Public Health.

Background: International collaboration is crucial to the future of research in dementia with Lewy bodies (DLB). Although it is hoped that a single global DLB cohort can be created through combination of data from the many longitudinal studies conducted internationally, a high likelihood of phenotypic heterogeneity may prohibit harmonisation and analysis of datasets. Our primary objective is to determine whether significant heterogeneity is observed in the sociodemographic and clinical characteristics of people with DLB globally.

Method: Longitudinal DLB studies and multi-centre consortia were invited through existing professional networks to contribute limited descriptive clinical and sociodemographic data for their samples using a standardised proforma. One way analysis of variance (ANOVA) with summary statistics as input was used to compare mean scores. χ test was used to compare categorical variables.

Results: Descriptive data from 10 projects in seven countries, and one multinational consortium, were shared, comprising a pooled sample of 4157 individuals with DLB. The proportion of males (χ = 135, df = 7, p< .01) and the duration of formal education (F(3,6) = 19.41, p<.02) varied significantly between DLB projects. No significant difference in mean age (F(3,7) = 0.87, p = .61) or duration of DLB symptoms at baseline (F(2,6) = 1.46, p = 0.46) was noted, nor was it observed in scores of baseline Mini-Mental State Examination (MMSE)(F(2,5) = 0.73, p = 0.66), Clinical Dementia Rating (CDR) (F(1,4) = 2.04, p = .48), or Part III of the Unified Parkinson's Disease Rating Scale (UPDRS) (F(2,6) = 12.9, p = .73).

Conclusion: Despite a broad range of project methodologies and a diverse group of subjects, little sociodemographic or clinical heterogeneity was observed between longitudinal DLB projects. Interrogation of subject-level data, rather than descriptive data for each project, and more detailed clinical data on participants would allow further examination of heterogeneity. This could support harmonisation of individual DLB datasets to create a single global DLB cohort.