Importance: Accurate diagnosis of neurodegenerative movement disorders is challenging because of a lack of in vivo biomarkers, overlapping clinical features and a delay in the emergence of pathognomonic features.
Objective: To evaluate clinicopathological correlation, diagnostic accuracy, genetic association with pathology, and ancestry-related differences in a multi-ancestry brain bank cohort.
Design: Multicentre retrospective autopsy cohort study on donors enrolled between 1985 - 2024.
Setting: 11 academic brain banks in the UK, US and Australia.
Participants: Brain donors identified from participating brain banks with available brain tissue and a clinical diagnosis of Parkinson's disease, Parkinson's disease dementia, dementia with Lewy bodies, progressive supranuclear palsy, corticobasal syndrome, multiple system atrophy, or neurologically normal controls.
Exposure: Genetic variant carrier status and clinical diagnostic category.
Main outcome: Clinical diagnostic accuracy; Lewy body and Alzheimer's disease pathology burden; survival; association with genetic variants and genetically inferred ancestry.
Results: We studied 3,353 brain donors (1281 [38.2%] female, mean [SD] age at death, 76.8 [10.6] years). Misdiagnosis rates for movement disorders ranged approximately from 10%-20%. Clinical diagnoses of dementia with parkinsonism (PDD/DLB) were more strongly associated with Lewy body pathology than Parkinson's disease without dementia (OR = 1·96, 95% CI = 1·30 - 3·04, p = 7·2e-04). Lewy pathology was identified in 4% of neurologically normal controls. Alzheimer's disease co-pathology was present in 40% of cases with Lewy body disease. GBA1 variant carriers exhibited greater Lewy body burden compared with noncarriers (OR = 1·94, 95% CI = 1·24 - 3·03, p = 0·01) or LRRK2 carriers (OR = 7·44, 95% CI = 2·16 - 25·64, p = 0·01). Pathological diagnoses differed by ancestry, with South Asian donors more likely to have progressive supranuclear palsy pathology and Ashkenazi Jewish donors more likely to have Lewy body disease (p < 0.0001), independent of GBA1 and LRRK2 mutation status.
Conclusion and relevance: Our findings highlight the value of integrating genetic and pathological data to improve diagnostic accuracy. The high prevalence of Alzheimer's disease co-pathology and ancestry-related differences in pathology point to the need for biologically informed diagnostic tools. These results support the integration of genetically and pathologically stratified approaches, correlating pathology with in vivo biomarkers, for future therapeutic trials.
Funding: Medical Research Council, Global Parkinson's Genetic Program/Aligning Science Across Parkinson's.