Glomerular Filtration Rate in Asphyxiated Neonates Under Therapeutic Whole-Body Hypothermia, Quantified by Mannitol Clearance

Abstract

Background: Therapeutic hypothermia (TH) is an established intervention to improve the outcome of neonates with moderate-to-severe hypoxic-ischemic encephalopathy resulting from perinatal asphyxia. Despite this beneficial effect, TH may further affect drug elimination pathways such as the glomerular filtration rate.

Objectives: The objective of this study was to quantify the effect of TH in addition to asphyxia on mannitol clearance as a surrogate for the glomerular filtration rate.

Methods: The effect of asphyxia and TH (mild vs moderate/severe) on mannitol clearance was assessed using a population approach, based on mannitol observations collected in the ALBINO (ALlopurinol in addition to TH for hypoxic-ischemic Brain Injury on Neurocognitive Outcome) trial, as some were exposed to a second dose of 10 mg/kg intravenous mannitol as placebo to ensure blinding. Pharmacokinetic analysis and model development were conducted using NONMEM version 7.4.

Results: Based on 77 observations from 17 neonates (TH = 13), a one-compartment model with first-order linear elimination best described the observed data. To account for prenatal glomerular filtration rate maturation, both birthweight and gestational age were implemented as clearance covariates using an earlier published three-quarters power function and a sigmoid hyperbolic function. Our final model predicted a mannitol clearance of 0.15 L/h for a typical asphyxia neonate (39.5 weeks, birthweight 3.25 kg, no TH), lower than the reported value of 0.33 L/h for a healthy neonate of similar age and weight. By introducing TH as a binary covariate on clearance, the additional impact of TH on mannitol clearance was quantified (60% decrease).

Conclusions: Mannitol clearance was decreased by approximately 60% in neonates undergoing TH, although this is likely confounded with asphyxia severity.

Trial registration: ClinicalTrials.gov identifier NCT03162653.

Fig. 1

a–f Covariate-η plots of the covariates (gestational age, birthweight, alanine aminotransferase, lactate dehydrogenase, and therapeutic hypothermia) that showed a clear trend (|corr| > 0.3 or p < 0.05). CL clearance, V volume of distribution

Fig. 2

a Observed vs population-predicted concentrations and b observed vs individual-predicted concentrations. The solid blue line represents loess smoothing. c Conditional weighted residuals (CWRES) vs population-predicted concentrations and d time. The solid blue line represents loess smoothing. e Distribution of the normalized prediction distribution errors (NPDEs). The solid red line represents a normal distribution, whereas the solid blue line represents the actual distribution of the NPDEs. f QQ plot of NPDEs