The Difference Between Cystatin C– and Creatinine-Based Estimated GFR in Heart Failure With Reduced Ejection Fraction: Insights From PARADIGM-HF

Rationale & Objective The clinical implications of the discrepancy between cystatin C (cysC)– and serum creatinine (Scr)–estimated glomerular filtration rate (eGFR) in patients with heart failure (HF) and reduced ejection fraction (HFrEF) are unknown. Study Design Post-hoc analysis of randomized trial data. Setting & Participants 1,970 patients with HFrEF enrolled in PARADIGM-HF with available baseline cysC and Scr measurements. Exposure Intraindividual differences between eGFR based on cysC (eGFRcysC) and Scr (eGFRScr; eGFRdiffcysC-Scr). Outcomes Clinical outcomes included the PARADIGM-HF primary end point (composite of cardiovascular [CV] mortality or HF hospitalization), CV mortality, all-cause mortality, and worsening kidney function. We also examined poor health-related quality of life (HRQoL), frailty, and worsening HF (WHF), defined as HF hospitalization, emergency department visit, or outpatient intensification of therapy between baseline and 8-month follow-up. Analytical Approach Fine-Gray subdistribution hazard models and Cox proportional hazards models were used to regress clinical outcomes on baseline eGFRdiffcysC-Scr. Logistic regression was used to investigate the association of baseline eGFRdiffcysC-Scr with poor HRQoL and frailty. Linear regression models were used to assess the association of WHF with eGFRcysC, eGFRScr, and eGFRdiffcysC-Scr at 8-month follow-up. Results Baseline eGFRdiffcysC-Scr was higher than +10 and lower than −10 mL/min/1.73 m2 in 13.0% and 35.7% of patients, respectively. More negative values of eGFRdiffcysC-Scr were associated with worse outcomes ([sub]hazard ratio per standard deviation: PARADIGM-HF primary end point, 1.18; P = 0.008; CV mortality, 1.34; P = 0.001; all-cause mortality, 1.39; P < 0.001; worsening kidney function, 1.31; P = 0.05). For a 1–standard-deviation decrease in eGFRdiffcysC-Scr, the prevalences of poor HRQoL and frailty increased by 29% and 17%, respectively (P ≤ 0.008). WHF was associated with a more pronounced decrease in eGFRcysC than in eGFRScr, resulting in a change in 8-month eGFRdiffcysC-Scr of −4.67 mL/min/1.73 m2 (P < 0.001). Limitations Lack of gold-standard assessment of kidney function. Conclusions In patients with HFrEF, discrepancies between eGFRcysC and eGFRScr are common and are associated with clinical outcomes, HRQoL, and frailty. The decline in kidney function associated with WHF is more marked when assessed with eGFRcysC than with eGFRScr. Plain-Language Summary Kidney function assessment traditionally relies on serum creatinine (Scr) to establish an estimated glomerular filtration rate (eGFR). However, this has been challenged with the introduction of an alternative marker, cystatin C (cysC). Muscle mass and nutritional status have differential effects on eGFR based on cysC (eGFRcysC) and Scr (eGFRScr). Among ambulatory patients with heart failure enrolled in PARADIGM-HF, we investigated the clinical significance of the difference between eGFRcysC and eGFRScr. More negative values (ie, eGFRScr > eGFRcysC) were associated with worse clinical outcomes (including mortality), poor quality of life, and frailty. In patients with progressive heart failure, which is characterized by muscle loss and poor nutritional status, the decline in kidney function was more pronounced when eGFR was estimated using cysC rather than Scr. The clinical implications of the discrepancy between cystatin C (cysC)– and serum creatinine (Scr)–estimated glomerular filtration rate (eGFR) in patients with heart failure (HF) and reduced ejection fraction (HFrEF) are unknown. Post-hoc analysis of randomized trial data. 1,970 patients with HFrEF enrolled in PARADIGM-HF with available baseline cysC and Scr measurements. Intraindividual differences between eGFR based on cysC (eGFRcysC) and Scr (eGFRScr; eGFRdiffcysC-Scr). Clinical outcomes included the PARADIGM-HF primary end point (composite of cardiovascular [CV] mortality or HF hospitalization), CV mortality, all-cause mortality, and worsening kidney function. We also examined poor health-related quality of life (HRQoL), frailty, and worsening HF (WHF), defined as HF hospitalization, emergency department visit, or outpatient intensification of therapy between baseline and 8-month follow-up. Fine-Gray subdistribution hazard models and Cox proportional hazards models were used to regress clinical outcomes on baseline eGFRdiffcysC-Scr. Logistic regression was used to investigate the association of baseline eGFRdiffcysC-Scr with poor HRQoL and frailty. Linear regression models were used to assess the association of WHF with eGFRcysC, eGFRScr, and eGFRdiffcysC-Scr at 8-month follow-up. Baseline eGFRdiffcysC-Scr was higher than +10 and lower than −10 mL/min/1.73 m2 in 13.0% and 35.7% of patients, respectively. More negative values of eGFRdiffcysC-Scr were associated with worse outcomes ([sub]hazard ratio per standard deviation: PARADIGM-HF primary end point, 1.18; P = 0.008; CV mortality, 1.34; P = 0.001; all-cause mortality, 1.39; P < 0.001; worsening kidney function, 1.31; P = 0.05). For a 1–standard-deviation decrease in eGFRdiffcysC-Scr, the prevalences of poor HRQoL and frailty increased by 29% and 17%, respectively (P ≤ 0.008). WHF was associated with a more pronounced decrease in eGFRcysC than in eGFRScr, resulting in a change in 8-month eGFRdiffcysC-Scr of −4.67 mL/min/1.73 m2 (P < 0.001). Lack of gold-standard assessment of kidney function. In patients with HFrEF, discrepancies between eGFRcysC and eGFRScr are common and are associated with clinical outcomes, HRQoL, and frailty. The decline in kidney function associated with WHF is more marked when assessed with eGFRcysC than with eGFRScr.