RESUMEN
BACKGROUND: After a first afebrile seizure, EEG in addition to cMRI is recommended for pediatric patients. Once indications requiring immediate treatment are excluded, it is of interest to determine if the results provide a prognostic tool for seizure relapses. METHODS: Patients aged between 1 month and 18 years who had a first afebrile seizure between 2006 and 2008 were retrospectively studied and monitored for another 48 months. RESULTS: Out of 248 patients, 62.5% had generalized and 36.3% focal seizures. 34.7% of the EEG results were pathological. 176 patients had a cMRI that showed in 23.3% probable epileptogenic lesions. 3 patients with benign cerebral tumours needed surgical therapy. In the following 48 months 29.4% of the children showed seizure relapses. There was a correlation between epileptic patterns in the EEG and further seizures (p=0.0001). However, the sensitivity of the EEG based diagnoses was 0.6, the specificity 0.78 and the positive predictive value 0.52. There was no correlation between epileptogenic lesions and the probability of seizure relapses. The sensitivity of the cMRI to this effect was 0.36, the specificity 0.74 and the positive predictive value 0.34. DISCUSSION: The EEG is superior to cMRI for predicting seizure relapses. The percentage of noticeable cMRI findings is high but this has low therapeutic relevance and is assumed to largely represent "incidental findings". It is important to question the value of MRI investigations for sedated small children except in the case of emergencies. The key question is whether the cMRI should be deployed to diagnose epilepsy, the probability of seizure recurrences or to classify the entity of a most likely epilepsy.