Among patients with heart failure (HF) and reduced ejection fraction (HFrEF), assessment of kidney function is commonly performed using serum creatinine-based estimates. However, the high prevalence of sarcopenia and malnutrition and reduced physical activity in that patient population may limit the accuracy of serum creatinine as a marker for glomerular filtration.
Cystatin C, an endogenous protease inhibitor, is less influenced by muscle mass, diet, and demographic characteristics, including race, allowing for a more accurate estimation of kidney function. When factors unrelated to kidney function are present, there may be discrepancies between cystatin C-based estimated glomerular filtration rate (eGFRcysC) and serum creatinine-based eGFR (eGFRScr).
According to Alberto Pinsino, MD, and colleagues, there are few data available of the discrepancy between eGFRcysC and eGFRScr in patients with HFrEF. The researchers conducted a post-hoc analysis of data from the PARADIGM-HF (Prospective Comparison of Angiotensin Receptor-Neprilysin Inhibitor With Angiotensin-Converting Enzyme Inhibitor to Determine Impact on Global Mortality and Morbidity in Heart Failure) trial to examine that discrepancy. Results were reported in the American Journal of Kidney Diseases.
The analysis included data on 1970 patients with HFrEF enrolled in PARADIGM-HF who had available baseline measurements of cystatin C and serum creatinine. The clinical outcomes of interest were the PARADIGM-HF primary end point (composite of cardiovascular mortality or hospitalization for HF), cardiovascular mortality, all-cause mortality, and decline in kidney function. Secondary outcomes included poor health-related quality of life (HRQoL), frailty, and worsening HF, defined as HF hospitalization, visit to the emergency department, or outpatient intensification of therapy between baseline and 8-month follow-up.
Clinical outcomes on baseline difference between eGFRcysC and eGFRScr (eGFRdiffcysC-Scr) were regressed using Fine-Gray subdistribution hazard models and Cox proportional hazards models. The association of baseline eGFRdiffcysC-Scr with poor HRQoL and frailty was examined using logistic regression. The association of worsening HF with eGFRcysC, eGFRScr, and eGFRdiffcysC-Scr at 8-month follow-up was assessed using linear regression models.
Mean eGFRdiffcysC-Scr at baseline and at 8-month follow-up were –6 mL/min/1.73 m2 and –5 mL/min/1.73 m2, respectively. At baseline, 35.7% of patients (n=703) had an eGFRdiffcysC-Scr lower than –10 mL/min/1.73 m2 (negative group), 51.3% (n=1011) had an eGFRdiffcysC-Scr between –10 and +10 mL/min/1.73 m2 (midrange group), and 13.0% (n=256) had an eGFRdiffcysC-Scr higher than +10 mL/min/1.72 m2 (positive group).
Compared with the midrange and positive groups, those in the negative group were older, more likely to be White, be smokers, and receive diuretic agents. They also had higher New York Heart Association (NYHA) class, heart rate, N-terminal pro-brain natriuretic peptide levels, urinary albumin-creatinine ratio, and prevalence of atrial fibrillation. They were more likely to receive beta-blockers and mineralocorticoid receptor antagonists. There were independent associations between a baseline eGFRdiffcysC-Scr lower than –10 mL/min/1.73 m2 and older age, White race, current smoking, NYHA class III/IV, and higher heart rate.
The PARADIGM-HF primary end point, CV mortality, all-cause mortality, and decline in kidney function, occurred, in 22.3% (n=439), 11.8% (n=233), 16.9% (n=333), and 2.6% (n=51) of patients, respectively.
In both crude and adjusted models, there were associations between more negative values of baseline eGFRdiffcysC-Scr and increased risk for the PARADIGM-HF primary end point (hazard ratio [HR] per 1-standard deviation [SD] decrease, 1.18; P=.008); CV mortality (HR per 1-SD decrease, 1.34; P=.001); and all-cause mortality (HR per 1-SD decrease, 1.39; P<.001). When baseline eGFRdiffcysC-Scr was modeled as a categorical variable, with the negative group consistently associated with higher events compared with the midrange and positive groups, the associations persisted.
Baseline eGFRdiffcysC-Scr was also associated with decline in kidney function (HR per 1-SD decrease, 1.79; P<.001). Following multiple adjustments, the association was attenuated (HR per 1-SD decrease, 1.31; P=.05). When baseline eGFRdiffcysC-Scr was modeled as a categorical variable, the risk for worsening kidney function was higher in the negative group than in the midrange and positive groups; however, statistical significance was lost in adjusted models.
Of the patients assessed at baseline, 15.5% (n=301) met the definition of poor HRQoL, and 67.0% (n=1301) were classified as frail.
There were associations between more negative values of baseline eGFRdiffcysC-Scr and higher prevalence of poor HRQoL (adjusted odds ratio [aOR] per 1-SD decrease, 1.29; P=.001) and frailty (aOR per 1-SD decrease, 1.17; P=.008). When baseline eGFRdiffcysC-Scr was modeled as a categorical variable, the prevalences of poor HRQoL and frailty were significantly higher in the negative group than in the positive group.
Worsening HF was detected in 5.7% (n=93) of 1631 patients who had samples available at baseline and at 8-month follow-up. There was an association between worsening HF and a decline in kidney function at 8 months when assessed using serum creatinine or cystatin C. However, changes in kidney function at 8-month follow-up were more pronounced with considering eGFRcysC (adjusted between-group difference, –8.08 mL/min/1.73 m2; 95% CI, –10.50 to –5.66; P<.001) rather than eGFRScr (adjusted between-group difference, –3.69 mL/min/1.73 m2; 95% CI, –6.06 to –1.32; P=.002). In adjusted models that included eGFRdiffcysC-Scr at baseline, there was an association between worsening HF and a change in 8-montgh eGFRdiffcysC-Scr of –4.67 mL/min/1.73 m2 (95% CI, –6.94 to –2.41; P<.001).
The researchers cited some limitations to the study, including the lack of a gold-standard measure for kidney function, and the small number of patients of non-White races and NYHA class IV.
In summary, the authors said, “In a cohort of ambulatory patients with HFrEF, more negative values of eGFRdiffcysC-Scr were associated with worse clinical outcomes, poor HRQoL, and frailty. The decline in kidney function associated with worsening HF was more marked when assessed with eGFRcysC than with eGFRScr, resulting in more negative values of eGFRdiffcysC-Scr. Our findings challenge the current paradigm of kidney function assessment in the HF population, which relies heavily on serum creatinine, and call for a more widespread incorporation of cystatin C and of measures of discrepancy between those two markers in clinical practice and in future research.”
Takeaway Points
- Researchers conducted a post-hoc analysis of data from a randomized trial to examine the discrepancy between estimated glomerular filtration rate based on serum creatinine (eGFRScr) and eGFR based on cystatin C (eGFRcysC) in patients with heart failure and reduced ejection fraction.
- Discrepancies between eGFRcysC and eGFRScr were common and more negative values of differences between the two equations were associated with worse clinical outcomes.
- More negative values of the differences between eGFRcysC and eGFRScr were associated with poor health-related quality of life and frailty.
© 2025 Mashup Media, LLC, a Formedics Property. All Rights Reserved.