Introduction: This study sought to determine whether 10 years of assignment to intensive lifestyle intervention (ILI) relative to diabetes support and education leads to better cognition. We examine intervention effects overall and among clinical subgroups, and report correlations between computer-administered and interviewer-administered cognitive batteries.
Methods: The Action for Health in Diabetes (Look AHEAD) was a 16-site randomized controlled trial with overweight/obese individuals (aged 45-76) who had type 2 diabetes. The NIH Toolbox Cognition Battery tests developed to measure cognition across the lifespan were used to evaluate cognition. Results were compared with standard paper-and-pencil tests. The Toolbox and paper-and-pencil tests were administered an average of 10.9 years after randomization to 1002 participants.
Results: Toolbox measures significantly correlated with interviewer-administered measures, with the strongest correlations between the Toolbox Fluid Cognition Composite and Trails B (r = -0.64, P < .0001) and Digit Symbol Coding (r = 0.63, P < .0001), and between the Toolbox Dimensional Change Card Sort (r = 0.55, P < .0001) and the Digit Symbol Coding test. Overall, ILI and diabetes support and education groups had similar adjusted mean cognitive outcomes (P > .05 for all). Subgroup analyses identified different intervention effects within baseline body mass index groups for Picture Sequence Memory (P = .01), within baseline cardiovascular disease groups for Picture Vocabulary (P = .01) and Fluid Cognition Composite (P = .02) measures, and within baseline age groups for Picture Vocabulary (P = .02).
Discussion: Correlations between Toolbox and interviewer-administered outcomes provide a measure of internal validity. Findings suggest no overall effect of the intervention on cognition and that an ILI resulting in weight loss may have negative implications for cognition in individuals aged ≥60, with previous history of cardiovascular disease, and those with body mass index ≥40.
Keywords: Aged; Body mass index; Cognition; Diabetes mellitus; Neuropsychological tests; Obesity; Randomized controlled trial; Weight loss.