Abstract BACKGROUND AND AIMS Age-adapted thresholds of estimated glomerular filtration rate (eGFR) have been proposed for the diagnosis of chronic kidney disease (CKD) [1] to acknowledge decline in kidney function with increasing age [2]. Proponents of age-adapted thresholds suggest that subjects over the age of 65 with eGFR 45–59 mL/min/1.73 m2 and without albuminuria appear to be at low risk of adverse outcomes [1]; however, the risks associated with CKD extend beyond kidney failure [3]. We evaluated whether cystatin C testing (eGFRcys) could improve risk stratification in CKD without adopting age-adapted thresholds. METHOD Data were from participants in the UK Biobank cohort with eGFR based on serum creatinine (eGFRcr) ≥45 mL/min/1.73 m2 and without albuminuria. Participants with a history of cardiovascular disease or kidney failure at baseline were excluded. CKD status was initially categorized by eGFRcr, then concordance of eGFRcys < 60 mL/min/1.73 m2 was determined. Multivariable-adjusted Cox proportional hazards models evaluated associations of CKD status with death, myocardial infarction (MI), stroke and kidney failure. Analyses were stratified by age: older (65–73 years) and younger (<65 years) participants. RESULTS Among 76 629 older and 351 773 younger participants, followed over median 11.5 (IQR 10.8–12.2) years, there were 24 251 deaths, 6983 MI, 4081 strokes and 209 kidney failure events. Of 2016 older participants with eGFRcr 45–59 mL/min/1.73 m2 37% had eGFRcys > 60 mL/min/1.73 m2. Furthermore, eGFRcr alone did not detect 6278 (8%) of older participants with eGFRcys < 60 mL/min/1.73 m2. In cross-sectional analyses, the median baseline eGFR was lower at higher age for all equations: this was most pronounced for eGFRcys (Figure 1). CKD by eGFRcys only or with both markers experienced higher absolute incidence of MI, stroke and mortality compared to those without CKD by either marker (Figure 2). Compared to those without CKD, older participants with concordant eGFRcr and eGFRcys < 60mL/min/1.73 m2 had elevated adjusted risks of death (HR 1.8, 1.6–2.0), MI (HR 1.9, 1.5–2.4), stroke (HR 1.7, 1.2–2.3) and kidney failure (HR 4.2, 1.2–14.1). Findings were similar in younger participants with concordant eGFRcr and eGFRcys < 60 mL/min/1.73 m2 for death (HR 2.2, 1.9–2.6) and stroke (HR 2.1, 1.4–3.2), though not MI (HR 1.2, 0.8–1.8). When CKD was not confirmed by eGFRcys, eGFRcr 45–59 mL/min/1.73 m2 was not associated with increased hazards of death (HR 1.1, 0.9–1.4), MI (HR 1.1, 0.7–1.7) or stroke (HR 1.1, 0.7–1.9) in older participants, nor in younger participants for the same outcomes: death (HR 1.1, 0.8–1.3), MI (HR 1.2, 0.8–1.9) and stroke (HR 0.8, 0.4–1.6). CONCLUSION Among persons with CKD, cardiovascular disease and mortality are important risks that should be considered when determining CKD status. CKD categorization by eGFRcr alone includes a large proportion of individuals who have similar risks to persons without CKD. eGFRcys detects a substantial number of high-risk individuals not identified to have CKD by eGFRcr. In the absence of cystatin C testing, lower, age-adapted thresholds for diagnosis of CKD may inadequately detect the broader risks associated with kidney disease in older people.
Read full abstract