"Random" gentamicin concentrations do not predict trough levels in neonates receiving once daily fixed dose regimens

BACKGROUND: Monitoring plasma gentamicin concentrations in neonates 24 hours after a once daily dose (4 mg/kg) often necessitates additional blood sampling. In adults a nomogram has been developed enabling evaluation of gentamicin doses by sampling concentrations with other blood tests, 4 – 16 hours after administration. We attempted to develop a similar nomogram for neonates. METHODS: In addition to standard 24 hour sampling to monitor trough concentrations, one additional "random" gentamicin concentration was measured in each of 50 neonates <4 days of age (median gestation 33 weeks [28–41]), when other blood samples were clinically necessary, 4 – 20 hours after gentamicin administration. 24 hour concentrations of >1 mg/L were considered high, and an indication to extend the dosing interval. RESULTS: Highest correlation (r(2 )= 0.51) of plasma gentamicin concentration against time (4 to 20 hours) was with logarithmic regression. A line drawn 0.5 mg/L below the true regression line resulted in all babies with 24 hr gentamicin concentrations >1 mg/L having the additional "random" test result above that line, i.e. 100% sensitivity for 24 hour concentrations>1 mg/L, though only 58% specificity. Having created the nomogram, 39 further babies (median gestation 34 weeks [28–41]), were studied and results tested against the nomogram. In this validation group, sensitivity of the nomogram for 24 hr concentrations >1 mg/L was 92%; specificity 14%, positive predictive value 66%, and negative predictive value 50%. Prematurity (≤ 37 weeks) was a more sensitive (94%) and specific (61%) indicator of high 24-hour concentrations. 62 (87%) of 71 preterm babies had high 24-hour concentrations. CONCLUSION: It was not possible to construct a nomogram to predict gentamicin concentrations at 24 hours in neonates with a variety of gestational ages. Dosage tailored to gestation with monitoring of trough concentrations remains management of choice.