BACKGROUND: Long-term side-effects and cost of HIV treatment motivate the development of simplified maintenance. Monotherapy with ritonavir-boosted lopinavir (LPV/r-MT) is the most widely studied strategy. However, efficacy of LPV/r-MT in compartments remains to be shown. METHODS: Randomized controlled open-label trial comparing LPV/r-MT with continued treatment for 48 weeks in treated patients with fully suppressed viral load. The primary endpoint was treatment failure in the central nervous system [cerebrospinal fluid (CSF)] and/or genital tract. Treatment failure in blood was defined as two consecutive HIV RNA levels more than 400 copies/ml. RESULTS: The trial was prematurely stopped when six patients on monotherapy (none in continued treatment-arm) demonstrated a viral failure in blood. At study termination, 60 patients were included, 29 randomized to monotherapy and 13 additional patients switched from continued treatment to monotherapy after 48 weeks. All failures occurred in patients with a nadir CD4 cell count below 200/microl and within the first 24 weeks of monotherapy. Among failing patients, all five patients with a lumbar puncture had an elevated HIV RNA load in CSF and four of six had neurological symptoms. Viral load was fully resuppressed in all failing patients after resumption of the original combination therapy. No drug resistant virus was found. The only predictor of failure was low nadir CD4 cell count (P < 0.02). CONCLUSION: Maintenance of HIV therapy with LPV/r alone should not be recommended as a standard strategy; particularly not in patients with a CD4 cell count nadir less than 200/microl. Further studies are warranted to elucidate the role of the central nervous system compartment in monotherapy-failure.