BackgroundRecent estimates suggest that dementia incidence is decreasing in the US possibly due to better management of cardiovascular disease (CVD) risk factors, but these studies lack repeated cross-sectional assessment among a representative US sample. Our objective was to assess temporal trends in cognitive performance in relation to CVD risk factors among older National Health and Nutrition Examination Survey (NHANES) participants.MethodsWe used repeated cross-sectional assessment of 5711 participants ≥60 years of age from four NHANES cycles: 1999–2000, 2001–2002, 2011–2012 and 2013–2014. Cognitive function was assessed during each cycle with the Digit Symbol Substitution Test (DSST). We estimated mean DSST score at each cycle and annual trend in DSST before and after adjustment for age, sex, race/ethnicity, education, smoking status, blood pressure, glucose status and body mass index.ResultsDSST scores was significantly higher for 2011–2012 (difference: 6.7, 95% CI 4.4 to 9.0) and 2013–2014 (difference: 6.2, 95% CI 4.0 to 8.5), but not 2001–2002 (difference: 2.3, 95% CI −0.01 to 4.6) as compared with 1999–2000 before adjustment. We observed a linear trend for higher annual DSST score before adjustment (DSST/year: 0.44, 95% CI 0.31 to 0.57) and after adjustment for age, sex, race/ethnicity, educational attainment and CVD risk factors (DSST/year: 0.17, 95% CI 0.08 to 0.26). Educational attainment was most strongly associated with the attenuation in the trend in cognitive function (77% of trend attenuation and 20% of variance in DSST).ConclusionCognitive function is improving over time for US adults aged ≥60 years. These improvements are strongly associated with greater educational attainment and irrespective of the changing US demographic and cardiovascular health profiles.