DOI: 10.1161/circ.148.suppl_1.17090 ISSN: 0009-7322

Abstract 17090: Prognostic Value of Creatinine versus Cystatin C-Based Estimation of Glomerular Filtration Rate for Mortality in Heart Failure

Bethany Roehm, Jonathan Gordon, Meredith McAdams, Justin L Grodin, Susan Hedayati
  • Physiology (medical)
  • Cardiology and Cardiovascular Medicine

Introduction: Serum creatinine and cystatin C are filtration markers used to calculate estimated glomerular filtration rate (eGFR). Lower eGFR using both is associated with increased mortality in heart failure with reduced ejection fraction (HFrEF). Cystatin C may be more accurate in advanced HFrEF with sarcopenia. Prognostic ability of each (eGFRcr or eGFRcys) or their weighted average (eGFRcr-cys) has not been compared.

Hypothesis: eGFRcr-cys is more predictive of mortality in HFrEF than eGFRcr or eGFRcys.

Methods: We analyzed data from 310 participants from Registry Evaluation of Vital Information for Ventricular Assist Devices in Ambulatory Life with baseline serum creatinine and cystatin C. Primary outcome was all-cause mortality. Cox models were adjusted for age, sex, race, diabetes mellitus, and NYHA class.

Results: Median eGFR by cr, cys and cr-cys were 60, 58 and 60 ml/min/1.73m 2 . During 2-year follow-up, 45 participants died. Within each patient, the difference of eGFRcr and eGFRcys varied from -49 to 50 ml/min/1.73m 2 . Among all estimates, those with baseline eGFR <60 ml/min/1.73m 2 had worse survival (Figure). Lower baseline eGFRcr [HR 0.97, 95% CI (0.96-0.99), P =0.008, C= 0.67], eGFRcys [HR 0.98 (0.97-0.99), P= 0.006, C= 0.68], and eGFRcr-cys [HR 0.98 (0.96-0.99), P= 0.004, C= 0.68] were associated with higher mortality. C-statistic comparison did not show a significant difference for predictive ability: eGFRcr vs. eGFRcys, P=0.7; eGFRcr vs. eGFRcr-cys, P=0.9; eGFRcys vs. eGFRcr-cys, P=0.3).

Conclusions: Using all three estimates, lower eGFR was associated with higher mortality, although all had relatively poor predictive performance. eGFRcys was expected to be similar to or lower than eGFRcr due to decreased muscle mass, yet at the individual level there was wide variability between estimates which could in part explain their poor predictive abilities. There is need for more specific kidney filtration markers for prediction of mortality in HFrEF.

More from our Archive