Commentary

The study by Lees et al shows that initiating treatment with bromocriptine or levodopa does not affect mortality rates and that patients who began treatment with levodopa had slightly better motor scores during follow up but a higher incidence of dyskinesias. The study has methodological limitations, the main one being that outcome assessment was not blinded to treatment allocation. Potential bias is suggested by the much higher number of deaths from cerebrovascular disease in the bromocriptine group than in the levodopa group. Dyskinesia scores are rather subjective; the distinction between dystonia and dyskinesias is sometimes difficult.

This study describes a larger number of PD patients and has a longer follow up period than do other studies that compare levodopa and dopamine agonists. These results are in keeping with previous open label studies on bromocriptine¹ and with more recent controlled studies of new dopamine agonists.^2–4 The withdrawal rate in the dopamine agonist group in this study is higher than that reported in the controlled trials of cabergoline (16% at 3 y) and ropinirole (27% at 5 y). In this report, only 3% of patients receiving bromocriptine remained on initial treatment after 10 years.

Remarkably, no clinical differences other than potency have emerged among the various dopamine agonists. Therefore, the question arises on the possible generalisation of long term data on bromocriptine. A recently published meta-analysis acknowledged that determining how to start treatment would depend on the relative effect of improving motor disability compared with lessening motor complications.⁵ The present study raises further awareness and concern about the pros and cons of starting PD treatment with a dopamine agonist rather than with levodopa. The results are more pro-levodopa than those published in recent series, which may lead clinicians to consider low doses of levodopa as an alternative initial treatment in PD patients.