Abstracts

Status epilepticus (SE) is associated with an increased risk of death among US inpatients. We are conducting a series of studies investigating whether SE increases the risk of death among specific subpopulations of inpatients. This study investigates the impact of SE on the risk of death among a very large population of inpatients with carotid artery distribution stroke.
We used data from the National Inpatient Sample (NIS) of the Healthcare Cost and Utilization Project (HCUP-3), 1988-97, to ascertain discharge records with a diagnosis of carotid artery distribution stroke (ICD-9 433.1, 433.10, 433.11). The NIS is based upon a 20% stratified single stage cluster sample of US hospitals. NIS contains patient-level data, while maintaining patient privacy. Our analyses used the NIS unweighted data. Neonates were excluded. Descriptive analyses were performed, and unadjusted odds ratios were calculated using in-hospital death as the primary outcome variable and SE as the primary exposure variable. Multiple logistic regression models were developed using Stata 7.0, entering variables in a step-wise fashion that were both clinically relevant co-morbid conditions and potential confounding variables on univariate analysis.
403,942 inpatient discharge records had a diagnosis of stroke due to carotid occlusion, with an overall case fatality rate of 2.25%. Only 246 discharge records had a diagnosis of SE. The case fatality rate among those with both SE and carotid distribution stroke was 13.82%.The age distribution of those with carotid stroke was skewed towards the elderly; less than 0.1% of the total stroke population was less than 25 years of age. SE increased the risk of death among patients with carotid distribution stroke (unadjusted OR=6.92; 95% CI=4.82, 9.95). After adjusting simultaneously for multiple confounding variables, SE still increased the risk of death (adjusted OR=3.89; 95% CI=2.55, 5.91). Other factors that significantly increased the risk of death in the final model included cerebral edema (OR=15.47; 95% CI=11.48, 20.84), co-existing subarachnoid hemorrhage (OR= 12.10; 95% CI=9.54, 15.36), anoxia (OR= 11.25; 95% CI=8.75, 14.46), co-existing intracerebral hemorrhage (OR= 7.38; 95% CI=6.45,8.45), and hyponatremia (OR= 5.08; 95% CI=4.11, 6.29). Risk factors for death that were not as significant as SE in the final model, but considered important risk factors for death by clinicians included congestive heart failure (OR= 2.17; 95% CI=2.05, 2.30), and cardiac arrhythmia (OR= 1.11; 95% CI= 1.05, 1.17).
SE significantly increased the risk of death among a large sample of inpatients with carotid stroke, even after simultaneously adjusting for multiple co-morbid conditions. The very small percentage of stroke inpatients with SE (0.06%) in this study suggests that SE may be underdiagnosed among US inpatients with carotid disbribution stroke.
[Supported by: grant RO3 HS11453-01 (ET) from the Agency for Healthcare Research and Quality.]