Accuracy of clinical diagnosis in parkinsonism--a prospective study.

Clinical diagnosis of Parkinson's syndrome (PS) is reasonably easy in most cases but the distinction between different variants of PS may be difficult in early cases. The correct diagnosis is not only important for counselling and management of patients but also in conducting pharmacological and epidemiological studies. There is very little critical literature on the pathological verification of the clinical diagnosis in PS. We report our 22 years experience to address that issue. Between 1968 and 1990, 65 PS patients came to autopsy. Complete data are available in 59 (M-50, F-19) cases. The initial diagnosis made by a qualified neurologist was idiopathic Parkinson's disease (IPD) in 43 cases. Of those 28 (65%) had Lewy body pathology. After a mean duration of 12 years the final diagnosis was IPD in 41 cases which was confirmed in 31 (76%). The IPD could not be clinically distinguished from cases with severe substantia nigra neuronal loss without inclusions or from those with neurofibrillary tangle inclusions and neuronal loss at the anatomical sites typically involved in IPD. All progressive supra-nuclear palsy, olivopontocerebellar atrophy, Jakob-Creutzfeldt's disease and the majority of the multiple system atrophy cases were diagnosed correctly during life. The correct clinical diagnosis in most non-IPD variants of PS was possible within 5 years of onset (range: 2 months to 18 years). We recommend that studies aimed at including only the IPD cases restrict the enrollment to those cases that have had PS motor manifestations for five years or longer duration.