PCCM Fellow Physician Lucile Packard Children's Hospital Stanford Palo Alto, California, United States
Background: In adults with sepsis, slow electroencephalography (EEG) background activity is common. Lack of EEG reactivity is associated with mortality in adults with sepsis. However, data about EEG features in sepsis are limited in children. Objective: To determine the prevalence of abnormal EEG findings in children with sepsis and to test if lack of EEG reactivity is associated with mortality or new morbidity. Design/Methods: We performed a retrospective cohort study of children admitted to the Children’s Hospital of Philadelphia’s (CHOP) intensive care unit (ICU) from 2013-2019 who had an EEG within 14 days from sepsis recognition. EEG results were categorized by background activity, asymmetry, reactivity, and ictal activity. Primary outcome was death or new morbidity in survivors (change of >= 2 points in Functional Status Score at ICU discharge from baseline). Bivariable logistic regression evaluated the association between EEG findings and death or new morbidity. Results: EEG results were available for 271 children, 45% of whom were female. Median age was 5.3 years [IQR 1.5-12.9]. 87% required invasive mechanical ventilation and 61% required vasoactives. Co-morbidities included epilepsy (38%), immunocompromise (17%), and tracheostomy dependence (13%). Infections were bacterial in 36%, viral in 24%, culture negative in 22%, and multifactorial in 18%. Death or new morbidity occurred in 49% (death in 13%; new morbidity in 36%). EEGs were obtained at a median of 1.1 days [0.7-3.1] from sepsis recognition and were most often for seizures (54%) or altered mental status/encephalopathy (48%). EEG background was normal in 18%, slow-disorganized in 68%, discontinuous in 5%, burst-suppressed in 1%, and attenuated-featureless in 8%. Asymmetry was noted in 18%. Epileptiform discharges were present in 36%, electrographic seizures in 11%, and electrographic seizures with status epilepticus in 9%. EEG reactivity was absent in 26% and associated with 1.7-fold odds of death or new morbidity (95%CI 0.94-3.01, p=0.08). Clinical seizures occurred in 44% and were associated with 0.49-fold odds of death or new morbidity (95%CI 0.30-0.79, p=0.004). There was no association of EEG background or electrographic seizures/status epilepticus with death or new morbidity (p=0.25 and p=0.42, respectively).
Conclusion(s): EEG abnormalities were common in children with sepsis. While mortality and new morbidity was high, only the absence of EEG reactivity was associated with increased ICU mortality and morbidity. Studies are needed to identify children with sepsis who could benefit from routine EEG monitoring.
