One senior neurologist (JBB) from the Department of Neurology and Pediatrics, West Virginia University, Morgantown, W.Va., USA, and three neurology residents from the University of Oklahoma, Harvard Unitersity, and W.Va. University reviewed descriptions of seizures transcribed from medical records and classified them according to the International League Against Epilepsy (ILAE) Classification system, 1981. The overall agreement in classifying seizure types by the ILAE system was relatively poor and only slightly better than would have been expected by chance. Agreement was particularly weak for atypical absence, partial seizures with secondary generalization, and generalized motor seizures. It was better for simple and complex partial, simple absence and infantile spasms. Approximately 22% of descriptions were insufficient for classification by the ILAE system. The authors conclude that the development and testing of more explicit criteria for the diagnosis of specific seizure types may be especially useful in improving the reliability of seizure classification. 
COMMENT. A classification of seizure types is of little value without assessment of its reproducibility in practice. Surprisingly, this study appears to be the first test of the reliability of the ILAE system of seizure classification, first proposed in 1970 and generally employed in statistical trials of anticonvulsant drugs. The relatively low level of concordance among observers using the ILAE system in the present study is disturbing, and the substitution of a more specific seizure classification should be entertained especially in therapeutic trials.