Background Several non-motor features may individually contribute to identify prodromal Parkinson’s disease (PD), but little is known on how they interact.

Methods We conducted a case–control study nested within the Health Professionals Follow-up Study in a large cohort of men age 40–75 at recruitment in 1986. Cases (n=120) had confirmed PD, were<85 in January 2012, returned a 2012 questionnaire with questions on probable rapid eye movement sleep behaviour disorder (RBD) and constipation sent to all cohort participants and completed in 2014 the Brief Smell Identification Test and a questionnaire assessing parkinsonism and other non-motor PD features (including depressive symptoms, excessive daytime sleepiness, impaired colour vision and body pain). Controls (n=6479) met the same criteria as cases, except for the PD diagnosis.

Results Concurrent constipation, probable RBD and hyposmia were present in 29.3% of cases and 1.1% of controls, yielding an age-adjusted OR of 160(95%CI 72.8to353) for three features versus none. The odds of PD increased exponentially with additional non-motor features (OR for 6–7 features versus none: 1325; 95%CI333to5279). Among men without PD, the number of non-motor features was associated with odds of parkinsonism (OR for 6–7 features versus none: 89; 95%CI21.2to375). We estimated that in a population with a prodromal PD prevalence of 2%, concurrent constipation, probable RBD and hyposmia would have a maximum sensitivity of 29% and a positive predictive value (PPV) of 35%. The PPV could increase up to 70% by including additional features, but with sharply decreased sensitivity.

Conclusions Concurrent constipation, probable RBD and hyposmia are strongly associated with PD. Because these features often precede motor symptoms and their co-occurrence could provide an efficient method for early PD identification.