DOI: 10.1002/alz.076468 ISSN: 1552-5260

Alzheimer’s disease genetic risk score and neuroimaging biomarkers in the FINGER multidomain lifestyle randomized controlled trial

Gazi Saadmaan Hossain, Maria Carolina Dalmasso, Alfredo Ramirez, Mikko Hiltunen, Nina Kemppainen, Jenni Lehtisalo, Tiia Ngandu, Juha O. Rinne, Hilkka Soininen, Ruth Stephen, Miia Kivipelto, Alina Solomon
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



A comprehensive genetic risk score (GRS) for Alzheimer’s disease (AD‐GRS) was recently developed based on 83 genome‐wide significant variants (Bellenguez et al. Nat Genet. 2022), excluding APOE. This study investigated AD‐GRS and APOE4 in relation to neuroimaging biomarkers in the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER).


FINGER included at‐risk individuals without dementia from the general population, aged 60‐77 years. Participants were randomized to either 2‐year multidomain lifestyle‐based intervention or regular health advice. At baseline, 132 participants underwent MRI scans, of which 48 also underwent PET scans (PiB/FDG). Two years later, 112 and 39 repeated their scans, respectively. MRI measures included hippocampus, total gray matter and white matter lesion volumes, and cortical thickness. Composite indices were calculated for PiB and FDG uptake values. We report preliminary results from linear regressions (standardized β‐coefficients, p‐values).


At baseline, higher AD‐GRS tended to relate to lower hippocampus volume (β = ‐0.15, p = 0.07) and FDG‐PET index (β = ‐0.23, p = 0.08); this remained after APOE adjustment. APOE4 was associated with lower hippocampus volume (β = ‐0.27, p = 0.001) and higher PiB‐PET index (β = 0.48, p = 0.001). Overall, AD‐GRS was not related to the 2‐year change in neuroimaging measures, while APOE4 was related to decreasing hippocampus (β = ‐0.27, p = 0.01) and total gray matter (β = ‐0.25, p = 0.01) volumes, and cortical thickness (β = ‐0.28, p = 0.003). There were no significant interactions of AD‐GRS or APOE4 with randomization group in relation to neuroimaging changes. However, in analyses stratified by AD‐GRS (below vs above median), there was less PIB index increase in intervention vs control in the higher AD‐GRS group (β = ‐0.60, p = 0.03); no intervention control‐difference was found in the lower AD‐GRS group (β = 0.04, p = 0.86). In analyses stratified by APOE4, PIB index tended to increase less in intervention vs control in non‐carriers (β = ‐0.38, p = 0.078), with no intervention‐control difference in carriers (β = 0.05, p = 0.89).


Unlike APOE4, AD‐GRS was not related to 2‐year change in AD‐related neuroimaging measures. AD‐GRS and APOE4 may also have very different impact on potential intervention effects on amyloid, i.e. less accumulation in the higher‐risk group (AD‐GRS) vs lower‐risk group (APOE). However, these exploratory findings need to be verified in larger studies.

More from our Archive