The multidomain intervention trial for prevention of dementia among older adults with type 2 diabetes: a multi‐center, randomized, 18‐month controlled trial

Abstract Background The aim of the Japan‐Multidomain Intervention Trial for Prevention of Dementia in Older Adults with Diabetes (J‐MIND‐Diabetes) was to examine the efficacy of multi‐domain intervention in prevention of cognitive decline in older adults with type 2 diabetes and mild cognitive impairment (MCI). Method The J‐MIND‐Diabetes is an 18‐month, multi‐centered, randomized controlled trial. Outpatients with type 2 diabetes aged 70‐85 years with MCI were recruited and randomized into multi‐domain intervention (management of metabolic and vascular risk factors, physical exercise, nutritional guidance, and promotion of social participation) and control groups (general instructions on dementia prevention) at a 1:1 allocation ratio using the stratified permuted‐block randomization methods. The primary outcome was a change from baseline to 18‐month follow‐up in a global composite score combining several neuropsychological tests, which include the MoCA‐J, the MMSE, the Rey‐Osterrieth Complex Figure Test (ROCFT), delayed recall test of a 10‐word list, the Digit Span, the Trail Making Test, the Digit Symbol Substitution Test, and the Letter Word Fluency Test. Result Between March 13, 2019, and May 8, 2020, 361 patients were screened and 154 were randomly assigned to the multi‐domain intervention group (n = 81) or control group (n = 73). 110 (71%) participants completed the trial. The full analysis set included 136 patients. In the primary efficacy analysis, there was no significant differences in the change of composite score from baseline to 18‐month follow‐up between the intervention group and the control group (mean difference 0.066 [95% CI ‐0.089 to 0.222]). In the analysis for each neuropsychological test, scores of immediate and delayed recall tests of the ROCFT in the intervention group significantly improved by a mean of 1.627 points and 1.793 points, respectively. The mean difference in the change of score of delayed recall test of the ROCFT between intervention group and control group was significant (mean difference 1.803 [95% CI 0.169 to 3.436]). In subgroup analyses, intervention effect on composite score was significant among patients with moderate to severe white matter hyperintensities (mean difference 0.351 [95% CI 0.088 to 0.614]). Conclusion The multi‐domain intervention could improve memory function in older adults with type 2 diabetes and MCI.