Bayesian pharmacokinetic analysis of a gentamicin nomogram in neonates: a retrospective study

Abstract

Background: Although gentamicin is used extensively within the first week of life for suspected sepsis in neonates, little is known about the performance of gentamicin dosing nomograms in this population.

Objective: The goal of our study was to retrospectively assess the performance of a gentamicin dosing nomogram in neonates given gentamicin during the first week after birth.

Methods: In this retrospective study, gentamicin therapeutic drug monitoring data were collected during routine clinical care for all neonates who were born in St. Boniface General Hospital (Winnipeg, Manitoba, Canada) between January 1999 and April 2001 and given gentamicin during the first week after birth. We used Bayesian pharmacokinetic analysis to retrospectively assess the performance of our gentamicin dosing nomogram in neonates born at gestation ages <32 weeks, between 32 and 34 weeks, and >34 weeks. Bayesian pharmacokinetic values for parameters within groups were compared and used to explore predicted peak and trough serum gentamicin concentrations based on the institutional dosing nomogram.

Results: In a total of 58 neonates, those neonates born at ≤34 weeks' gestation had a weight-normalized apparent volume of gentamicin distribution 1.6 times larger than infants born after 34 weeks' gestation (P<0.001), as identified by Bayesian analysis. Weight-normalized gentamicin clearance was 22% lower in the youngest age category (P<0.01). Only 33% of predicted peak serum gentamicin concentrations were >6 mg/L for neonates born at ≤34 weeks' gestation, whereas 90% were therapeutic in neonates born at >34 weeks' gestation (P<0.001). With the present nomogram, the likelihood of an indication for adjustment of the dosing regimen was 12.4-fold higher (95% CI, 3.5-43.7) for those neonates born at ≤34 weeks' gestation.

Conclusions: These results have important clinical implications with regard to the advisability of determining peak serum gentamicin concentrations in neonates born at ≤34 weeks' gestation. Sampling of peak serum concentrations is indicated in this population to avoid underdosing and potential loss of therapeutic efficacy.

Keywords: Bayesian pharmacokinetics; gentamicin; neonatal sepsis; nomograms.

Figure 1

Bayesian-derived parameter estimates for gentamicin elimination rate constant (k_e) in relation to gestational age (GA) at birth. The dashed line represents the a priori pharmacokinetic values from the literature that were incorporated in the Bayesian estimation.

Figure 2

Bayesian-derived parameter estimates for volume of distribution (V_d) of gentamicin in relation to gestational age (GA) at birth. The dashed line represents the a priori pharmacokinetic values from the literature that were incorporated in the Bayesian estimation.

Figure 3

Performance of the institutional dosing nomogram with respect to predicted steady-state peak serum gentamicin concentrations (peak) and gestational age (GA) at birth. The symbols represent individual values from Bayesian pharmacokinetic analysis. Average concentration data are presented as mean (95% CI). ∗P<0.001 versus the group born at GA >34 weeks.

Figure 4

Performance of the institutional dosing nomogram with respect to predicted steady-state trough serum gentamicin concentrations (trough) relative to gestational age (GA). The symbols represent individual values from Bayesian pharmacokinetic analysis. Average concentration data are presented as mean (95% CI).