OBJECTIVES: To limit exposure to anti-HIV drugs and minimize risk of long-term side effects, studies have looked at the possibility of simplified maintenance strategies. Ritonavir-boosted protease-inhibitor (PI)-monotherapies are an attractive alternative, but limited compartmental penetration of PI remains a concern. DESIGN: Non-comparative 24-week pilot study. METHOD: Ritonavir-boosted atazanavir (ATV/r) monotherapy administered to fully suppressed patients (>3 month HIV RNA < 50 copies/ml). Plasma was obtained every 4 weeks and cerebrospinal fluid (CSF) and semen at W24. RESULTS: Two patients (7%) failed ATV/r monotherapy. One patient was subsequently identified as a protocol violator since he had a previous history of treatment failure under indinavir. The second patient deliberately decided to stop treatment after W20. Excluding failing patients, individual measurements of HIV RNA in patients having occasional viral 'blips' was found in five patients. At W24, 3/20 patients had elevated viral loads in CSF (HIV RNA > 100 copies/ml), and 2/15 in semen, despite viral suppression in plasma (< 50 copies/ml). Samples with elevated HIV RNA (> 500 copies/ml) in CSF were all wild type. The mean ATV drug concentration ratio (CSF/blood, n = 22) was 0.9%. Indicators of altered immune activation (CD8CD38 C-reactive protein) remained unchanged. CONCLUSION: This study supports previous results indicating the potential use of PI-based mono-maintenance therapies. However, our results in CSF cautions against the uncontrolled use of PI-based monotherapies.