Abstracts

Rationale: Although the epidemiology of convulsive status epilepticus (CSE) has been well described, few studies have addressed risk factors for NCSE. It has been hypothesized that NCSE patients with a history of epilepsy have a better prognosis than those with acute medical problems as an etiology. The purpose of this study was (1) to describe the clinical characteristics associated with NCSE, as compared with CSE, and (2) to assess whether NCSE in the setting of epilepsy has a lower mortality compared with other etiologies of NCSE. Methods: Prospectively identified cases of CSE and NCSE in the Richmond, Virginia metropolitan area SE database were reviewed. NCSE included simple partial SE and complex partial SE without associated tonic-clonic or clonic activity, absence SE, subtle SE, and electrographic SE in patients with coma. CSE included primary and secondarily generalized convulsive SE, myoclonic SE, and localization-related seizures with clonic or tonic-clonic manifestations. Mortality was assessed at 30 days. Chi Square analysis was used to evaluate associations between single variables. The Kruskall-Wallis test was used to compare SE durations. Multivariate regression analyses were performed to identify independent risk factors.Results: Of 809 cases of SE, 10% were NCSE, and 90% were CSE. There were no significant differences in age, race, or seizure history between the two groups. Among NCSE cases, those with a history of seizures had significantly lower mortality than those without a history of seizures (p = 0.0013, OR 0.1, CI 0.025 – 0.410), and females had lower mortality than males (p=0.0271, OR 2.7, CI 0.085 – 0.863). Among CSE cases, those with a history of seizures also had a significantly higher survival rate. In-hospital onset of SE occurred in 61% of NCSE and 30% of CSE (p <0.0001). Multivariate analysis identified SE onset location as the only independent risk factor for NCSE (p = 0.0002, OR 2.8, CI 1.6 – 4.9). Etiologies of hypoxia/anoxia and CNS acute occurred in a higher proportion of NCSE compared with CSE (p <0.0001) but did not independently predict SE type or mortality in NCSE. Median SE duration was 210 minutes for NCSE and 95 minutes for CSE (p<0.0001). Mortality was 41.7% for NCSE, and 23.0% for CSE (OR 2.4, CI 1.5 – 3.8). With multivariate regression analysis of NCSE and CSE groups, and mortality as outcome, increasing age, etiology, in-hospital onset and longer SE duration significantly predicted mortality, while neither SE type (NCSE vs. CSE), nor seizure history were independent risk factors for mortality.Conclusions: In-hospital SE onset occurs significantly more frequently with NCSE than CSE. Among cases of NCSE, female gender and history of seizures are independently associated with a lower risk of mortality. Compared with CSE, NCSE has a different distribution of etiologies, and a longer duration. A history of seizures was associated with increased survival following NCSE and CSE. The odds ratio for mortality is 2.4 times higher for NCSE than CSE, but SE type does not independently predict mortality.