Abstracts

Rationale: Early onset childhood epilepsy is associated with poor neurodevelopmental outcome. Few studies of the natural history exist however; determining predictors of poor outcome is important in identifying those at risk and for developing appropriate intervention strategies. We report data from a prospective population-based study evaluating neurodevelopmental outcome of children with epilepsy onset < 2 years. Methods: Children < 2 years of age presenting with recurrent unprovoked seizures from a defined area of North London were enrolled over two years. Clinical assessments and neurodevelopmental evaluation were performed at enrolment using Bayley Scales of Infant and Toddler Development-III and at follow-up using the Leiter-R. Neuroimaging and EEG data at presentation were evaluated. Univariate nonparametric statistical analysis (Mann-Whitney, Kruskal-Wallis and Spearman s Rho) were used to investigate developmental outcome in relation to infantile spasms, MRI and EEG. Results: Of 69 infants enrolled, data from baseline and 3 year follow-up were available for 36 (23 male, age onset M=6.6 m, age follow-up M=49.5 m).At follow-up 61.1% had an IQ > 2SD below the mean. Cognitive baseline scores and follow-up IQ scores were strongly correlated r = 0.758 (36), p < 0.01. The presence or absence of an aetiological relevant finding on MRI was related to IQ: relevant finding (47.2%) M IQ=59, nonspecific abnormality (25%) M IQ= 78, normal (22%) M IQ=92 (p=0.013). Children with epileptiform discharges on EEG or abnormal background with no epileptiform activity had mean IQs (58, 60) significantly lower than children with a normal EEG (105; p = 0.006). Children with a history of infantile spasms (33%) had significantly lower IQ scores at follow up than children with no history (p= 0.043). Conclusions: 60% children with early onset epilepsy are developmentally delayed 3 years following initial presentation. Developmental status close to presentation predicts function at follow up. Aetiologically relevant structural brain abnormalities, epileptiform activity and abnormal background activity on initial EEG, and a history of infantile spasms were associated with significantly lower IQ scores.