Abstracts

Rationale: Status Epilepticus (SE) is a group of diverse conditions with significant variability in severity, response to treatment and consequently,outcome. We identify predictors of poor outcome in patients with SE admitted to NeuroIntensive care unit (NICU). Methods: This is a retrospective analysis of prospectively collected data from SE patients admitted to the NICU from November 2013 to January 2016. Patients' demographics, status semiology at presentation and during the course of hospitalization, intractability to treatment, underlying etiology, past history of epilepsy and admission benzodiazepine use were the variables that were analyzed using STATA software. Glasgow outcome scale (GOS) was used to identify bad outcome (1-3) and good outcome (4&5). Results: Characteristics and variables of 100 patients, aged 45 to 65 years (mean 58) were analyzed (Table1). Median length of stay (LOS) was 7 days [[nterquartile Range (IQR) 4-17] and median NICU LOS was 3 days (IQR 2-7). Most patients had good outcome with median GOS of 4 (IQR 3-5). Overall, 31 patients had poor outcome (16 males and 15 females)(p=0.853). In patients with poor outcome, median age was 59 years (p=0.2699). Neither age nor gender predicted poor outcome. When etiology was analyzed, there was no difference in outcome between structural and non-structural causes of SE. However, prior history of epilepsy was a strong negative predictor of poor outcome with only 14 of 70 patients (20% with prior history of epilepsy having poor outcomes (p < 0.01). Sixty-five patients had refractory SE, 25 (38%) of which had poor outcome (p < 0.01). Outcome was analyzed based on status semiology on initial presentation, revealing poor outcome in 16 of 37 complex partial SE (CPSE) (p=0.04), 9 of 48 generalized convulsive SE (GCSE), all patients with Myoclonic SE (n=2), and 3 of 9 nonconvulsive SE (NCSE) (p < 0.01). NCSE at any time during the hospital course (including at presentation) was seen in 31 patients and 14 (45.2%) of these patients had poor outcome (p=0.02). Mean number of days on ventilator was higher in patients with NCSE than those without( 9.2 vs 1.6; p=0.0001) and in those with new onset seizures than those without (7.8 vs 2.9; p= 0.001). Methods of treatments revealed that only 7 of 31 (22.5%) patients who received adequate benzodiazepine dosing had poor outcome (p=0.2247). A logistic regression analysis with backward selection for the significant outcomes yielded history of epilepsy as negative predictor of bad outcome (OR 0.2 [95% CI 0.08-0.64]; p < 0.01) and refractory status approached clinical significance (OR 3.3 [ 95% CI 0.98-11.08];p=0.05). Conclusions: Outcome from SE tends to be poor in patients with CPSE, refractory SE or in patients who had NCSE either from the onset or anytime during the course. Impact of these factors is statistically significant. Past history of epilepsy was associated with statistically significant lower incidence of poor outcome. This may represent the sensitivity to adequate treatment in patients who may present with SE due to medication non-adherence. Our cohort represents SE seen in a NICU and skewed towards those with a tendency to have refractory SE or underlying etiology leading to a higher rate of bad outcomes. Further prospective studies may be useful to provide insight in modifying predictors to improve outcome in SE. Funding: None