OBJECTIVE: This work aimed at building a population pharmacokinetic (PK) model for lamivudine (LMV), stavudine (STV) and zidovudine (ZDV), estimating their inter and intraindividual PK variability and investigating the influence of different covariates. METHODS: Population PK of LMV, STV and ZDV was separately evaluated from plasma concentrations obtained in 54, 39 and 27 HIV1-infected patients, respectively, enrolled in the COPHAR1-ANRS102 trial. The primary objective of this trial was to study the pharmacokinetics of indinavir (IDV) and nelfinavir (NFV) in treated patients with a sustained virological response. Concentrations of nucleoside analogs (NA) were measured in plasma as a secondary objective. A one-compartment model with first-order elimination was used, with zero-order absorption for LMV and first-order absorption for STV and ZDV. RESULTS: Mean parameters [interpatient variability in coefficient of variation (CV%)] of LMV, STV and ZDV were: oral volume of distribution (V/F) 145 l (52%), 24 l (81%) and 248 l (80%), oral clearance (Cl/F) 32 l/h, 16 l/h (74%) and 124 l/h (51%), respectively. For LMV, absorption duration (Ta) was 1.46 h (64%). For STV and ZDV, ka was 0.46 h(-1) and 2.9 h(-1), respectively. We found a systematic effect of combination with NFV vs. IDV. We found that intrapatient variability was greater than interpatient variability (except for STV) and greater than 55% for the three drugs. CONCLUSION: This trial enabled the estimation of the population PK parameters of three NA in patients with a sustained virological response, and the median curves could be used as references for concentration-controlled strategies. We observed, as for the protease inhibitors, a great variability of PK parameters.