Accuracy of the clinical diagnosis of postencephalitic parkinsonism: A clinicopathologic study

The accuracy of the clinical diagnosis of postencephalitic parkinsonism (PEP) is unknown. We determined the validity of the clinical diagnosis of PEP by presenting 105 records with neuropathologic diagnoses of PEP (n = 7), progressive supranuclear palsy (n = 24), Parkinson's disease (n = 15), dementia with Lewy bodies (n = 14), multiple system atrophy (n = 16), corticobasal degeneration (n = 10), Creutzfeldt-Jakob disease (n = 4), and other dementia disorders (n = 15), as clinical vignettes to six neurologists unaware of the autopsy findings. The neurologists' own clinical diagnoses were compared with neuropathologic diagnoses for measures of diagnostic accuracy, including reliability (K statistics), sensitivity and positive predictive values for the first and last visits. The group reliability for the diagnosis of PEP was almost perfect (κ = 0.91, 0.9). The mean sensitivity at the first visit was 86% (range, 71-100%) with minimal change at the last visit (83%; range, 71-100%). Positive predictive values remained unchanged (100%). The high reliability, sensitivity and positive predictive values of the clinical diagnosis of PEP indicate that neurologists identify this disorder even when they report that they have never evaluated a case. In our data set, the best predictors for the diagnosis of PEP included onset below middle age; symptom duration lasting more than 10 years, and the presence of oculogyric crisis. History of encephalitis lethargica, present in most PEP cases, was an important individual diagnostic predictor.