Background Hyperfiltration refers to the enhanced re-nal elimination of circulating solute. It is an increasingly recognised phenomenon in critically ill adults, leading to subtherapeutic treatment of renally cleared drugs. Although its existence has also been suggested in critically ill children, concrete data are currently lacking.

Objectives The primary aim of this pilot study was to investigate the incidence and risk factors of hyperfiltration in a paediatric intensive care setting. Additionally, a comparison of different methods for glomerular filtration rate (GFR) assessment in critically ill children was made.

Methods The HYPIC study was a single centre, prospective, observational study, conducted at the paediatric intensive care unit (PICU) and the cardiac surgery intensive care unit (CSICU) of the Ghent University Hospital, Belgium, enrolling patients between 1 month and 15 years of age. GFR was estimated by means of a calculated 24 hour creatinine clearance (24 hour CrCL). Creatinine in serum and urine were determined using the Jaffe’s reaction, and corrected for interfering total protein concentration accord-ing to Speeckaert et al.1 The Larsson formula was used for cystatin C-based estimation of GFR.2 Hyperfiltration was defined as a GFR exceeding normal values for age plus two standard deviations. Logistic regression analysis was used to evaluate risk factors for hyperfiltration. GFR assessment methods (24 hour CrCl, modified Schwartz formula and Larsson formula) were compared using Bland-Altman plots.3

Results Data were collected from 58 patients (median age: 20 months; age range: 1 month to 15 years). Hyper-filtration was present in 80.8% of patients. Body length was identified to be an independent risk factor for hyper-filtration (p=0.05). Although not statistically significant, body surface area (p=0.12) and a neurological admission reason (p=0.12) also seem related to the development of hyperfiltration. A systematic difference between calcu-lated creatinine clearance (24 hour CrCL) and the estimated GFR (eGFR) using the modified Schwartz formula was observed (mean difference 28.9 ml/min/1.73 m²; SD60.4 ml/min/1.73 m²). The Schwartz formula was accurate at low GFR, but underestimated the GFR at higher values. The mean difference of GFR between the Larsson formula and the 24 hour CrCl was very low (3.67 ml/kg/m²; SD66.9 ml/min/1.73 m²).

Conclusion Hyperfiltration is a common phenomenon in critically ill children. The modified Schwartz formula is likely to underestimate GFR in case of hyperfiltration. Cystatin C seems a promising alternative renal biomarker but needs further investigation.