Comparison of Cystatin C-Based and Serum Creatinine-Based Renal Function Estimates Against Timed Urine Collection in Critically Ill Patients

Background: Creatinine-based estimation equations for renal function, especially the Cockcroft-Gault creatinine clearance (CG-CrCL) equation, may be inaccurate in critically ill patients. We sought to determine if a cystatin C (CysC)-based (CKD-EPI CysC-serum creatinine (SCr) estimated glomerular filtration rate [eGFRcr-cys]) calculation better estimated renal clearance compared to CG-CrCL against measured creatinine clearance (mCrCL) from timed urine creatinine collection in adult critically ill patients. Materials and methods: This was a single-center retrospective cohort study of adult patients admitted to an intensive care unit (ICU) between July 1, 2022, and July 31, 2023, who had concurrent creatinine and cystatin C measured within 24 hours of a timed urine creatinine collection. The primary outcome was the mean difference in bias between eGFRcr-cys and C-G CrCL compared to timed urine creatinine collection. eGFR was adjusted for body surface area (BSA) and reported as mL/min. Results: Forty-five patients were included in this study. The mean bias between C-G CrCL and mCrCL was 23.8 mL/min (95% CI −40.3 to 87.8 mL/min) and 3.3 mL/min (95% CI −43.3 to 49.9 mL/min) between eGFRcr-cys and mCrCL. The mean difference between the biases was 20.5 mL/min ( P = .001). Conclusions: In critically ill patients, eGFRcr-cys was subject to less bias compared to C-G CrCL. Future research is needed to better understand the clinical implications of using eGFRcr-cys over C-G CrCL, particularly for medication-related decision making in critically ill patients.