Background: Migraine patients want acute treatment to provide complete relief of the migraine attack within 30 minutes. Traditionally, “speed of onset of effect” is evaluated by estimating the time-point for first statistical separation of drug and placebo. The estimated onset of effect can be a few percent difference of patients being pain free in very large randomised, controlled trials. This difference, however, can be clinically irrelevant.

Methods: Placebo-controlled randomised, controlled trials with pain freedom results from 30 min to 2–4 hours were retrieved from the literature. For each time-point, the therapeutic gain (drug minus placebo) (TG) was calculated. Therapeutic gain for being pain free of 5% was chosen for the definition of “onset of action”, since this is approximately 1/3 of the 16% TG and 1/4 of 21% of TG for sumatriptan 50 mg and 100 mg, respectively.

Results: A total of 22 time-effect curves based on randomised, controlled trials were analysed. Based on the “onset of action” of 5% pain freedom, the evaluated drugs and administration forms can be classified as follows: i) Early time to onset, ≤30 min (three randomised, controlled trials); ii) medium time to onset, 60 min (nine randomised, controlled trials); iii) delayed time to onset, 90–120 min (10 randomised, controlled trials).

Conclusion: Only three non-oral administration forms with a triptan (subcutaneous sumatriptan and nasal zolmitriptan) resulted in an “onset of action” at ≥30 min; in the future, early onset of action should be a priority in the development of new drugs or new administration-forms for the treatment of acute migraine attacks.