Authors' Reply: Interpreting Rapid eGFR Decline in the Generation 100 Study

We appreciate the feedback on our JASN article “Long-Term Physical Exercise for Preventing CKD in Older Adults: A Randomized Clinical Trial,”1 which has provided us the opportunity to clarify our approach and address the concerns raised by Cabrera and Segovia.2 As noted, it seems somewhat surprising that 30% experienced eGFRcys decline >5 ml/min per year in the control arm of our trial, which comprised relatively healthy community-living older adults with mean age of 72 years and eGFRcys of 96 ml/min per 1.73 m2. In contrast, among community-based elders evaluated in the cardiovascular health study (mean age of 72; mean eGFRcys 79 ml/min per 1.73 m2), the upper quartile of eGFRcys decline was >3 ml/min per year despite a greater prevalence of comorbidities.3 In our opinion, this finding reflects the absence of studies with measured GFR in older adults without established CKD, potentially leading to imperfect calibration of eGFRcys equations in the higher GFR range. However, the mean eGFRcys changes were −2.2 (4.1) in our trial versus −1.8 (2.6) in cardiovascular health study, suggesting that actual annual changes in blood cysC concentrations and true GFR loss were likely very similar. Measurement error is unlikely to be a major contributor because we demonstrated a coefficient of variation of 1.8% for our cysC method. More importantly, any bias in eGFR would apply to all arms of the trial and would not impact our primary results. We chose the binary eGFR decline definition as our primary outcome because of its greater clinical importance; but we also reported on eGFR slope as a continuous secondary outcome. Compared with the control group, the high-intensity interval training group had a 0.3 ml/min per year better slope (95% confidence interval [CI], −0.06 to 0.6). When evaluating the observed changes in physical activity, the group that increased their activity had a 0.8 ml/min per year better slope (95% CI, 0.04 to 1.6). These continuous outcomes are consistent with our primary finding of 25% lower risk of rapid eGFR decline in the high-intensity interval training group compared with the control group (multiadjusted relative risk 0.75 with 95% CI, 0.59 to 0.95). Thank you for raising these important points. We conclude that physical activity slowed the decline of kidney function using both dichotomous and continuous outcomes, and we believe our findings are robust.