A 84-month follow up of adherence to HAART in a cohort of adult Senegalese patients

OBJECTIVEs To assess long-term adherence of the first HIV-1 patients receiving highly active antiretroviral therapy (HAART) in Senegal, and to identify the main determinants of adherence. METHODs The first 180 patients enrolled in the Senegalese HAART initiative between August 1998 and April 2001 followed up for at least 30 days were eligible. Adherence was assessed monthly at each drug dispensation between November 1999 and November 2006 by a pharmacist using a pill count completed by a questionnaire. Adherence was expressed as the proportion of tablets taken to prescribed tablets. An adherence of 95% was considered to be good. A random-intercept logit model was fitted to identify the main determinants of adherence. RESULTs Adherence data were available for 158 of 167 eligible patients. Twenty-nine patients died during the study period and 10 were lost to follow-up. Median treatment duration was 78 months, accruing to 6657 person-months of observation. Overall, mean adherence reached 91% [median: 100%, interquartile range (IQR) 96-100%] and adherence exceeded 95% in 78% [95% Cl 77-79%] of observations. After 4 years of treatment mean adherence stabilized around 90 % and adherence >= 95 % stabilized around 70%. Treatment duration and protease inhibitor (PI)-based regimen (indinavir) had a negative effect on adherence, but adherence tended to improve with time for patients receiving a PI. Patient-level variance was highly significant and accounted for a third of total variance. CONCLUSIONs This work demonstrates that good long-term adherence can be achieved in the sub-Saharan context given close monitoring and adherence support measures, confirms the worse adherence for indinavir and underlines the importance of patient heterogeneity.