“Is it a seizure?” Prediction tool for seizure likelihood in children aged 1–24 months admitted for electroencephalographic monitoring of paroxysmal, rhythmic, and repetitive events

Abstract

AbstractObjectiveParoxysmal, rhythmic, and repetitive events (PRREs) during infancy can be concerning for possible seizures, especially following perinatal brain injuries. The workup for establishing whether a PRRE represents a seizure involves the use of continuous video‐electroencephalography (cVEEG) for event characterization. This study aims to determine the diagnostic yield of cVEEG for evaluating events concerning for seizures in children aged 1–24 months.MethodsWe performed a single‐center retrospective chart review (January 1, 2019–December 31, 2020) and included all children aged 1–24 months admitted for PRRE capture and characterization using cVEEG. Chart abstraction included demographics, birth and family history, known brain injury, event semiology, duration, and frequency, as well as interictal electroencephalographic (EEG) features. For each of these variables, odds ratios for seizure prediction were calculated.ResultsA total of 243 patients were identified for this study. On cVEEG, n = 160 (65.4%) had a target event of concern captured during an initial admission with cVEEG. Whereas n = 41 (25.8%) patients with events captured were confirmed to have seizures, most patients (n = 119) were confirmed to have nonepileptic events. Several variables predicted seizure likelihood during the initial admission with cVEEG, including event duration (>1 min), frequency (occurring ≥3 times per week), and presence of abnormal interictal findings on cVEEG. For patients who did not receive a diagnosis at the time of initial admission with cVEEG, the likelihood of a subsequent epilepsy diagnosis was associated with specific PRRE semiology (motor active or motor passive), longer event duration (>1 min duration), and the presence of interictal abnormal EEG features on initial cVEEG admission. Prediction tools utilizing scoring systems to stratify risk in infants with suspected seizures due to PRREs are included.SignificanceUnique patient attributes and PRRE characteristics, as well as the presence of EEG interictal abnormalities, can provide valuable insights for discerning children with a higher likelihood of epilepsy diagnosis following cVEEG admission.