Effect of multidomain interventions on estimated dementia and cardiovascular risk reduction: An individual‐participant data meta‐analysis from FINGER, MAPT, and pre‐DIVA

Abstract

AbstractBackgroundMultidomain interventions based on management of vascular and lifestyle risk factors could potentially reduce the risk of dementia and cardiovascular disease. Given the slow progressive nature of these conditions, dementia in particular, clinical endpoints may not be achievable within trials’ time‐frame. Well‐established risk scores could be useful surrogate outcomes. In this individual‐participant data meta‐analysis, we assessed the effect of three multidomain dementia prevention interventions on the CAIDE and FINRISK risk scores for dementia and cardiovascular disease, and the impact of baseline characteristics on the response to such interventions.MethodBaseline and 24‐month data from the FINGER, MAPT, and preDIVA dementia/cognitive decline prevention trials carried out in three European cohorts of older adults, were pooled at individual‐participant level (N=6465). Linear mixed‐model with maximum likelihood estimation and random effect by trial and study‐site was used to estimate changes in CAIDE and FINRISK risk scores in the intervention group vs control. Subgroup‐treatment effect interactions were tested to assess the influence of sex, baseline age, education, MMSE, and dementia or cardiovascular disease risk on the response to the interventions. The effect of the interventions on individual modifiable risk factors relevant for both risk scores was also investigated.ResultThe interventions led to a significant reduction of both CAIDE (estimated effect: ‐0.173; 95% CI: ‐0.268 to ‐0.077; P<0.001) and FINRISK (estimated effect: ‐0.108; 95% CI: ‐0.181 to ‐0.035; P=0.001) scores. A significantly larger reduction of CAIDE score was reported in women compared to men (P value for interaction=0.007) and a similar, but not significant, pattern was reported for FINRISK score. Interactions by other baseline characteristics were also not significant. Significant reduction of blood pressure (P≤0.001), and waist circumference (P=0.027), but not of BMI nor serum cholesterol, was observed. No baseline characteristic affected the response on these outcomes.ConclusionMultidomain interventions had beneficial effects on the estimated risk of dementia and cardiovascular disease, measured by the CAIDE and FINRISK scores. Women benefited more than men in terms of dementia risk, whereas effects of baseline age, education, cognitive status, and risk of dementia or cardiovascular disease on the response to the interventions could not be demonstrated.