Matthew Bower, MD
Neurology Resident (PGY-4), UC Irvine
Incoming Neurocritical Care Fellow 2022
Johns Hopkins University
Sara Stern-Nezer, MD, MPH
Assistant Clinical Professor
Departments of Neurology & Neurosurgery
University of California, Irvine
Ruijter B, Keijzer M, et al. Treating Rhythmic and Periodic EEG Patterns in Comatose Survivors of Cardiac Arrest. NEJM (2022). DOI: 10.1056/NEJMoa2115998
Background: Up to 10-35% of comatose post cardiac arrest patients have epileptiform activity such as periodic generalized discharges (GPDs) or rhythmic activity. Electrographic or clinical seizures are rare, but when they do occur it portends a poor neurological prognosis. However, the significance of epileptiform activity and whether treatment is beneficial remains unclear.
Methods: The Treatment of Electroencephalographic Status Epilepticus After Cardiopulmonary Resuscitation (TELSTAR) trial was a multicenter, open label, randomized control trial with blinded outcome evaluation to evaluate whether escalating antiseizure therapy compared with the standard of care improved neurological outcome. The trial included adults aged ³18 years who were comatose based on a Glasgow Coma Scale £8 with EEG findings of periodic discharges, rhythmic delta activity, or spike and wave discharges that persisted for at least 30 minutes or 5 minutes occurring twice within a 60 minute period. Antiseizure therapy was escalated as follows: 1.) one antiseizure drug (phenytoin, valproate, or levetiracetam) and one sedative agent 2.) a second antiseizure drug and a second sedative agent 3.) addition of a barbiturate. The goal was >90% of activity suppressed for at least 48 hours. Classification of the baseline EEG and the degree of suppression was determined by two authors who were aware of the group assignments.
Outcome metrics: The primary outcome was the Cerebral Performance Category (CPC) at 3 months determined by an evaluator blinded to group assignment. Secondary outcomes included mortality at 3 months, ICU length of stay, and mechanical ventilation duration. Statistical analysis was an intention-to-treat comparison of the groups in terms of risk difference for a poor outcome defined as CPC 3-5.
Results: In total, 88 patients were included in the antiseizure treatment group and 84 patients in the control group. For the primary outcome, 79 of 88 (90%) patients in the treatment group and 77 of 84 (92%) patients in the control group had a poor outcome (risk difference 2%; 95% CI -7 to 11; P=0.68). Mortality at 3 months occurred in 70 (80%) patients in the treatment group and 69 (82%) patients in the control group (risk difference 3%; 95% CI -9 to 14). While 3-month mortality was the same between groups, there was a lower mortality rate within the first 24 hours after randomization for the treatment group (9% vs. 24%). The mean length of stay in ICU was longer in the treatment group (8.7 vs. 7.5 days), and duration of ventilation was longer in the treatment group (7.8 vs. 6.6 days).
Commentary: This multi-center, open label, randomized control trial demonstrated that suppression of epileptiform activity post cardiac arrest did not improve outcomes. The primary finding is in contrast with a prior observational study by Beretta et. al. demonstrating that stepwise antiseizure treatment resulted in a good outcome in 16 of 36 (44%) patients with rhythmic and periodic EEG patterns. However, in this trial, 0 of 13 (0%) patients with GPDs had a good outcome with treatment. The TELSTAR subgroup analysis also suggested a trend towards fewer good outcomes in the treatment group for patients who had GPDs, but the study was not powered for this analysis. Studies postulate that outcomes with GPDs may be poor regardless of antiseizure treatment as it represents severe anoxic brain injury rather than epileptic activity.
While this study’s strengths included a prospective and randomized design with blinded outcome evaluators, there are multiple limitations including a significant proportion of the control group that achieved some degree of EEG suppression, missing data on time from arrest to return of spontaneous circulation (ROSC), and the possible effect on withdrawal of care considerations by the treating physicians who were aware of group assignment. Suppression of activity for 48 hours was only achieved in 56% of the treatment group. Two percent of the control group also achieved suppression for 48 hours; however, many of the control patients also achieved either partial (17%) or full suppression (47%) for at least 24 hours. The use of routine sedative medications for patients on mechanical ventilation may have contributed to suppression of EEG patterns and the lack of difference in outcomes. Baseline characteristics were not different between the groups, but there was a significant proportion of missing data in both groups regarding time from cardiac event until initial resuscitation (treatment group 25%; control 18 of 84, 21%) and time from resuscitation to ROSC (treatment 26 of 88, 30%; control group 26%). Therefore, it is uncertain whether the groups were balanced in terms of duration of arrest and by extension the potential severity of anoxic injury. Given the higher mortality in the control group the day after randomization, it is possible that withdrawal of care discussions may have been biased as treating physician was aware of the group assignment. Finally, the authors caution that the wide confidence interval for the primary outcome does not rule out a mild benefit or harm. Due to the limitations, further studies are needed to assess whether suppressing specific patterns, especially those other than generalized periodic discharges, may improve neurological outcomes in comatose survivors of cardiac arrest.