(1413) Comparison of Renal Function Monitoring Modalities in Critically Ill Patients

Introduction: Serum creatinine (SCr) is a biomarker commonly used to estimate renal function and select medication dosing, but there may be limitations to use in critically ill patients. Cystatin C (CysC) is an alternative biomarker less affected by age, sex, weight, and muscle wasting and therefore potentially more reliable. Studies comparing SCr and CysC-based renal estimates against the gold-standard timed urine collection in critically ill patients are lacking. This study aimed to determine if a CysC-based (CKD-EPI CysC-SCr estimated glomerular filtration rate) or a SCr-based (Cockroft-Gault creatinine clearance) calculation better estimated renal clearance compared to a timed urine collection in critically ill patients.

Methods: This was a single-center retrospective cohort study of adult critically ill patients admitted between July 1, 2022, and July 31, 2023. Patients who had a SCr, CysC, and timed urine collection obtained within 24 hours of each other were included. The primary outcome was mean difference in bias between CysC-based and SCr-based estimated renal clearance compared to timed urine collection measured renal clearance.

Results: Forty-five patients were included. The mean bias between SCr-based estimation and timed urine collection was 23.8 mL/min (95% CI -40.3-87.8 mL/min) and 3.3 mL/min (95% CI -43.3-49.9 mL/min) between CysC-based estimation and timed urine collection. The mean difference between the biases was 20.5 mL/min (p=0.001). Serum creatinine-based and CysC-based estimations were on average 63% and 8.7% higher than timed urine collection, respectively. Thirty-six (80%) patients were receiving renally dosed medications; of the medications administered when the timed urine collection was completed, 60.3% would be correctly dosed based on the SCr-based estimate while 75.8% would be correctly dosed based on the CysC-based estimate (p=0.07).

Conclusions: In critically ill patients, CysC-based renal function calculation led to less bias compared to timed urine collection while SCr-based calculation tended to overestimate clearance. Future research is needed to better understand the clinical implications of using CysC-based renal function calculations in critically ill patients, particularly for medication dosing.