Abstracts

RATIONALE: Animal models, population-based epidemiological studies, and clinical series have suggested that status epilepticus (SE) contributes to brain injury and increases the risk of death. Using data from a very large sample of inpatients from US community hospitals, we have recently demonstrated that SE increases the risk of death after controlling for multiple co-morbid conditions. [At the end of this activity, the participant will be able to discuss how the risk of death among inpatients with SE differs by age.]
METHODS: The National Inpatient Sample of the Healthcare Cost and Utilization Project (1988-95) is a very large population-representative 20% sample of US hospital discharges, which contains over 50 million discharge records with data available at the patient level while protecting privacy. We extracted all discharge records with a diagnosis of SE (ICD-9 codes 345.2, 345.3, 345.7), and then obtained a random sample (stratified by year) of discharge records without a diagnosis of SE for a case:control ratio of 1:3. Discharge records with only a diagnosis of non-convulsive SE and discharge records of neonates (less than 1 month of age) were excluded from this analysis. Descriptive analyses were performed, and unadjusted odds ratios (OR) were caclulated using convulsive SE as the primary exposure and death as the primary outcome variable. Step-wise multiple logistic regression models were developed for all ages together as well as multiple age groups, with SE as the primary exposure variable and inpatient death as the outcome. Co-morbid conditions and other potential confounding variables were examined and considered for the logistic regression model; final regression models were developed based upon the best model fit.
RESULTS: 39,649 inpatient records with a diagnosis of SE and 118,383 discharge records without a diagnosis of SE were ascertained. The inpatient mortality was 10.4% among those with SE and was 2.9% among those without SE. The unadjusted and adjusted (after multiple logistic regression) OR are shown in the Table.
CONCLUSIONS: The increased risk of death associated with SE among inpatients varies by age. Children under the age of 1 year appear to be the most vulnerable. Children ages 1-4 years and adults have significant increased risk of death, and school-age children and younger teenagers do not appear to have significant increased risk of death associated with SE in our study population.[table1]
[Supported by: 1 RO3 HS11453-01 (ET) from the Agency for Healthcare Research and Quality.]