Epidemiological survey of dementia in parkinsonism and control population.

In a 13-year (1967-1979) survey of Rochester, Minnesota residents, 138 new cases of parkinsonism were recognized. For each case, 2 age- and sex-matched residents from the community were identified who attended the same medical facility as the patient did. Original medical records for 40 years preceding the diagnosis and control selection were reviewed on all patients and controls. Diagnosis of dementia was made three times more frequently in patients than in the controls, but the age at diagnosis was similar in the two groups. Those cases where neither of the three matched individuals had dementia at the time of initial recognition of parkinsonism were further followed. Dementia emerged significantly more frequently in patients than in controls. The subgroup of patients treated with levodopa had higher frequency of dementia than the controls. It is concluded that dementia is significantly more common in parkinsonian patients than in a comparable population. Dementia is believed to be either an integral part of the disease process or these patients are prone to some other illness that produces dementia. Our data conclusively show that dementia is not a reflection of old age. Higher incidence of dementia in levodopa-treated cases may be due to case-selection artifact or an as yet unrecognized effect of levodopa.