Abstracts

Epilepsy surgery offers the possibility of freedom from seizures. Most information about prediction of surgical outcome is from small studies done at single centers, often from the pre-MRI era. The definitions of seizure outcome are often eclectic. Information based upon large, prospectively assembled patient series that have been carefully evaluated under similar protocols is necessary to obtain precise estimates of postsurgical seizure outcome and predictors of outcome.
In a seven-center study, 396 individuals underwent resective surgery (348 had temporal resections - ATL) for intractable focal epilepsy. They were enrolled prospectively at the time of initial surgical evaluation (1996-2000) and are being followed (median followup=3.5 years). Remission was defined as no seizures for [ge]2 years. Relapse was the occurrence of any seizure after 2-year remission. Survival techniques are used for analysis.
Of the 348 ATL patients 262 (75%) have been followed [ge]2 years and are thus eligible for this analysis. A total of 169 (65%) had a 2-year remission. The probability of attaining a 2-year remission at 2, 3, and 4 years after surgery was 49% (95% CI:43-55%), 62% (56-68%), and 67% (61-73%). Unilateral hippocampal atrophy (HA) [103/142 (73%) with versus 45/87 (52%) without (p=0.001)] and history of febrile seizures (FSZ) [57/79 (72%) with versus 112/183 (61%) without (p=0.09)] were associated with increased remission. Generalized tonic clonic seizures (GTC)[120/201 (60%) with versus 49/61(80%) without p[lt]0.01] were associated with decreased remission. There was no significant effect of age at onset. In a multivariable proportional hazards model, HA (Rate Ratio (RR)=1.6 (1.1- 2.2) and GTC (RR= 0.6 (0.4-0.9) were independently associated with remission.
Relapses occurred in 34/169 (20%) of those who remitted. The probability of relapse was 12% (CI=7-17%) and 24% (CI=15-32%) at 1 and 2 years after remission. No predictors of relapse were identified in this cohort.
Two thirds of ATL patients achieve a prolonged remission. HA, FSZ, and young age at onset, have been individually identified as important predictors of outcome in previous smaller studies. Our study is sufficiently large to examine these factors simultaneously with the result that unilateral HA appears to be the single most important factor and explains away virtually all of the effects of the FSZ. The findings for GTC were reported in two previous independent series analyzed by our group (Berg et al., Epilepsy Research 1998; 29:185-194) and may reflect diffuse and bilateral involvement. A substantial proportion of patients relapse. Further investigations will focus on the role of drug withdrawal in relapse after remission.
[Supported by: NIH NS32375]