Clinical and immunological outcomes according to adherence to first-line HAART in a urban and rural cohort of HIV-infected patients in Burkina Faso, West Africa
Emanuele Focà, Silvia Odolini, Giorgia Sulis, Stefano Calza, Virginio Pietra, Paola Rodari, Pier Francesco Giorgetti, Alice Noris, Paul Ouedraogo, Jacques Simpore, Salvatore Pignatelli, Francesco Castelli
BackgroundAim of our study is to investigate the clinical and immunological outcomes according to first-line HAART adherence in a large cohort of HIV-infected patients in Burkina Faso.MethodsA retrospective study was conducted between 2001 and 2009 among patients from two urban medical centers [St. Camille Medical Center (CMSC) and “Pietro Annigoni” Biomolecular Research Center (CERBA)] and 1 in the rural District of Nanoro (St. Camille District Hospital). Socio-demographical and clinical data were analyzed. Adherence was evaluated through a questionnaire investigating 5 key points related to drugs, consultations and blood exams, by assigning 0 to 2 points each up to 10 points overall. Data were collected at baseline and regularly thereafter. Adherence score was considered as a continuous variable and classified in optimal (8–10 points) and sub-optimal (0–7 points). Immunological outcome was evaluated as modification in CD4+ T-cell count over time, while predictors of death were explored by a univariate and multivariate Cox model considering adherence score as a time-varying covariate.ResultsA total of 625 patients were included: 455 (72.8%) were females, the median age was 33.3 (IQR 10.2) years, 204 (32.6.%) were illiterates, the median CD4+ T-cell count was 149 (IQR 114) cells/μl at baseline. At the end of the observation period we recorded 60/625 deaths and 40 lost to follow-up. The analysis of immunological outcomes showed a significant variation in CD4+ T-cell count between M12 and M24 only for patients with optimal adherence (Δ=78.2, p<0.001), with a significant Δ between the two adherence groups at M24 (8–10 vs 0–7, Δ=53.8, p=0.004). Survival multivariate analysis revealed that covariates significantly related to death included being followed at CERBA (urban area) or Nanoro (rural area), and receiving a regimen not including fixed dose combinations, (p=0.024, p=0.001 and p<0.001 respectively); conversely, an increasing adherence score as well as an optimal ad