Intravenous (n = 21) and inhalational maintenance anaesthesia (n = 21) were compared by random allocation in patients with the aspirin intolerance syndrome undergoing endoscopic nasal procedures. Premedication was with oral midazolam and intravenous methylprednisolone sodium succinate 10 mg kg-1. Anaesthesia was induced in both groups with etomidate and alfentanil and ventilation was controlled. Anaesthesia was maintained in the intravenous group by infusion of alfentanil 1-1.5 micrograms kg-1 min-1 and injections of midazolam 2.5-5 mg h-1, and in the inhalational group by isoflurane up to 2%. Moderate arterial hypotension (70 mmHg) was achieved with nitroglycerine 0.5-5 micrograms kg-1 min-1 in the intravenous group, and with isoflurane up to 2% in the inhalational group. Adrenaline 1: 200 000 with 2% lignocaine was injected into the operative field. One patient in the inhalational group developed a resistant tachyarrhythmia but there was no overall significant difference (P = 0.34) in the frequency of dysrhythmias precipitated by adrenaline and lignocaine between the two groups. In one patient of each group methylprednisolone precipitated bronchospasm. On later challenge testing, 125 mg of intravenous methylprednisolone significantly reduced the peak expiratory flow (P < 0.05) in one of these patients. The results suggest that intravenous and inhalational maintenance anaesthesia are equally suitable for patients with aspirin intolerance syndrome. Corticosteroids during surgery should be given by the same route used pre-operatively (spray, oral, or spray plus oral) because intravenous injection may have adverse effects.