Predicted and measured aminoglycoside pharmacokinetic parameters in critically ill patients.

We conducted a prospective study to determine whether predicted aminoglycoside pharmacokinetic parameters (based on population averages) correlate with measured values in critically ill patients. There was wide interpatient variability for all pharmacokinetic parameters. Only predicted and measured volumes of distribution (18.7 +/- 6.5 versus 22.9 +/- 7.7 liters [mean +/- standard deviation], respectively), with a mean of 0.32 +/- 0.09 liter/kg of dosing body weight, were significantly different. There were no relationships between pharmacokinetic parameters and documented infection, death, or intubation status. The results indicate that volume of distribution is commonly underestimated in intensive care unit patients, whereas elimination rates may be adequately predicted based on population averages. We therefore recommend that aminoglycoside volume of distribution estimates for intensive care unit patients take fluid and adipose excess into account and be based on 0.32 liter/kg rather than the usual 0.25 liter/kg.