ObjectivesTo describe plasma concentrations of indinavir alone or combined with ritonavir, and of nelfinavir and its active metabolite M8, and to measure their variabilities in HIV-infected patients treated with a stable antiretroviral regimen and experiencing a sustained virological response for at least 12 months.Patients and methodsIn this prospective trial, blood samples were drawn during a 6-hour time interval between two doses at enrolment to assess protease inhibitor (PI) pharmacokinetic parameters, and 4 months later to assess plasma trough and peak concentrations. Safety and adherence assessments and laboratory data were collected during an 8-month period. PI pharmacokinetic characteristics were analysed using a non-compartmental approach. Inter- and intrapatient variabilities were estimated using a linear mixed-effect model. The impact of different covariates on plasma trough concentrations was investigated. Eighty-eight patients were analysed: 42 treated with indinavir and 46 with nelfinavir.ResultsThe interquartile range (IQR) of the plasma trough concentration corrected for the sampling time (Ccalc) was 116–374 μg/L for indinavir alone and 163–508 μg/L for indinavir/ritonavir. Ritonavir significantly increased indinavir elimination half-life and plasma exposure. For nelfinavir, the IQR of Ccalcwas 896–2059 μg/L for three-times-daily administration and 998–2124 μg/L for twice-daily administration. Variabilities were high for both PIs. Intrapatient variability for indinavir alone (and indinavir + ritonavir) was 76% (107%) and interpatient variability was 58% (10%) in adherent patients. Intrapatient variability for nelfinavir three times daily (and twice daily) was 41% (74%) and interpatient variability was 62% (50%). Intrapatient variability was lowered in patients with a high adherence level.ConclusionAlthough performed in a homogeneous population, this study documented a high interpatient but also intrapatient variability of indinavir and nelfinavir pharmacokinetics, which should be taken into account when interpreting therapeutic drug monitoring. Once patients have reached a sustained virological response, plasma PI monitoring may have a limited impact.