BACKGROUND: Concerns have been raised that marginalised populations may not achieve adequate compliance to antiretroviral therapy. Our objective was to describe the long-term virological, immunological and mortality outcomes of providing highly active antiretroviral therapy (HAART) with strong adherence support to HIV-infected female sex workers (FSWs) in Burkina Faso and contrast outcomes with those obtained in a cohort of regular HIV-infected women. METHODS: Prospective study of FSWs and non-FSWs initiated on HAART between August 2004 and October 2007. Patients were followed monthly for drug adherence (interview and pill count), and at 6-monthly intervals for monitoring CD4 counts and HIV-1 plasma viral loads (PVLs) and clinical events. RESULTS: 95 women, including 47 FSWs, were followed for a median of 32 months (interquartile range [IQR], 20-41). At HAART initiation, the median CD4 count was 147 cells/μl (IQR, 79-183) and 144 cells/μl (100-197), and the mean PVLs were 4.94 log10 copies/ml (95% confidence interval [CI], 4.70-5.18) and 5.15 log10 copies/ml (4.97-5.33), in FSWs and non-FSWs, respectively. Four FSWs died during follow-up (mortality rate: 1.7 per 100 person-years) and none among other women. At 36 months, the median CD4 count increase was 230 cells/μl (IQR, 90-400) in FSWs vs. 284 cells/μl (193-420) in non-FSWs; PVL was undetectable in 81.8% (95% CI, 59.7-94.8) of FSWs vs. 100% (83.9-100) of non-FSWs; and high adherence to HAART (> 95% pills taken) was reported by 83.3% (95% CI, 67.2-93.6), 92.1% (95% CI, 78.6-98.3), and 100% (95% CI, 54.1-100) of FSWs at 6, 12, and 36 months after HAART initiation, respectively, with no statistical difference compared to the pattern observed among non-FSWs. CONCLUSIONS: Clinical and biological benefits of HAART can be maintained over the long-term among FSWs in Africa and could also lead to important public health benefits.