A case-control comparison of 114 transient global amnesia (TGA) patients using both normal community-based controls (n = 109) and transient ischaemic attack (TIA) controls (n = 212) showed no evidence of an increased risk of TGA associated with any of the conventional risk factors for cerebrovascular disease. However, migraine was significantly more common in the TGA patients than in either control group. A prospective longitudinal study using actuarial (life table) analysis confirmed that the prognosis of the TGA patients was strikingly better than that of the TIA controls. An important minority (7%) of TGA cases developed epilepsy, usually within one year of presentation. On the basis of this evidence, a thromboembolic aetiology for TGA can be confidently rejected, at least for the vast majority of cases. There are good theoretical reasons to link migraine and TGA casually, and occasionally epilepsy may masquerade as TGA. In the remaining cases the cause remains unknown.