Variables of interest to predict glomerular filtration rate in preterm newborns in the first days of life

Abstract

Background: Measurement of neonatal renal function is challenging, and accurate, easy-to-use markers to estimate glomerular filtration rate (eGFR) are lacking. This study aimed to evaluate principal determinants of GFR in neonates and develop a predictive equation.

Methods: GFR was measured, using single injection inulin clearance, at median day 3 of life in 48 newborns. Associations of clearance with height, gestational age, weight, creatinine, and cystatin C were explored and a multivariable model to estimate GFR developed. We also evaluated preexisting GFR equations (Schwartz, Zappitelli, combined Zappitelli).

Results: Forty-four clearances were measured, 36 very preterm neonates (28-32 weeks); 5 extremely preterm (< 28 weeks), and 3 term newborns. No patient presented acute renal insufficiency. Median inulin clearance in preterm infants was 18.83 ml/min/1.73 m² (IQ 15.29; 24.99). Inulin clearance correlated with weight (ρ 0.74), gestational age (ρ 0.72), height (ρ 0.49), and creatinine (ρ - 0.42), but not cystatin C. In the multivariable model, predicted GFR equation was 2.32* (weight (g))^0.64/(creatinine (mcmol/l))^0.62. Mean error in predicting clearance was - 0.8 ml/min/1.73 m² (- 3.0-1.4) ranging from - 14.9 to 13.3 ml/min/1.73 m². Mean prediction error with Zappitelli and combined Zappitelli equations were 28.5 ml/min/1.73 m² (95% CI 24.6-32.3) and 28.3 ml/min/1.73 m² (95% CI 24.9-31.7), respectively, and 2 ml/min/1.73 m² (95% CI - 0.6-4.6) for Schwartz equation.

Conclusions: Weight and gestational age are crucial determinants of GFR in neonates. The Zappitelli models were not validated in our population. Our predictive model and Schwartz models performed better. Our model should be evaluated in another preterm population, particularly in those presenting renal insufficiency.

Keywords: Estimated glomerular filtration rate; Glomerular filtration rate; Newborns; Preterm.