Objective: To estimate mortality by drug use in a cohort of patients with PD relative to age and sex-matched comparators. Methods: Two longitudinal cohorts of patients with 7 and 11 years' duration of PD were constructed with matched comparators in Tayside, Scotland. Subjects were eligible for inclusion if they received a first prescription for an anti-Parkinson's drug from July 1989 to December 1995, with no PD drug prescription in the previous 6 months. Those who had previously taken a neuroleptic drug or were younger than 40 years of age were excluded. Results: Overall, subjects with PD in relation to comparators had higher mortality with a rate ratio (RR) of 1.76 (95% CI 1.11, 2.81) in the 7-year cohort. There was significantly greater mortality in patients with PD who received levodopa monotherapy (RR = 2.45, 95% CI 1.42, 4.23) relative to the comparators, adjusting for previous cardiovascular drug use and diabetes. However, there was no significant difference in mortality in those with PD receiving combination therapy of selegiline with levodopa and other drugs in relation to the comparators (RR = 0.92, 95% CI 0.37, 2.31). Conclusions: Subjects with PD had twice the rate of mortality relative to age- and sex-matched comparators. However, those subjects who received selegiline at any time in combination with co-careldopa or co-beneldopa showed no significant difference in mortality compared with the comparators. Monotherapy with levodopa was associated with the highest mortality.