Rationale:
Recent studies estimate that delays in electroencephalography (EEG) monitoring may lead to missed seizures in up to 17% of high-risk patients (Fatima et al., 2022 J Clin Neurophys 10:1097). Delays in status epilepticus diagnosis and treatment has been associated with poor health outcomes (Hill et al., 2017 Ann of Neurology 82(2): 155-165). Point-of-care EEG (POC EEG) has shown to shorten time to EEG and triage in critically-ill patients (Vespa et al., 2020 Critical Care Medicine, 48[9]). We aimed to assess the impact of POC EEG on length of stay (LOS) in the intensive care unit (ICU) in large academic centers with 24/7 coverage of conventional EEG (cEEG). We pose that patients who received POC EEG instead of or before cEEG would have overall shorter ICU LOS, potentially due to reduction in unnecessary treatments and faster triage.
Methods:
Data were collected as part of a multicenter retrospective study (SAFER-EEG), where adult patients were assigned to three cohorts: monitored with cEEG only, monitored with POC EEG only, and those who started with POC EEG and were followed by cEEG. We included patients from four centers who were admitted to the ICU during their hospital stay and started an EEG study before ICU discharge. For the first analysis, we compared ICU LOS within sites that had access to both POC EEG and cEEG. In a second analysis, we compared ICU LOS between a site that only had access to cEEG and the subgroup at other sites who had POC EEG first, followed by cEEG. In order to further explore the relationship between ICU LOS and EEG type, we performed subanalyses in patients with diagnosed seizures, in those who survived their ICU stay, and in those for whom EEG was started after hours.
Results:
Records from 658 patients were included in the analysis (cEEG = 437, POC-cEEG = 221). Two sites had access cEEG and POC EEG, and another site only had access to cEEG. First, we compared ICU LOS from sites with both POC EEG and cEEG available and found that ICU LOS in the POC-EEG groups was significantly lower than cEEG group (median
POC-EEG = 4.48 [2.15, 10.99] d vs.median
cEEG = 8.13 [2.98, 16.87] d;
p = 0.007).
A majority of the sub-analyses in patients who survived their ICU stay, those with diagnosed seizures, and those for whom EEG was started after-hours showed significantly lower ICU LOS in groups receiving POC EEG.
In the site with only cEEG, we observed a significantly higher ICU LOS when compared to the patients from other sites who received POC EEG followed by cEEG (median
cEEG = 6.65 [3.38, 12.42] d, median
POC-cEEG = 4.46 [2.39, 10.87] d;
p = 0.006). See Figure 1. The trend of shorter ICU LOS for those receiving POC EEG held among the sub-analyses when compared against a site with access only to conventional EEG.
Conclusions:
Patients who first received a POC EEG had, as a group, about three days shorter ICU LOS when compared to those who received cEEG, within sites and across different institutions. Multiple sub-analyses demonstrated the same trend. These data highlight that even in hospitals with 24/7 cEEG coverage, POC EEG can aid in patient assessment and early seizure detection, and can significantly shorten ICU stay.
Funding: Study funding provided by Ceribell, Inc.