In a prospective study, we have followed 347 children identified at th
e time of a first unprovoked seizure for a mean of 46 months. To date,
135 (39%) have experienced a seizure recurrence. In this study, we an
alyzed recurrence risk as a function of whether the child was asleep o
r awake at the time of the first seizure. The cumulative recurrence ri
sks for children whose first seizure occurred in sleep was 28% at 0.5
years, 39% at 1 year, 53% at 2 years, and 55% at 4 years, compared wit
h recurrence risks of 18%, 23%, 30%, and 35% at the same intervals in
children whose first seizure occurred while awake (p < 0.001). The ass
ociation of a first seizure during sleep with an increased recurrence
risk was present primarily in children with idiopathic seizures. It oc
curred in both those with a normal and an abnormal EEG. On multivariab
le analysis, sleep state, etiology, and the EEG were statistically sig
nificant predictors of recurrence risk. In children who experienced a
seizure recurrence, the recurrences occurred in the same sleep state i
n 73% of the cases (p < 0.0001). This was also true of subsequent recu
rrences. We conclude that the occurrence of a first seizure in sleep i
s associated with an increased risk of recurrence. Subsequent seizures
, if they do occur, usually occur in the same sleep state as the initi
al seizure.