Abstract
BACKGROUND: Frozen shoulder may be an early preclinical symptom of Parkinson's disease (PD). OBJECTIVE: To examine PD risk after frozen shoulder diagnosis and to evaluate this disorder as a possible manifestation of parkinsonism preceding the clinical recognition of PD and possible target for screening. METHODS: Danish population-based medical registries were used to identify patients aged ≥40 years with a first-time frozen shoulder diagnosis (1995-2016). A comparison cohort was randomly selected from the general population matched on age and sex. To address detection bias and the specificity of frozen shoulder diagnosis, we performed a sensitivity analysis, using similar matching criteria to select a cohort of patients with back pain diagnosis. The outcome was incident PD. Cumulative incidences and adjusted hazard ratios (HRs) were estimated with 95% confidence intervals (CIs). RESULTS: We identified 37,041 individuals with frozen shoulder, 370,410 general population comparators, and 111,101 back pain comparators. The cumulative incidence of PD at 0-22 years follow-up was 1.51% in the frozen shoulder cohort, 1.03% in the general population cohort, and 1.32% in the back pain cohort. For frozen shoulder versus general population, adjusted HRs were 1.94 (CI: 1.20-3.13) at 0-1 years and 1.45 (CI: 1.24-1.70) at 0-22 years follow-up. For frozen shoulder versus back pain, adjusted HRs were 0.89 (CI: 0.54-1.46) and 1.01 (CI: 0.84-1.21), respectively. CONCLUSION: Patients with frozen shoulder had an increased PD risk compared with the general population, although the absolute risks were low. Frozen shoulder might sometimes represent early manifestations of PD. Detection bias probably cannot account for the increased PD risk during the long-term follow-up.