A study was conducted to clarify whether ultra-early referral of patients with subarachnoid haemorrhage (SAH) is effective for improving the management outcome. The subjects were 455 patients who were admitted within 6 h after initial SAH. Of these patients, 289 were treated surgically, 159 of them within 24 h. At 6 months, 228 patients (50%) had a favourable outcome including good recovery or modrate disability, 37 (8%) had severe disability, and 190 (42%) had an unfavourable outcome including vegetative state or death. Of 214 patients with an admission grade IV or V, 47 (22%) had a favourable outcome. In 10 patients, emergency procedures such as haematoma removal or ventriculostomy were definitely effective, and in 13, early surgery may have been the reason for the improved outcome. However, in 24 patients, the reasons for a favourable outcome were not related directly to ultra-early referral; in 19 of them, there was spontaneous improvement of clinical grade and/or no SAH on computed tomography. Of 218 patients with admission grade I or II, 30 (14%) had an unfavourable outcome, and in 12 of them, this was ascribed to rebleeding. The rebleeding rate acid severity of vasospasm were not significantly reduced by surgery carried out within 24 h after SAH, in comparison with surgery carried out between 24 and 48 h, and there was no significant difference in surgical outcome between them. It is concluded that although ultra-early referral of patients with SAH was expected to improve the outcome in emergency cases, no substantial improvement in overall management outcome seems to have been achieved by this policy.