Association of the ATN Research Framework With Clinical Profile, Cognitive Decline, and Mortality in Patients With Dementia With Lewy Bodies.

Abstract

The ATN framework has been developed to categorize biological processes within the Alzheimer disease (AD) continuum. Because AD pathology often coincides with dementia with Lewy bodies (DLB), we aimed to investigate the distribution of ATN profiles in DLB and associate ATN profiles in DLB with prognosis. We included 202 patients with DLB from the Amsterdam Dementia Cohort (68±7 years; 19% female; Mini-Mental State Examination 24 ± 3; abnormal DAT-SPECT 105/119). Patients were classified into 8 profiles according to the ATN framework, using CSF β-amyloid (Aβ)42 (A), CSF p-tau (T), and medial temporal atrophy scores (N). We compared presence of clinical symptoms in ATN profiles and used linear mixed models to analyze decline on cognitive tests (follow-up 3±2 years for n = 139). Mortality risk was assessed using Cox proportional hazards analysis. Analyses were performed on the 8 profiles as well as 3 clustered categories (normal AD biomarkers, non-AD pathologic change, and AD continuum). Fifty (25%) patients with DLB had normal AD biomarkers (A-T-N-), 37 (18%) had non-AD pathologic change (A-T+N- 10%, A-T-N+ 6%, A-T+N+ 3%), and 115 (57%) were classified within the AD continuum (A+T-N- 20%, A+T+N- 16%, A+T-N+ 10%, A+T+N+ 9%). A+T+N+ patients were older and least often had REM sleep behavior disorder symptoms. Parkinsonism was more often present in A+T- compared to A-T+ patients (independent of N). Compared to patients with normal AD biomarkers, patients in A+ categories showed steeper decline on memory tests and higher mortality risk. Cognitive decline and mortality did not differ between non-AD pathologic change and normal AD biomarkers. In our DLB cohort, we found clinically relevant associations between ATN categories and disease manifestation. Patients within the AD continuum had steeper cognitive decline and shorter survival. Implementing the ATN framework within patients with DLB aids in subtyping patients based on underlying biological processes and could provide targets for future treatment strategies (for example, AD-modifying treatment). Expanding the framework by incorporating markers for α-synucleinopathy would improve the use of the framework to characterize patients with dementia with mixed pathology, which could enhance proper stratification of patients for therapeutic trials.