Authors’ Reply

• chronic kidney disease

Gupta and Burns¹ propose the null association between baseline eGFR <60 ml/min per 1.73 m² and risk for dementia in SPRINT may be attributable to the diminished predictive power of a single eGFR measurement in a population at high vascular risk.

Due to the bidirectional relation between kidney disease and vascular disease, disentangling their respective contributions to the development of cognitive impairment is challenging in observational studies. A strength of SPRINT, in this regard, is that two major vascular risk factors that confound prior studies of kidney disease and cognition, diabetes and hypertension, are limited by the study’s exclusion criteria and controlled by the intervention, respectively.

The SPRINT analysis also differed from two of the prior studies they cite in its use of adjudicated dementia and mild cognitive impairment (MCI) as the primary outcomes. The Health, Aging and Body Composition Study and the INVADE study each used a surrogate for MCI, change in a single cognitive test score, as the primary outcome.¹ Although elevated vascular risk was an inclusion criteria in SPRINT, it should be noted that the incidence of dementia in SPRINT, 7.2–8.6 cases per 1000 person-years, was lower than that observed in prior cohort studies, such as the Cardiovascular Health Study (26 cases per 1000 person-years)² and the 3C Study (12 per 1000 person-years).³ This may be explained by the younger average age and exclusion of individuals with diabetes in SPRINT. Thus, compared with prior observational studies that have reported an association between eGFR and cognitive impairment, SPRINT differed both in its ability to control for potential confounders, and in the frequency and stringency of the outcome. Both of these factors may have contributed to the observation of a null association.

As indicated in the supplemental material, although we did not observe a significant association between eGFR <60 ml/min per 1.73 m² and risk for dementia or MCI in SPRINT, there was a significant association between baseline eGFR <45 ml/min per 1.73 m² and risk for MCI (hazard ratio, 1.37; 95% CI, 1.04 to 1.81), and the composite outcome (hazard ratio, 1.34; 95% CI, 1.06 to 1.69). We hypothesize, similar to prediction of kidney failure risk,⁴ trajectory of eGFR and current eGFR may add complementary information to prediction of dementia risk.

• Copyright © 2020 by the American Society of Nephrology