Searching for responders to multidomain dementia prevention in late life: A pooled analysis of individual participant data from the MAPT and preDIVA trials

Abstract

INTRODUCTION

It is unknown in which, if any, subgroups of older adults multidomain interventions are effective at reducing long-term dementia incidence.

METHODS

We pooled up to 12 years of follow-up data from 5205 participants aged > 70 from the Multidomain Alzheimer Preventive Trial (MAPT) and Prevention of Dementia by Intensive Vascular Care (preDIVA) studies. The primary outcome was incident all-cause dementia. Pre-specified subgroups were defined by dementia risk factors (age, sex, education, apolipoprotein E [APOE] genotype, cognitive status, and cardiovascular risk factors).

RESULTS

Four hundred eighty-six participants developed dementia during 37,782 person-years of follow-up. Higher incidence was associated with baseline age, APOE ε4 genotype, physical inactivity, Mini-Mental State Examination, and blood pressure. Multidomain intervention had no effect on incident dementia overall (hazard ratio = 0.98, 95% confidence interval 0.80–1.21), or in any pre-specified subgroup. A recursive partitioning algorithm also did not detect any subgroups, defined by single or multiple risk factors, showing a differential intervention effect.

DISCUSSION

We did not identify any subgroups of older adults in whom multidomain interventions significantly reduced incident dementia.

CLINICAL TRIAL REGISTRATION

MAPT: NCT00672685 (clinicaltrials.gov); PreDIVA: ISRCTN29711771 (ISRCTN registry)

Highlights

• We pooled up to 12 years of follow-up data from two multidomain prevention trials.
• Five thousand two hundred five participants aged ≥ 70 were included.
• Subgroups were pre-defined by modifiable and non-modifiable dementia risk factors.
• A data-driven recursive partitioning algorithm was also used.
• Multidomain intervention did not lower incident dementia overall or in any subgroup.