Introduction.Adherence to highly active antiretroviral therapy is required to obtain an optimal long term virologic response rate of HIV-1-infected children. Plasma concentrations of protease inhibitors (PIs) outside the limits of the reference values indicate nonadherence to antiretroviral therapy in adults. We studied during a 2-year follow-up period routinely taken plasma protease inhibitor concentrations to assess adherence to antiretroviral therapy in HIV-1-infected children.Methods.In 40 children (ages 3 months to 18 years) blood samples were taken at regular outpatient visits every 12 weeks after the start of highly active antiretroviral therapy and analyzed for plasma concentrations of indinavir or nelfinavir by high performance liquid chromatography and for plasma HIV-1 RNA load. The percentage of samples fulfilling the criteria for adherence was assessed for each child by three methods. For each sample a concentration ratio was calculated by dividing the concentration in that sample by the time-adjusted population value. According to Method 1 concentration ratios below or above concentration ratio limits (CORALs) of population data obtained in adults were highly indicative of nonadherence. Because many children have high PI levels, Method 2 evaluated plasma samples of PIs with only the lower CORAL. According to Method 3 only children with plasma samples below the limit of quantification (0.04 mg/l) were considered noncompliant. Differences in adherence rate between virologic responders and virologic nonresponders and between adherence rates and the two protease inhibitors were analyzed. The cumulative incidence of HIV-1 RNA levels >500 copies/ml in children was calculated.Results.Thirty-one children started treatment with indinavir, and nine children started treatment with nelfinavir. The median adherence rates for indinavir as determined by methods 1, 2 and 3 were 54% [interquartile range (IQR), 25 to 69%], 67% (IQR 50 to 92%) and 80% (IQR 63 to 100%), respectively. For nelfinavir median adherence rates of 60% (IQR 39 to 75%), 100% (IQR 67 to 100%) and 100% (IQR 100 to 100%) were observed. Adherence rates calculated with Method 2 were significantly higher in virologic responders (P= 0.04). Adherence rates calculated with Methods 2 and 3 were significantly lower in children using indinavir compared with those using nelfinavir (P= 0.02 andP= 0.02, respectively).Conclusion.Calculation of adherence rates using the lower limit of CORALs of indinavir or nelfinavir in children may be a useful measurement for the assessment of nonadherence to antiretroviral therapy in children.