Abstract
Creatinine-based estimated glomerular filtration rate (eGFRcr) may not fully capture mortality-related vulnerability in adults with dysglycemia. We examined whether cystatin C-based eGFR (eGFRcys) discordance and category reclassification were associated with mortality in 2 population-based cohorts analyzed separately. Adults with dysglycemia from National Health and Nutrition Examination Survey (NHANES) 1999 to 2002 and China Health and Retirement Longitudinal Study (CHARLS) 2011 were followed for mortality through 2019 and wave 4/2018, respectively. eGFRdiff was defined as eGFRcys minus eGFRcr. Occult kidney dysfunction was defined as eGFRcr ≥60 and eGFRcys 2
. Reclassification to a worse eGFR category was defined as assignment to a lower category by eGFRcys than by eGFRcr. The cohorts were analyzed separately using multivariable Cox models; NHANES analyses incorporated the cystatin C surplus-sample survey design. NHANES included 1531 participants and 981 deaths through 2019; CHARLS included 5231 participants and 579 deaths through 2018. Occult kidney dysfunction was present in 12.1% and 17.9%, respectively, and reclassification to a worse eGFR category occurred in 30.4% and 56.1%. In fully adjusted models, each 10 mL/min/1.73 m
2
lower eGFRdiff was associated with all-cause mortality in NHANES (hazard ratio [HR], 1.24; 95% confidence interval [CI], 1.14–1.35) and CHARLS (HR, 1.15; 95% CI, 1.08–1.23). Occult kidney dysfunction was also associated with mortality in NHANES (HR, 1.58; 95% CI, 1.25–1.99) and CHARLS (HR, 1.27; 95% CI, 1.01–1.60). NHANES heart disease mortality showed concordant associations for eGFRdiff and occult kidney dysfunction. Some categorical CHARLS estimates were attenuated in sensitivity analyses, and direct comparisons indicated limited incremental information from discordance measures beyond eGFRcys alone. Cystatin C-based eGFR discordance was associated with mortality among adults with dysglycemia. It may identify higher-risk subgroups when eGFRcr appears preserved, although its incremental prognostic information beyond eGFRcys alone was limited.