Abstracts

Rationale: To determine the utility of death certificates as a means of finding deceased patients with epilepsy (PWE) whose death might be attributed to sudden unexpected death (SUDEP). We examined characteristics of death certificates for persons with definite, probable and possible SUDEP and their medical records and compared that information to that of PWE that died of other causes. Methods: Names, social security numbers, dates of birth and dates of death (if known) for 2,762 PWE were submitted to the state bureau of vital statistics. These PWE were seen in a single epilepsy practice and were followed for one to 30 years. There were 110 deaths known to the clinicians. However, 411 deaths were recorded as of December 31, 2005 for this cohort, and death certificates were obtained for 390. Logistic regression was performed to evaluate 24 items of information found in the death certificates for persons who died of definite/probable or possible SUDEP. Standardized mortality ratios (SMR) were calculated for all cause mortality, suicide, homicide, accidental, and natural causes.Results: Of all deaths, 18.2% were seizure related, and 12.5% were due to definite/ probable or possible SUDEP. Death certificates of 88.2% of the definite/probable and 78.1% of the possible SUDEP cases mentioned either seizure or epilepsy. 70.6% of definite/probable, 57.1% of possible, and 28.8% of non-SUDEP cases died at home. Death certificate predictors of definite/probable SUDEP were place of death, age, and mention of seizure/epilepsy. Predictors of possible SUDEP were time of death, race, autopsy, and mention of seizure/epilepsy. Persons in the epilepsy cohort experienced 21% more mortality over what would be expected in the South Carolina population. All age groups experienced increased mortality except for persons over 75 years. The overall SMR for males was significant at 1.39, but females did not show increased mortality. The SMR for drowning was 9.94 (95% CI: 4.54-18.87), and SMRs were significantly elevated for ages 15-54 years. Conclusions: Death certificates can provide useful information in identifying definite/probable or possible SUDEP if multiple variables are searched. Persons with epilepsy experience more mortality than the general population, especially among males and persons younger than 75. Accidental drowning is a significant problem in this population. Contrary to previous studies, in this long-term follow-up cohort, the incidence of suicide was not elevated over what would be expected. In light of the records indicating only 27% of the deaths, a better system of periodic surveillance of epilepsy clinic populations should be undertaken to ensure that all deaths are captured. (Source of funding: Epilepsy Services Research, Inc.)