Abstracts

Rationale: Hyperventilation (HV) is a convenient bedside method for eliciting absence seizures (AS) and is often used both to establish a new diagnosis of AS and to monitor the effectiveness of ongoing drug therapy of AS. We set out to measure the sensitivity of HV in two types of gold standard groups presumed to have active AS by different criteria: 1) a high confidence group of children who manifest spontaneous AS on routine EEG testing (outside of HV) and 2) a lower confidence group of children whose AS was judged active by parental report. Methods: The EEG diagnosis log at Children s Hospital Boston was screened from January 2006 to June 2010 for all patients with 3 Hz generalized spike-wave discharges or absence seizures. The subset of these patients matching ILAE criteria for childhood or juvenile absence epilepsy formed the study group (n=68). Two comparative gold standard groups of EEGs (all with adequate HV) in children with presumed active AS were formed: 1) those EEGs which showed spontaneous AS (outside of HV) and 2) those EEGs done at a time that the child s parents reported active AS in their child. Hyperventilation was carried out for a minimum of 3 minutes. An AS was defined as an episode of generalized spike-wave discharges associated with clinical change or unresponsiveness. Results: 53 EEGs that captured spontaneous AS were found in which HV was also done (in 44 unique patients). In 41 (77%) of these EEGs HV also elicited an episode of AS (71% of those untreated [n=28] and 84% of those treated [n=25] at the time of the EEG). 95 EEGs were identified (in 55 unique patients) whose parents reported active AS at the time of the EEG. Of these, HV elicited AS in 59 (62%) EEGs (71% of those untreated [n=41] and 56% of those treated [n=54]). Of the 84 EEGs done at a time that parents reported no active AS, AS was seen spontaneously or during HV in 30 (36%) tracings. Conclusions: Even in children with active AS by the strict criterion of the occurrence of spontaneous AS in the EEG, only 77% had an AS during HV, implying that this technique will miss approximately one quarter of children with active AS. When positive parental report was used as the gold standard for presumed active AS, only 62% of children manifested AS during HV; the different rates imply that parents may overestimate the activity of AS approximately one quarter of the time in children who are effectively treated. Conversely, in patients in whom AS was definitely still active by EEG criteria, parents underreported AS activity 36% of the time. This percentage represents a minimum estimate since some children with active AS may not have exhibited their seizures during the period of the EEG recording. These findings imply that parental report has an approximate false-positive rate of one quarter and an approximate false-negative rate of more than one third.