Examining the predictive performance of different equations to estimate glomerular filtration rate in paediatric oncology patients.

Aims: The National Kidney Foundation guidelines recommend the use of equations to estimate glomerular filtration rate (GFR) in paediatrics and adults, some of which require serum creatinine plasma concentrations. However, serum creatinine may not always be reflective of renal function in paediatrics, as a lack of muscle mass may lead to very low serum creatinine concentrations, which may be misinterpreted as good renal function. This study aims to evaluate the predictive performance of five equations to estimate GFR in paediatric oncology patients in whom a measured GFR had been obtained from the excretion of 51Cr-EDTA.

Methods: Data were collected retrospectively from the medial records of paediatric oncology patients at the Lady Cilento Children Hospital in Brisbane, Australia from 2007 to 2013. Patients had undergone a radioisotopic method to measure their GFR prior to treatment with cisplatin, carboplatin, tacrolimus, cyclosporine or a bone marrow transplant. Estimates of GFR (mL/min) for these patients were obtained from five equations suggested to be useful in children (1-4). GFR predicted by the five equations was compared to GFR measured by excretion of 51Cr-EDTA and equation predictive performance was examined in terms of bias, precision and accuracy. To examine how best to handle serum creatinine concentrations < 30 µmol/L which is below the assay limit of quantification these values were either replaced by 15 µmol/L, 30 µmol/L or an expected mean serum creatinine concentration based on patient age and gender according to Ceriotti et al (5). 

Results: Sixty-eight children (median age 3.7 years) contributed 118 measurements of measured GRF. Measured GFR (mL/min) values ranged from 7 to 146 (median: 38.5). The equation proposed by Rhodin et al. (2) yielded the least bias, the highest precision and the highest accuracy in estimating GFR reflected by a mean prediction error, a root mean square error and a mean relative percentage prediction error of -0.13 mL/min, 12.3 mL/min and 7.96%, respectively. The Schwartz equation (1) and the equation proposed by Cole et al.(3), also performed reasonably well with a mean relative percentage prediction error of 14.6% and 25%, respectively. The least accurate model was the equation proposed by Leger et al. (4). 60% of collected serum creatinine concentrations were reported as < 30 µmol/L. Setting those serum creatinine concentrations to 15 µmol/L resulted in the least accurate and the most biased estimate of GFR for all tested equations that required a measurement of serum creatinine concentration.

Conclusions: This study shows that an equation by Rhodin et al. which does not use serum creatinine concentration to estimated GRF had the best agreement with measured GFR obtained by 51Cr-EDTA clearance in oncology children. This equation which describes maturation of renal function in children using total body weight and post-menstrual age avoids problems associated with serum creatinine concentrations below the assay quantification limit.


1.Schwartz GJ, Munoz A, Schneider MF, Mak RH, Kaskel F, Warady BA, et al. New equations to estimate GFR in children with CKD. Journal of the American Society of Nephrology : JASN. 2009;20(3):629-37.

2.Rhodin MM, Anderson BJ, Peters AM, Coulthard MG, Wilkins B, Cole M, et al. Human renal function maturation: a quantitative description using weight and postmenstrual age. Pediatric nephrology. 2009;24(1):67-76.

3.Cole M, Price L, Parry A, Keir MJ, Pearson AD, Boddy AV, et al. Estimation of glomerular filtration rate in paediatric cancer patients using 51CR-EDTA population pharmacokinetics. British journal of cancer. 2004;90(1):60-4.

4.Leger F, Bouissou F, Coulais Y, Tafani M, Chatelut E. Estimation of glomerular filtration rate in children. Pediatric nephrology. 2002;17(11):903-7.

 5.Ceriotti F, Boyd JC, Klein G, Henny J, Queralto J, Kairisto V, et al. Reference intervals for serum creatinine concentrations: assessment of available data for global application. Clinical chemistry. 2008;54(3):559-66.