Establishing discordance rate of estimated glomerular filtration rate between serum creatinine-based calculations and cystatin-C-based calculations in critically ill patients.

INTRODUCTION

The use of serum creatinine (SCr) for drug dosing has significant limitations and is influenced by many non-kidney factors. Cystatin C (cysC) is an alternative or additional marker of kidney function that is less affected by non-kidney factors. Although cysC may be useful in hospitalized patients, the use of cysC to calculate drug dosing in critically ill patients has been incompletely investigated.

OBJECTIVE

The objective of this study was to determine the rate of discordance in estimated glomerular filtration rate (eGFR) between SCr-based calculations and SCr/cysC-based calculations that affect drug dosing in critically ill patients.

METHODS

This was a single-center, retrospective, observational cohort study at an academic medical center including critically ill adult patients admitted in 2023 with SCr and cysC ordered. Data were collected via chart review. Demographic data were analyzed via descriptive statistics. Discordance, defined as the percentage of times at which there is at least one discrepancy in kidney dosing for a medication using Cockcroft-Gault (CG) creatinine clearance versus Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) eGFR creatinine-cystatin C (eGFRcr-cys) equations, was analyzed via Wilcoxon matched pair signed ranked sum. eGFR calculations were normalized for patients' body surface area for comparison.

RESULTS

The study population included 232 patients (53.02% female; mean age 58.7 +/- 14.9 years; with 62.5% in medical, 23.28% in surgical, and 8.62% in neurological intensive care) with a median SCr of 0.94 mg/dL IQR [0.57-1.58] and median cysC of 1.92 mg/L IQR [1.27-2.77]. The median clearance rates were 68.5 mL/min (45.3-111.5) for CG and 53.9 mL/min (30.9-80.7) for CKD-EPI eGFRcr-cys; p < 0.001. The discordance rate across all study drugs was 32.3% (75/232). The four most common study drugs demonstrating discordance were cefepime 40.6% (52/128), vancomycin 38.3% (46/120), levetiracetam 35.1% (13/37), and piperacillin/tazobactam 11.6% (5/43).

CONCLUSION

Clinically significant discordance exists between SCr and SCr/cysC-based estimates of kidney function. This study established a discordance rate, as defined by drug dosing, of 32.3% in adult patients admitted to the ICU.