Comparison of classifications of seizures: a preliminary study with 28 participants and 48 seizures

Our aim was to compare three available seizure classifications (SCs), namely, the international classification of epileptic seizures published in 1981 (ICES; Epilepsia 1981;22:489-50); the semiological seizure classification (SSC) by H. Luders, J. Acharya, C. Baumgartner, et al. (Epilepsia 1998;39:1006-13; Acta Neurol Scand 1999;99:137-41); and the proposal of a new diagnostic scheme for seizures (PDSS) by J. Engel, Jr. (Epilepsia 2001;42:796-803) published in 2001. The three SCs were compared with respect to diagnostic success rates, usefulness, and consistency by a large group of neurologists in this preliminary study.

After a training period, 28 blindfed participants with different levels of experience with epilepsy classified videos or written descriptions of 48 randomly selected seizures according to the three SCs. Definite diagnoses of the seizures were established based on all clinical, ictal/interictal EEG, and MRI data. All the participants answered a questionnaire concerning their preferences for SCs after the study.

The overall diagnostic success rates were 81.4% for ICES, 80.5% for PDSS, and 87.5%, for SSC. Various parameters concerning experience with epilepsy affected success rates positively, without reaching statistical significance, whereas experience with epilepsy surgery appeared to be a parameter significantly affecting the success rate in all SCs. In reliability analysis, Cronbach's alpha was 0.94 for ICES, 0.88 for PDSS, and 0.70 for SSC, all showing good agreement in the group. Nineteen reviewers chose SSC, eight chose ICES, and one chose PDSS as their preference in the questionnaire, completed after the end of the study.

The results of this preliminary study demonstrate that with proper training, physicians treating epilepsy patients can handle new SCs, and emphasize the need for revision of the current classification.