REM sleep behaviour disorder: How useful is it for the differential diagnosis of parkinsonism?

• REM sleep behaviour disorder is more frequent among synucleinopathies.
• pRBD is relatively frequent in PSP, a tauopathy, which may indicate that its’ occurrence better correlates with the topography and extent of degeneration than with the molecular pathology.
• pRBD has good positive predictive value for synucleinopathies.
• Diagnostic performance of pRBD for synucleinopathies measured in terms of sensitivity and specificity was disappointing.

BackgroundREM sleep behaviour disorder (RBD) is typically linked to synucleinopathies (SP). In this study we analyzed the utility and performance of RBD as a tool for the differential diagnosis of the most common forms of degenerative parkinsonism, including SPs and tauopathies.MethodsPatients with a syndromic diagnosis of degenerative parkinsonism matched for gender, age, and disease stage were assessed using a structured protocol with demographic and clinical data, including the diagnosis of probable RBD (pRBD), ascertained clinically using established criteria.ResultsOne hundred cases of Parkinson's disease (PD), 87 with progressive supranuclear palsy (PSP), 72 with the parkinsonian form of multiple system atrophy (MSA), 50 with dementia with Lewy bodies (DLB), and 18 with corticobasal degeneration (CBD) were included. pRBD was found in 58 (58%) of the PD patients, 59 (81.9%) of those with MSA, 37 (74%) with DLB, 32 (36.7%) with PSP, and one (5.5%) with CBD. Among the SPs, pRBD was significantly more common in MSA when compared with PD patients. Differences were also significant individually for all SPs when compared to PSP. The positive predictive value (PPV) of pRBD for a SP was 82.3%, but sensitivity was 69.4% and specificity 68.6%.ConclusionsIn our sample, pRBD was more frequent in SPs than in PSP and CBD, however, its’ frequency in PSP was significant. Although pRBD had a good PPV for a SP, all other measurements used for determine diagnostic performance were disappointing.