#4162 COMPARISON OF EGFR CYSTATIN C AND EGFR CREATININE IN RECENTLY HOSPITALISED INDIVIDUALS

Abstract

Abstract Background and Aims Creatinine and cystatin C are endogenous biomarkers that can be used to estimate glomerular filtration rate (GFR). Cystatin C may be less dependent on age, gender and muscle mass. The comparative performance of these markers in patients recently hospitalised patients with acute kidney injury (AKI), who may have experienced changes in body composition, is unclear. Method Two matched cohorts of hospitalised individuals who had survived to 90 days after admission were recruited. The cohorts consisted of people who had sustained AKI during hospital admission and those who had not, and were matched 1:1 for age, baseline eGFR stage and diabetes. Serum creatine and serum cystatin C were measured at 3 months after hospitalisation and eGFR calculated using CKD EPI creatinine (eGFRcr) and CKD EPI cystatin C (eGFRcys) equations. Difference between the measures (eGFRdiff) was calculated as eGFRcys-eGFRcr and percentage difference (eGFR%diff) was relative to the mean of both eGFR measures. Primary outcome was mortality after 5 years of prospective follow up. Univariable survival analyses were conducted with the Kaplan Meier method and comparisons were performed with the log rank test, and a fully adjusted multivariable analysis performed with Cox proportional hazards model. Results 854 individuals were recruited, matched, and had paired creatinine and cystatin C measurements. 427 (50%) had sustained AKI. Median eGFRcys was lower than eGFRcr (53.5ml/min/1.73 m2 [34.4-85.4] vs 68.4ml/min/1.73 m2 [52.6-84.7], p<0.001). eGFRcys and eGFRcr were correlated (r = 0.486, p<0.001) but Bland Altman analysis showed variable bias across the eGFR range. More individuals had an eGFR <60ml/min using eGFRcys (57%) compared with eGFRcr (35.6%). There was a graded relationship between eGFR%diff and outcome, with a shorter survival time (Figure 1) and higher proportion of deaths at 5 years (Q1 64 (30.0%), Q2 54 (25.4%), Q3 41 (19.2%), Q4 (31 (14.6%), p<0.001) in the first quartile. Cox proportional hazards analysis showed that eGFR%diff was an independent predictor of 5-year mortality after adjustment for age, Charlson index score and albuminuria (adjusted OR 0.995 [0.992-0.998], p = 0.002). Conclusion Cystatin C eGFR measures were lower than creatinine eGFR measures in recently hospitalised individuals. Lower eGFRcys relative to eGFRcr was independently associated with increased mortality. A larger difference between eGFRcys and eGFRcr may reflect loss of muscle mass (resulting in relatively lower serum creatinine) in somebody who has recently had an acute illness. Further, these results confirm how eGFRcr may over-estimate kidney function and potentially miss patients with failed renal recovery at 3-months after AKI. The difference between eGFRcys and eGFRcr may be helpful in identifying people at greater risk of early mortality after an acute illness.