Abstracts

Rationale: Staring as a suspicion for seizures is a very common reason for referral to child neurology. Many children admitted to a monitoring unit have been found to be for staring. The use of inpatient long term video EEG monitoring can be expensive and add a burden to healthcare costs. We performed a retrospective chart review of all children between 2009 and 2012 at Nationwide Children's Hospital who received long term video EEG monitoring (LTM) for the purposes of characterization of staring. The goals were to assess whether a LTM would be useful and cost effective in the analysis of staring as a symptomatic complaint. We wanted to evaluate if predictive factors are present that would increase the yield of a diagnostic result on the long term video EEG study. Methods: We encoded several characteristics of the most recent prior EEG. We categorized each patient as having had a normal prior EEG, an abnormal prior EEG, or unknown (i.e., either no baseline EEG or unavailable prior result). Binary indicators of focal epileptiform abnormalities, generalized epileptiform abnormalities, generalized slowing, and occipital intermittent rhythmic delta activity (OIRDA) was included. Spells were characterized as follows: Duration was short (less than 1 minute), long (longer than 1 minute), or unknown. Frequency was uncommon (3 or fewer per week), common (more than 3 per week), none in the past two months, or unknown. Automatisms were either present, absent, or unknown. Breakability was either: ever present, never present, or unknown. And post-ictal mental status was: ever changed, never changed, or unknown. We included indicators of the following previous diagnoses: epilepsy, attention deficit disorder (ADD), psychogenic non-epileptic seizures (PNES), pervasive developmental disorder (PDD), learning disability (LD), developmental delays (DD), bipolar disorder, or any other axis 1 mood disorder. Finally, we included indicators if the child had any current prescriptions for medications for a psychiatric condition or for epilepsy. The primary outcome was if the LTM admission captured a staring spell and diagnosed it as a seizure. Results: Of the 276 admissions, 29 (11%) captured a staring spell and diagnosed it as seizure (i.e., had a positive LTM study). Variables were present that increased yield of a diagnostic study. Based on this information, we created a scoring system as follows: -3 points if the previous EEG was normal, -1 point if the child takes a medication for a psychiatric condition, +1 point if the child takes an anti-epileptic drug for epilepsy, and +1 point if the spells last less than 1 minute. Reviewing the empirical data suggests the following interpretation of this score. If the total score is negative, the staring spells will only rarely be diagnosed as seizures (1 of 99 cases; 1%). If the score is 1, the staring spells were uncommonly diagnosed as seizures (3 of 55; 5%). If 2, the staring was occasionally (13 of 83; 15%) diagnosed and if 3, commonly (12 of 39; 30%). Conclusions: Based on the results of our study, it appears the yield of long term video EEG for staring is rather low. In summary, long term video EEG (LTM) testing is not a cost effective and efficient manner to diagnose if a child's staring is due to an epileptic seizure. If a patient has a score of less than 2 on the proposed scoring system, it is unlikely this test would be useful and strong consideration of utilizing clinical judgement only should occur. Consideration of other factors and variables that may improve the utility of the score would provide further improvement on which factors are needed prior to determining if a patient should undergo a prolonged LTM study. A prospective study validating the proposed scoring system is needed to substantiate our findings. Funding: No funding was received.