Purpose: To determine how the onset of unprovoked seizures influences recurrence of seizures with fever in children followed from the time of a first febrile seizure.
Methods: In a prospective cohort of children (n = 428) identified at the time of a first febrile seizure, predictors of a second seizure with fever were identified. The occurrence of a first unprovoked seizure was treated as a time-dependent covariate in a Cox regression model rather than as a censoring point as it traditionally has been in the past.
Results: One hundred forty-three (33.4%) children had further seizures. Seven had further seizures with fever only after onset of unprovoked seizures. After adjustment was made for the four previously described predictors of recurrent febrile seizures (age at onset, family history, height of fever, and duration of fever), the onset of unprovoked seizures was associated with a rate ratio of 3.47 (p = 0.0015), indicating a large increase in the risk of further seizures with fever after onset of unprovoked seizures.
Conclusions: Young children who develop unprovoked seizures after a febrile seizure are at substantial risk for further seizures with fever. This may represent part of the spectrum of benign febrile seizures or it may represent the so-called "epilepsy triggered by fever" spectrum. It affects only a small proportion of children with febrile seizures; however, in some children, it may be useful information to consider when making treatment decisions.