The accuracy of the diagnosis of paroxysmal events in children.

OBJECTIVE

To assess the accuracy of the diagnosis of epileptic seizures in children.

METHODS

The Dutch Study of Epilepsy in Childhood is a prospective hospital-based study of 881 children referred because of possible seizures. The diagnosis was based on predefined descriptive criteria, as applied by a panel of three pediatric neurologists. Children with a definite other diagnosis were excluded. All children with unclear events were followed up for 1 year and children with seizures were followed up for 2 years to assess the accuracy of the diagnosis.

RESULTS

In 170 of 224 children seen after a single event, the incident was classified initially as epileptic, in 54 as unclear. In none of the 170 children did the diagnosis prove to be wrong. In four of the 54 children, recurrent episodes enabled a definite diagnosis of epilepsy. In 412 of the 536 children seen with multiple events, an initial diagnosis of epilepsy was made. After follow-up, this initial diagnosis was probably incorrect in 19. In contrast, seven of 124 children with multiple unclear episodes at intake later received the diagnosis epilepsy.

CONCLUSIONS

A false-positive diagnosis of epilepsy was made in 4.6%, whereas a definite diagnosis of epilepsy or seizure was delayed in 5.6% of children with multiple unclear events and in 7.4% of children with one unclear event.