Measured and Estimated Glomerular Filtration Rate in the ICU: A Prospective Study.

OBJECTIVES: To compare estimated glomerular filtration rate using classical static and kinetic equations with measured glomerular filtration rate assessed by plasma iohexol clearance in a mixed population of critical care patients. PATIENTS: Unselected patients older than 18 and admitted to a general ICU. DESIGN: Interventional prospective single center study. INTERVENTION: Measurement of glomerular filtration rate by the plasma clearance of an IV single dose of iohexol and estimation of glomerular filtration rate with creatinine or cystatin C-based standard and kinetic equations as well as urinary creatinine clearance. MEASUREMENTS AND MAIN RESULTS: Sixty-three patients were included with a median age of 66 years old. The median measured glomerular filtration rate was 51 mL/min/1.73 m (interquartile range, 19-85 mL/min/1.73 m). All used equations displayed significant biases, high errors, and poor accuracy when compared with measured glomerular filtration rate, overestimating renal function. The highest accuracy and lowest error were observed with cystatin C-based chronic kidney disease epidemiology collaboration equations. Both modification of diet in renal disease and Cockcroft-Gault equations displayed the lowest performance. Kinetic models did not improve performances, except in patients with unstable creatinine levels. Creatinine- but not cystatin C-based estimations largely derived over ICU stay, which appeared more related to sarcopenia than fluid balance. Finally, estimated glomerular filtration rate misclassified patients according to classical glomerular filtration rate categories in approximately half of the studied cases. CONCLUSIONS: All known estimated glomerular filtration rate equations displayed high biases and unacceptable errors when compared with measured glomerular filtration rate in a mixed ICU population, with the lowest performance related to creatinine-based equations compared with cystatin C. In the ICU, we advocate for caution when using creatinine based estimated glomerular filtration rate equations. Drifting of serum creatinine levels over time should also be taken into consideration when assessing renal function in the ICU.