Abstracts

Rationale: The impact of continuous EEG monitoring (cEEG) in the intensive care unit (ICU) on the outcomes of children has not been thoroughly investigated. Our institution implemented cEEG in 2014 and were in a position to compare patient populations before and after this process.  Methods: We conducted a retrospective study of critically-ill children with a neurological condition admitted to Seattle Children’s Hospital Pediatric ICU (PICU) where cEEG has been a standard of care for high risk patients since 2014. The study included three cohorts: (1) cEEG group received cEEG from 9/2014–6/2018; (2) contemporary control cohort did not receive cEEG from 9/2014–6/2018; and (3) historical control cohort did not receive cEEG from 1/2011-8/2014. Outcomes included mortality and Pediatric Cerebral Performance Category (PCPC) decline from baseline to ICU discharge and Pediatric Outcome Performance Category (POPC) and PCPC scores at ICU discharge. Data were extracted from our local Virtual Pediatric Systems database. EEG data were collected from the electronic medical record. We compared cEEG and contemporary control cohorts using Pearson’s chi-square test and ANOVA. Subgroup analyses were performed among patients with a diagnosis of seizure and of status epilepticus based on ICD 9/10 at discharge.  We used propensity score matching to compare study cohort to historical control cohort patients to evaluate the potential impact of cEEG on the outcomes. Generalized linear models were used to analyze the associations between cEEG implementation and outcomes in patients with seizures and status epilepticus in all of the cohorts.  Statistical tests were performed in either R version 3.1.3 or SAS 9.4. For all statistical tests, two-sided alpha <0.05 was considered significant.  Results: The cEEG cohort included 234 cases and contemporary control cohort included 2294 cases without cEEG. Among those with cEEG, 96 (41%) patients had seizures during cEEG, including 30 (31% among patients with seizures) with status epilepticus during cEEG. Patients who had cEEG had higher in-hospital mortality (17.5%) compared to contemporary control (3.1%) (p<0.001), and worse PCPC (2.82 vs 1.86) and POPC (3.31 vs 1.49) scores at ICU discharge (P<0.001). PCPC declined from baseline to ICU discharge in the cEEG group (1.04 vs 0.16) (p<0.001). There were no significant difference in baseline PCPC and POPC scores between the two groups. Using propensity scores, we matched 132 cases with seizure in the cEEG group with 132 historical controls and found similar mortality (9.8% cEEG group vs 7.6% control group, P=0.662), discharge PCPC (2.58 vs 2.41, p=0.18), discharge POPC (2.98 vs 2.79, p=0.056), and PCPC decline (0.6 vs 0.42, p=0.274). We matched 38 patients with status epilepticus in the cEEG group with 38 historical controls, and cEEG patients were more likely to have PCPC decline (0.55 vs 0.03 p=0.016). In general linear regression analysis, we found cEEG implementation was independently associated with worse discharge PCPC in patients with diagnosis of seizures (p=0.002).    Conclusions: Patients who received cEEG are at high risk of mortality and PCPC decline at ICU discharge. Further study is needed to determine the impact of cEEG guided management on outcomes, as the perceived clinical need for cEEG is likely biomarker of children with more severe neurologic disease.  Funding: No funding