Impact of Serum Cystatin C–Based Glomerular Filtration Rate Estimates on Drug Dose Selection in Hospitalized Patients

Abstract

Study Objective Serum creatinine ( S c r ) concentration is used to calculate estimated glomerular filtration rate (e GFR ) for medication dosing. Serum cystatin C (CysC) concentration has been proposed as an adjunct or alternative to S cr . This study sought to evaluate the possible impact of using CysC in eGFR equations on drug dose recommendations in hospitalized patients with infections. Design Retrospective analysis of prospectively collected data. Setting Large academic tertiary care medical center. Patients A total of 308 adults with suspected or documented infections and stable kidney function who were hospitalized between 2012 and 2015. Measurements and Main Results Standardized S c r and CysC measured at the time of antibiotic dosing were used to estimate GFR from the three Chronic Kidney Disease Epidemiology Collaborative ( CKD ‐ EPI ) equations using S c r (eGFR Cr ), CysC( eGFR CysC ), or a combination of S c r and CysC (eGFR Cr‐CysC ), and these values were compared with estimated creatinine clearance (eCl cr ) from the Cockcroft‐Gault equation (standard of care for drug dosage adjustments). The eGFR s were categorized into five common dosage adjustment strata (lower than 20, 20–49, 50–79, 80–130, and higher than 130 ml/min), and agreement between equations was tested with the weighted κ statistic. Recommended drug doses varied considerably between the eC l cr and the CKD ‐ EPI equations (weighted κ [95% confidence interval]: eGFR Cr 0.73 [0.68–0.79], eGFR CysC 0.42 [0.35–0.5], eGFR Cr‐CysC 0.65 [0.6–0.71]). If eGFR Cr, eGFR CysC , or eGFR Cr‐CysC were used instead of eC l cr to dose drugs, 11%, 12%, and 8% of doses, respectively, would be higher, and 12%, 38%, and 24% of doses, respectively, would be lower. Conclusion Significant discordance in drug doses was observed when the CKD ‐ EPI equations were used in place of eCl cr . When CysC was included in eGFR equations, recommended doses were often lower. Further study is needed to develop and test drug‐specific dosing guidelines that incorporate alternate renal biomarkers and/or more contemporary eGFR equations.