Only a few data have thus far been published on the pharmacokinetics of amikacin in neonates. To gain further information on this issue, we studied a series of 32 neonates who were treated with amikacin for suspected or documented bacterial infection. Nineteen neonates were preterm (mean gestational age = 32.0 +/- 3.6 weeks, mean body weight = 1.74 +/- 0.81 kg) while the remaining 13 were full-term (mean body weight = 3.19 +/- 0.82 kg). The 32 neonates were given amikacin by intramuscular route. A total of 121 concentrations were measured (average number of concentrations per patient = 3.8; range 3-6). To estimate amikacin pharmacokinetic parameters, the serum concentration values of the drug were fitted to the one-compartment pharmacokinetic model. The calculated pharmacokinetic parameters were the following (mean +/- SD): clearance = 64.6 +/- 30.8 ml/h/kg; volume of distribution = 0.655 +/- 0.414 liters/kg; half-life = 7.6 +/- 4.4 h. These results are similar to the values reported previously, with the important exception of the volume of distribution, which was considerably higher in our study. The intraindividual variability of amikacin pharmacokinetics was evaluated by the standard two-stage method yielding an intraindividual variability coefficient of 28.9%. No previous estimate of this parameter has as yet been published. The population parameters of amikacin in neonates, derived from the present study (i.e. coefficient for intraindividual variability and means +/- SD for clearance and volume of distribution), can be applied to a further series of neonates to facilitate the prospective use of the bayesian method for individualizing amikacin dosage.