OBJECTIVE To determine intracellular concentrations of indinavir (IDV) and investigate the relationship between plasma and intracellular IDV pharmacokinetics in HIV-infected patients. METHODS A pharmacokinetic study of 10 patients receiving IDV plus dual nucleoside analogue therapy. Peripheral blood mononuclear cells were isolated by density gradient centrifugation and cell counts estimated. IDV was extracted from cells in the presence of 60% methanol and evaporated to dryness. Both plasma and intracellular IDV samples were assayed by high performance liquid chromatography linked to mass spectrometry. Data were subjected to non-compartmental pharmacokinetic analysis. RESULTS The mean intracellular IDV area under the curve over 8 h (AUC0-8) was lower than the plasma AUC0-8 (7574 +/- 1003 vs 25060 +/- 4171 ng/ml/h; P<0.004). However, both the elimination half-life (t1/2) and the mean residence time (MRT) of IDV intracellularly were prolonged compared with plasma (t1/2: 2.0 +/- 0.3 vs 1.2 +/- 0.09 h; MRT: 3.6 +/- 0.6 vs 2.1 +/- 0.1 h; P<0.05). All patients were responsive to therapy at the time of the study, as assessed by HIV plasma RNA levels. Individual plasma versus intracellular time course results suggest that, due to the prolonged intracellular half-life, some patients may achieve acceptable intracellular IDV concentrations despite sub-therapeutic plasma levels. Similarly, potentially inadequate intracellular concentrations may occur despite therapeutic plasma concentrations. CONCLUSIONS There is no significant intracellular accumulation of IDV within the lymphocytes of HIV-1-infected patients relative to plasma. However, intracellular concentrations are compatible with reported IDV-free drug concentrations in plasma. The intracellular elimination half-life and mean residence time of IDV are significantly prolonged compared with plasma. This may in part explain why certain patients maintain adequate viral suppression despite sub-therapeutic plasma IDV levels.