Steroids have been beneficial in the treatment of demyelinating diseases with features similar to those of Guillain-Barré syndrome (GBS). However, steroid treatment of GBS has been disappointing; in an earlier trial oral prednisolone was ineffective, although the dose was low and the sample small. We assessed the benefit of a high-dose steroid regimen in a large sample of patients with GBS in a multicentre, randomised, double-blind trial. 242 adult patients were randomised to receive intravenous methylprednisolone (IVMP) 500 mg (124 patients) or a placebo (118) daily for 5 days. Patients were diagnosed by standard clinical criteria and entered the trial within 15 days of onset of neurological symptoms. All patients were too weak to run. Some patients received plasma exchange depending on the practice of their centre. Disability was graded on a scale from 0 (healthy) to 6 (dead) at intervals for 48 weeks. There was no significant difference in any outcome variable between patients treated with IVMP and those given placebo. The most important outcome was the difference between the groups in disability grade 4 weeks after randomisation, which was only a 0.06 grade (95% Cl -0.23 to 0.36) greater improvement in the IVMP than the placebo group. The 39 patients in the IVMP group who required ventilation did so for a median of 18 days, 9 days fewer than the 44 patients who had a placebo and required ventilation (95% Cl -9.6 to 27.6). Median time to walk unaided was 38 days in the IVMP patients and 50 days in the placebo patients (difference 12 days, (95% Cl -21.3 to 45.3). A short course of high-dose IVMP given early in GBS is ineffective.