Abstracts

Rationale: Time to treatment in status epilepticus (SE) is often longer than suggested by guidelines, even when SE starts in the hospital. These delays are independently associated with increased morbidity and mortality. We aimed to compare the time to treatment for pediatric convulsive SE [time to first benzodiazepine (BZD), and time to first non-BZD anti-seizure medication (ASM)] between tertiary hospitals with and without overnight in-hospital pediatric neurology resident coverage. Our working hypothesis was that time to treatment would be shorter in centers with overnight pediatric neurology resident. Methods: Prospective observational study in 9 tertiary centers. Inclusion criteria were: (1) admission to a pSERG tertiary hospital between January 2014 and January 2019; (2) SE onset in a tertiary hospital; (3) SE onset after 5pm and before 8am (shift time frame); (4) failure of first line ASM (BZD) for seizure control; (5) age from 1 month to 21 years; (6) focal or generalized convulsive seizures at onset; (7) known time of SE onset and (8) known time of ASM administration. Exclusion criteria were: (1) unconventional treatment (i.e., non-BZD ASM before BZD), and (2) non-convulsive SE at diagnosis. If a patient had >1 SE episode, then only the 1st episode was included. Statistical analysis. Wilcoxon rank sum test (to compare the median time to treatment between both groups), and time-to-event analysis (Gehan-Breslow-Wilcoxon test for univariate analysis) and Cox proportional hazards regression model for multivariate analysis, in which the outcome was time to treatment (to first BZD in one regression and to first non-BZD ASM in another regression), the main predictor was pediatric neurology resident in-hospital overnight (dichotomous yes/no) and the covariates were SE type (dichotomous intermittent/continuous), history of epilepsy (dichotomous yes/no), and prior SE (dichotomous yes/no). Continuous variables are presented as median (25th – 75th percentile). Results: We studied 56 patients (52% males) with a median (p25-p75) age of 4.1 (1.6-7.3) years. Table 1 shows patients’ demographic and clinical characteristics. First BZD. The median time to receive the first BZD between both groups was not statistically different [without resident 5 (3-10) vs with resident 8 (4-17), p=0.33]. The presence of an in-hospital pediatric neurology resident overnight was not a significant predictor in the univariate analysis (p=0.3), nor in the multivariate analysis (HR 1.62, 95% CI 0.82-3.17, p=0.16) (KM curves, Figures 1a-1b). None of the covariates analyzed were statistically significant in the multivariate analysis. First non-BZD ASM. The median time to receive the first non-BZD ASM between both groups was not statistically different [without resident 44 (27-55.5) vs with resident 37 (21-59), p=0.40]. The presence of an in-hospital pediatric neurology resident overnight was not a significant predictor in the univariate analysis (p=0.4), nor in the multivariate analysis (HR 0.88, 95% CI 0.46-1.68, p=0.70) (KM curves, Figures 1c-1d). None of the covariates analyzed were statistically significant in the multivariate analysis. Conclusions: In this study, the presence of pediatric neurology resident in-hospital overnight coverage did not impact treatment delays. Limitations included a small sample size, and uncontrolled for potential confounders, such as hospital effect, internal SE pathways or seizure action plans, and hospital location at SE onset (such as, ED, ICU, Neurology Ward, or other Wards). Funding: Funded by the Pediatric Epilepsy Research Foundation & the Epilepsy Research Fund