HIV Status Disclosure and Retention in Care in HIV-Infected Adolescents on Antiretroviral Therapy (ART) in West Africa

Objective: We assessed the effect of HIV status disclosure on retention in care from initiation of antiretroviral therapy (ART) among HIV-infected children aged 10 years or more in Cote d'Ivoire, Mali and Senegal. Methods: Multi-centre cohort study within five paediatric clinics participating in the IeDEA West Africa collaboration. HIV-infected patients were included in this study if they met the following inclusion criteria: aged 10-21 years while on ART; having initiated ART >= 200 days before the closure date of the clinic database; followed >= 15 days from ART initiation in clinics with >= 10 adolescents enrolled. Routine follow-up data were merged with those collected through a standardized ad hoc questionnaire on awareness of HIV status. Probability of retention (no death or loss-to-follow-up) was estimated with Kaplan-Meier method. Cox proportional hazard model with date of ART initiation as origin and a delayed entry at date of 10th birthday was used to identify factors associated with death or loss-to-follow-up. Results: 650 adolescents were available for this analysis. Characteristics at ART initiation were: median age of 10.4 years; median CD4 count of 224 cells/mm(3) (47% with severe immunosuppression), 48% CDC stage C/WHO stage 3/4. The median follow-up on ART after the age of 10 was 23.3 months; 187 adolescents (28.8%) knew their HIV status. The overall probability of retention at 36 months after ART initiation was 74.6% (95% confidence interval [CI]: 70.5-79.0) and was higher for those disclosed compared to those not: adjusted hazard ratio for the risk of being death or loss-to-follow-up = 0.23 (95% CI: 0.13-0.39). Conclusion: About 2/3 of HIV-infected adolescents on ART were not aware of their HIV status in these ART clinics in West Africa but disclosed HIV status improved retention in care. The disclosure process should be thus systematically encouraged and organized in adolescent populations.