The present paper compares, in terms of mortality, two treatment regimens for Parkinson's disease (PD), i.e., bromocriptine later combined with levodopa versus levodopa only. Between 1982 and 1989, 60 PD patients (29 treated with levodopa alone [group D] and 31 receiving first bromocriptine followed by an association of bromocriptine + levodopa [group B/D]) were recruited. Data were updated in January 2000. Survival functions were estimated using Kaplan Meier product-limit method and comparison between the two groups with the log-rank test. Mortality was also compared with that of the general French population using standardized mortality ratios (SMRs). The mean duration of follow-up was 10.3 ± 3.0 years. Seventeen patients died during the follow-up: nine in the group B/D and eight in the group D. The probability of survival at 10 years was 79.0% [95% confidence interval [CI]: 71.4–86.6] in group B/D and 72.9% [95% CI: 63.3–82.6] in group D. In comparison with the general French population, SMRs were not statistically different from 1, in the whole sample of PD patients (1.21, 95 % CI [0.71–1.95]), in group D (0.98 [0.42–1.93]), or in group B/D (1.53 [0.70–2.92]). In this population, we were unable to find any favourable effect of an early use of bromocriptine on mortality in PD in comparison with levodopa alone. © 2001 Movement Disorder Society.