A novel postoperative seizure classification for long-term mortality of patients with intractable epilepsy: comparison with the Engel system.

Abstract

To explore the association of clinical seizure pattern with the long-term prognosis after epilepsy surgery is important for patient counseling and risk management. To study the long-term mortality after surgery of intractable epilepsy patients with a novel classification of clinical patterns by long-term frequency and duration of seizure compared with the Engel classification. We conducted a longitudinal cohort study by enrolling 280 patients with epilepsy who underwent anterior temporal lobectomy between 1987 and 2002. In addition to the Engel classification, we proposed a novel classification of clinical patterns pertaining to at least 1 year of follow-up of the frequency and duration of seizures after surgery (inactive, delayed, intermittent, and intensive groups). We followed the vital status of these members until 2007 to obtain a 15-year survival rate for each classification. The standardized mortality ratio and hazard ratios with proportional hazards regression model by the extent of severity were estimated. The overall 15-year survival rate for patients who underwent anterior temporal lobectomy was 95.1%. The standardized mortality ratio estimates (95% confidence interval) for our classification showed a gradient relationship from 0.6 (0.1-2.3), 2.5 (0.7-6.3), 7.6 (0.8-27.3), and 8.9 (3.2-19.3) for inactive, delayed, intermittent, and intensive groups, respectively (Trend test, P = .04), whereas the corresponding estimates were 0.8 (0.2-2.2), 5.9 (1.2-17.2), 6.7 (2.5-14.7), and 7.2 (0.8-25.9) for Engel I to IV, respectively, which showed a less increasing trend (Trend test, P = .82). Similar findings were noted for hazard ratios for the 2 classifications. The proposed novel classification with long-term observed frequency and duration of seizures after surgery is more informative for predicting long-term mortality than the Engel classification.