Amoxicillin plasma concentrations, pharmacokinetic parameters, and the influence of demographic, anthropometric, and clinical covariates were investigated in 150 neonates. Gestational age (GA) ranged from 25 to 42 weeks and mean postnatal age (PNA) was 0.8 days. Amoxicillin concentrations were measured with reversed-phase HPLC in surplus plasma from routine assays of coadministered gentamicin. Mean total body clearance corrected for body weight (CL/W) was 0.096 +/- 0.036 L/kg(-1)h(-1), mean elimination half-life (t(1/2)) was 5.2 +/- 1.9 hours, and mean volume of distribution corrected for body weight (V/W) was 0.65 +/- 0.13 L/kg. Multiple regression equations were calculated for the prediction of CL/W amoxicillin. CL/W gentamicin, V/W gentamicin, and GA were significant predictors of CL/W amoxicillin. Amoxicillin peak and trough concentrations after the second dose and the time the concentration exceeds the minimum inhibitory concentration (T>MIC), reached with the current dosage regimen, were evaluated. Toxic plasma concentrations were reached in several patients. Therefore, the authors have proposed a lower dosage regimen, based on GA, population pharmacokinetic parameters, bacterial susceptibility (T>MIC), and possible toxicity: 15 mg/kg per 8 hours and 20 mg/kg per 8 hours for neonates with GA < or = 34 and GA>34 weeks, respectively. Simulation with this new dosage regimen indicated that satisfactory plasma concentrations were reached in all 150 neonates. Therefore, use of therapeutic drug monitoring and pharmacokinetic calculations for dosage adjustment is generally not necessary.