiated to cART between January 2006 and November 2007, with at least 1 follow-up visit. A treatment change or discontinuation was defined as change of regimen including single drug substitutions or a complete halting of cART.
RESULTS: There were 14,162 patients eligible for analysis and 10,313 person-years of follow-up, of whom 1376 changed or stopped their cART. Among these, 859 (62%) changed their regimen (including 514 patients who had a single drug substitution) and 517 (38%) completely discontinued cART. The overall incidence rate (IR) of cART changes or stops per 100 person-years was 13.3 [95% confidence interval (CI): 12.7-14.1]. The incidence was much higher in the first year of post-cART initiation (IR: 25.0, 95% CI: 23.6-26.3) compared with the second year (IR: 2.4, 95% CI: 2.0-2.8). The most commonly cited reason was toxicity (46%). In multivariate regression, individuals were more likely to discontinue cART if they were World Health Organization stage III/IV [adjusted hazard ratio (AHR): 1.37, 95% CI: 1.11-1.69] or were receiving a zidovudine-containing regimen (AHR: 4.44, 95% CI: 3.35-5.88). Individuals were more likely to change their regimen if they were aged > or =38 years (AHR: 1.44, 95% CI: 1.23-1.69), had to travel more than 1 hour to clinic (AHR: 1.34, 95% CI: 1.15-1.57), had a CD4 at cART initiation < or =111 cells/mm3 (AHR: 1.51, 95% CI: 1.29-1.77), or had been receiving a zidovudine-containing regimen (AHR: 3.73, 95% CI: 2.81-4.95). Those attending urban clinics and those receiving stavudine-containing regimens were less likely to experience either a discontinuation or a change of their cART.
CONCLUSIONS: These data suggest a moderate incidence of cART changes and discontinuations among this large population of adults in western Kenya. Mostly occurring within 12 months of cART initiation, and primarily due to toxicity, older individuals, those with more advanced disease, and those using zidovudine are at higher risk of experiencing a change or a discontinuation in their cART.