P25 Comparison of renal function estimation methods in critically ill children: a pilot study

Abstract

Background Accurate assessment of renal function is crucial in intensive care to guide therapy. Both acute kidney injury and augmented renal clearance (ARC) may compromise outcome. Common formulas to estimate glomerular filtration rate (GFR) are unreliable in critically ill adults.1 A comparison of a gold standard technique to assess GFR with these formula-based estimations has never been reported in pediatric intensive care (PICU) patients. Our aim was to evaluate the feasibility of measuring plasma iohexol clearance (CLIOHEX) for GFR assessment in critically ill children and to compare CLIOHEX with estimated GFR using the modified Schwartz formula (eGFRSchwartz). Methods A prospective, interventional study was conducted at the PICU of the Ghent University Hospital, Belgium. Critically ill children without chronic kidney disease were included. After injection of a weight-dependent bolus of iohexol, serial blood samples (n=6) were taken over a 6-hours interval. CLIOHEX was compared to eGFRSchwartz. Correlation between both methods was assessed by a Pearson's correlation coefficient (r). Bland-Altman plots were evaluated to assess bias and limits of agreement (LOA). ARC was defined as a GFR exceeding normal values for age plus two standard deviations. Results 40 patients, median age 16 months (range 15 days - 13,6 years), 72,5% males, were included. No adverse effects related to iohexol were observed. Median CLIOHEX was 121 ml/min/1.73m2 (range: 43-221 ml/min/1.73m2). ARC was present in 20 patients based on CLIOHEX. Median eGFRSchwartz was 81 ml/min/1.73m2 (range: 31-131 ml/min/ 1.73m2). Only 1 patient was identified with ARC by eGFRSchwartz. eGFRSchwartz was systematically lower than CLIOHEX. There was a good correlation between CLIOHEX and eGFRSchwartz (r = 0,69; p< 0,01). Bias was 34 ml/min/1.73m2 with LOA (-24,5; 93 ml/min/1.73m2) Conclusion CLIOHEX was safely used to measure true GFR in critically ill children. eGFRSchwartz systematically underestimates GFR, especially in patients with ARC and seems not reliable in this patient population.