Abstracts

Rationale: Infantile spasms is a potentially devastating form of childhood epilepsy, often but not always associated with the chaotic, high-voltage interictal EEG pattern known as hypsarrhythmia. In addition to seizure-control, the elimination of hypsarrhythmia is a critical factor in successful treatment. However, the pathogenesis of hypsarrhythmia is poorly understood. Although 18-fluorodeoxyglucose positron emission tomography (FDG-PET) has previously been used to implicate the bilateral caudate nuclei and temporal cortex in the pathogenesis infantile spasms, the specific contribution of hypsarrhythmia is unknown. Among a cohort with infantile spasms, we set out to examine the relationship between hypsarrhythmia and regional cerebral metabolism.Methods: Patients with video-EEG confirmed infantile spasms who underwent simultaneous EEG and FDG-PET as part of a presurgical evaluation between 2008 and 2013 were retrospectively identified. For each patient, we tabulated the presence or absence of hypsarrhythmia (including ‘modified’ hypsarrhythmia), high voltage (>200µV), and frequent multifocal independent spike discharges (MISD). Relative cerebral metabolism (normalized to pons) was ascertained in 47 prespecified regions of interest (ROIs). Cerebral metabolism in each ROI was compared between patients with and without (1) hypsarrhythmia, (2) high-voltage background, and (3) MISD, using unpaired t-tests. Adjustment for multiple comparisons was accomplished with the Bonferroni procedure.Results: We identified 72 patients with infantile spasms who underwent simultaneous EEG/PET, including children with hypsarrhythmia (n=10), high-voltage EEG background (n=23), and MISD (n=20). After adjustment for multiple comparisons, the presence of hypsarrhythmia was associated with a substantial increase in relative cerebral metabolism in left superior lateral temporal cortex (p < 0.001). We specifically did not observe a similar association for right superior lateral temporal cortex and there was no observed association between relative cerebral metabolism and high-voltage EEG background or multifocal independent spike discharges in any ROI.Conclusions: The identification of left superior lateral temporal cortical dysfunction in hypsarrhythmia is compatible with prior studies of infantile spasms which broadly implicate temporal cortex. The left-sided and superior temporal specificity is of interest as linguistic/communication functioning is disproportionately affected in children with unsuccessfully treated infantile spasms, in comparison to other aspects of neurodevelopment. Further study is warranted to ascertain which EEG features of hypsarrhythmia are the most critical determinants of relative cerebral metabolism. This study suggests that relative metabolic activity in left superior lateral temporal cortex may be a useful biomarker in the identification and treatment of infantile spasms.