Herpes zoster and risk of incident Parkinson’s disease in US Veterans: a matched cohort study
Problem
The complex aetiology of Parkinson’s disease (PD) is poorly understood. While some acute, severe infections may increase future PD risk, the role of herpes zoster (HZ) is unclear. Previous studies have provided conflicting evidence on the relationship between HZ and PD. While two matched cohort studies using the Taiwanese National Health Insurance Database over similar time periods found up to 80% in increased PD risk after HZ, a nested case-control study using US Medicare claims data showed a decreased risk of 12% after HZ. We therefore aimed to replicate the approach used in the Taiwanese studies to investigate the relationship between HZ and incident PD in the largest integrated healthcare system in the US – the Department of Veterans Affairs.
Approach
We performed a matched cohort study using electronic health record data from the Veterans Aging Cohort Study-National (VACS-National). Patients with incident HZ aged ≥40 years from 2008 to 2018 were matched with up to five individuals without HZ by age, sex, race/ethnicity, geographic site, and calendar time. The primary outcome was a new diagnosis of PD defined by one outpatient or inpatient ICD-9/10 code. We used Cox regression to assess any association between HZ and PD, adjusting for demographic and clinical characteristics. In secondary analyses, we excluded outcomes that occurred within 6, 12, and 24 months of follow-up (to mitigate potential of reverse causality) and stratified the HZ group by whether they received antiviral therapy (AVT) within 7 days of HZ diagnosis.
Findings
198,099 patients with HZ were matched to 976,660 patients without HZ (median age 67.3 years (IQR 61.4-75.7), 94% male) and followed for a median duration of 4.2 years (IQR 1.9-6.6). Overall incidence of PD was 1.90 per 1000 person-years. In primary analysis, HZ was not associated with PD diagnosis (adjusted HR 0.95, 95% CI 0.90-1.01). Findings were similar when excluding outcomes within 6 months (HR 0.95, 95% CI 0.90-1.01), 12 months (HR 0.96, 95% CI 0.90-1.02) and 24 months of follow-up (HR 0.99, 95% CI 0.92-1.06). Risk of PD did not vary by receipt of AVT (HR 0.95, 95% CI 0.88-1.04 for AVT; HR 0.96, 95% CI 0.89-1.03 for no AVT).
Consequences
HZ was not associated with increased risk of PD in this large US Veteran cohort. This finding persisted throughout all secondary and sensitivity analyses. Differences in results between our study and previous studies could be due to methodological choices that were not well defined in previous studies, including how control groups were selected. Ideally, future studies would include longer follow-up to avoid reverse causation due to PD’s long prodromal phase.