Aims: To determine the population pharmacokinetics of intramuscular (i.m.) gentamicin in African infants with suspected severe sepsis.
Methods: Samples were withdrawn 1 h after a single i.m. injection of 8 mg x kg(-1) gentamicin and the next morning prior to any further dosing. Concentration-time data were analysed with the population pharmacokinetic package NONMEM. Data were fitted using a one-compartment model with a log-normal model for interindividual variability and an additive residual error model. The influence of a range of clinical characteristics was tested on the pharmacokinetics of intramuscular gentamicin and the effect of incorporating interindividual variability on bioavailability was examined.
Results: The data set comprised 107 patients and 203 concentrations. Peak concentrations ranged from 3.0 mg x L(-1) to 19.8 mg x L(-1) (median 10.6 mg x L(-1)) and 'next day' samples from 0.3 mg x L(-1) to 6.2 mg x L(-1). The best models were clearance/bioavailability (CL) (L x h(-1)) = 0.0913 x weight (kg) x (age (days) + 1)/11)0.130 and volume of distribution/bioavailability (V) = 2.02 x (1 + 0.277 x (weight -3)). Therefore, an infant with the median weight of 3 kg and age 10 days would have a predicted CL of 0.274 L x h(-1) and V of 2.02 L. Interindividual variability in CL was 40% and in V was 42%. This model required a term for covariance between CL and V. When variability in bioavailability was introduced as an alternative model, interindividual variability in CL was 22%, in V 18% and in relative bioavailability 36%.
Conclusions: Intramuscular administration of 8 mg x kg(-1) gentamicin daily to infants gives mean 1 h peak concentration of 10.6 mg x L(-1) and a trough concentration of less than 2 mg x L(-1). Wide variability in the peak concentration may reflect variable absorption rate or bioavailability.