We carried out a meta-analysis of 22 trials to determine the comparative placebo effect of (a) subcutaneous vs. oral and (b) in-hospital vs. at-home administration in the treatment of migraine. The headache relief rates were combined from the placebo arms of these randomised clinical trials assessing the value of sumatriptan in acute treatment of migraine. The main outcome measure was the proportion of patients reclassified from severe or moderate headache severity to no or mild headache severity 2 h after the beginning of treatment. In the oral regimen 222 of 865 patients (25.7%) reported no or mild headache severity after 2 h, compared to 279 of 862 patients (32.4%) of those receiving subcutaneous placebo (6.7% difference; 95% CI 2.4-11.0%). Adjusting for treatment setting and severity of headache at baseline did not change the observed difference. After placebo treatment at home 285 of 1,054 patients (27.0%) reported no or mild headache severity after 2 h, compared to 216 of 673 patients (32.1%) among those receiving placebo in hospital (5.1 % difference; 95% CI 0.6-9.5%). When adjusted for route of administration and severity of headache at baseline, the difference in relief rates between home and hospital setting disappeared. These findings indicate that subcutaneous administration enhances the placebo effect of acute treatment of migraine. Future trials of antimigraine drugs assessing the relative efficacy of various routes of administration should use a double-dummy technique. The interpreting of placebo-controlled trial results must therefore consider that the effect in the drug arm of the trial depends in part on the route of administration.