Pre-Diagnostic Factors Associated with Coding-Based Incident Multiple System Atrophy: A Nested Case-Control Study of the UK Biobank.

BACKGROUND

Multiple system atrophy (MSA) is a rare neurodegenerative synucleinopathy whose epidemiological risk factors remain poorly characterized.

OBJECTIVES

To characterize pre-diagnostic features and identify candidate risk factors for coding-based incident MSA in the UK Biobank (UKB), compared with Parkinson's disease (PD) and healthy controls.

METHODS

Iterative logistic regression models were used to derive risk factors from sociodemographic, lifestyle, environmental, and medical factors, adjusting for baseline age and sex, with Benjamini-Hochberg correction in the fully adjusted model. Candidate risk factors were evaluated in multivariate models at baseline and using 5- and 10-year latency windows to assess reverse causation.

RESULTS

We identified 114 coding-based incident MSA diagnoses, 1139 age- and sex-matched incident PD cases, and 1075 matched controls. We identified several baseline associations that attenuated with latency windows, including antihistamine use, alcohol use, and reduced physical activity, suggesting reverse causation. Statin use was associated with reduced risk at baseline (odds ratio [OR] 0.34, 95% confidence interval [95% CI] 0.17-0.69) and 5 years' latency but attenuated at 10 years, and was discordant with findings of reduced risk with moderately elevated cholesterol (OR 0.41, 95% CI 0.24-0.72). Rural living was the only candidate exposure whose effect strengthened with longer latency windows (5-year OR 2.73, 95% CI 1.32-5.65).

CONCLUSIONS

Coding-based incident MSA in the UKB exhibits a long pre-diagnostic phase consistent with autonomic and lifestyle changes several years prior to the diagnosis of an autonomic or parkinsonian disorder. Rural living may be an environmental signal warranting further investigation. © 2026 International Parkinson and Movement Disorder Society.