Independent and interactive contributions of white matter hyperintensities and Alzheimer's disease imaging and plasma biomarkers to cognitive decline in older adults without dementia.

Accounting for common co-pathologies such as white matter hyperintensities (WMH) may improve the prognostic performance of imaging and blood-based biomarkers for cognitive decline in Alzheimer’s disease (AD). Plasma biomarkers including p-tau217, glial fibrillary acidic protein (GFAP), and neurofilament light chain (NfL) as well as imaging biomarkers including hippocampus volume, amyloid-β (Aβ) PET, and WMH have been reported to be associated with decline across multiple cognitive domains. However, a comprehensive comparison of these biomarkers in predicting domain-specific cognitive decline and their potential additive or interactive effects with WMH volume, is lacking. We analyzed data from 216 (median age [range]: 67.2 years [51–90]) community-dwelling older adults who underwent MRI, Aβ PET, and blood-based biomarker assessments. Participants were followed for a median of 5.3 years with annual clinical and neuropsychological evaluations. We created cognitive composite scores for episodic memory, executive functions, processing speed/attention, language, visuospatial abilities, and global cognition. Linear mixed-effects models were used to compare the predictive value of imaging and blood-based biomarkers individually and in interaction with WMH volume for decline across cognitive domains and clinical dementia rating-sum of boxes (CDR-SOB) progression. Multigroup path analysis was used to investigate multiple biomarker interactions simultaneously. Individual biomarkers significantly predict decline across multiple cognitive domains. WMHs independently predicted decline in processing speed/attention (β = -0.024, 95% CI -0.045 to -0.004, P = 0.022) and executive functions (β = -0.025, 95% CI -0.042 to -0.007, P = 0.005) and interacted with imaging and plasma biomarkers to accelerate cognitive decline across nearly all domains, with the strongest effect observed in episodic memory (β = -0.042, 95% CI -0.057 to -0.027, P < 0.001). Additionally, significant progression in the CDR-SOB was only observed in individuals with both elevated Aβ and WMH volume. Multigroup path analysis identified the interaction between Aβ PET and WMH volume as the most robust interaction associated with faster decline. Coexisting AD pathology and WMHs interacted to predict faster cognitive decline, most pronounced in episodic memory but also across other cognitive domains. Aβ PET was the main factor interacting with WMHs to predict episodic memory decline, but significant interactions were also observed for hippocampal volume, plasma p-tau217, and GFAP. Consideration of Aβ PET and blood-based biomarkers in the context of an individual’s WMH volume may improve the prognostic value of these biomarkers.