Usefulness of estimating individual pharmacokinetic data for aminoglycoside therapy in seriously ill patients.

This study proposes the use of individual pharmacokinetic parameters to predict effective dosages of aminoglycosides for patients with serious infections. Parameters were calculated for 51 patients using a one-compartment model. This model required a trough and three postinfusion aminoglycoside concentrations from each patient, as well as least squares linear regression. A total of 56 dosing regimens (37 gentamicin, 19 tobramycin) were predicted from these parameters and the efficacy of this approach was examined by comparing the predicted peak (8 mg/l) and trough (1.5 mg/l) concentrations with the observed peak and trough concentrations measured 24 h later. Aminoglycoside concentrations in serum were measured using a fluorescence polarisation immunoassay. Parameters varied widely, as the following ranges demonstrate: Volume of distribution 0.16-0.52 l/kg; clearance 0.04-0.17 l/kg/h; half-life 1.1-5.3 h. The doses predicted ranged from 3.2-16.9 mg/kg/day, with an average of 8.6 (standard deviation [SD] 3.4) mg/kg/day. The measured peak (mean 7.5; SD 1.6) and trough (mean 1.95; SD 0.57) concentrations closely approximated the predicted levels. The equations used underestimated trough concentrations mainly because of aminoglycoside accumulation. The changing pharmacokinetic parameters for each patient meant that regular drug assays were still required to fine-tune the dosing regimen. Two patients (4%) developed significant nephrotoxicity. Ototoxicity was not assessed. We conclude that individual pharmacokinetic data for aminoglycoside therapy can be used effectively to predict dosages for desired concentrations in seriously ill patients. The dose required was on average 1.5-2 times the normal recommended dosage.