Sudden Unexplained Death due to Epilepsy (SUDEP) is not uncommon. The incidence may be difficult to ascertain due to variations in reporting the cause of death investigated by coroners. To study this, we used a cuttings service to identify all cases of SUDEP reported in the UK press during 1992. These cases had been investigated by a number of different pathologists and coroners. Post-mortem reports, witness statements and other relevant information were examined in 40 cases. Inconsistencies were observed both in investigations performed and observations made at time of death. Varying degrees of detail concerning 'type' and history of epilepsy were found. In 70% of cases, type of epilepsy was either not known or not referred to. Similar inconsistencies were found with details of medication, position of body, toxicology reporting and detailed examination of organs, particularly the brain. Attributed cause of death varied considerably, with epilepsy stated as primary cause of death in less than half of the sample. Disparity was noted also in coroner's verdicts, with no distinct pattern emerging in relation to attributed cause of death. The wide variation in practice of individual coroners and pathologists in the investigation and registering of sudden deaths raises issues of quality assurance. As the majority of these deaths are unwitnessed, we rely on these investigations and the statements of officials in the process of registering deaths. If we are to build a profile of people who are at risk of SUDEP, epilepsy must be stated on the death certificate. Until this happens, many more of these deaths will go unrecorded.