Abstracts

RATIONALE: Most studies consider remission as the ultimate and final seizure outcome in epilepsy. In fact, seizure outcome, especially early in the course of the disorder, is far more complicated and varied. We examined the phenomenon of repeated remission and relapse to better characterize this pattern of seizure outcome and factors associated with it.
METHODS: A prospectively identified cohort of children (0-15 y) with newly diagnosed epilepsy (initial N=613) is being intensively followed. The probability of 1y remission and subsequent relapse was examined for up to three remission periods per child. Product-limit and Cox methods were used for analysis. Cases were censored if they had insufficient follow-up (from study entry for the first remission or previous relapse for subsequent remissions) to achieve 1y remission.
RESULTS: During a median 6y follow-up, 524/602 (87%) of those followed [gte]1 year achieved a 1y remission, and 231 (44%) then relapsed. Of 206 followed [gte]1 year after the first relapse, 163 (79%) had a 2nd remission, and 64 (39%) of those relapsed again. 35 of 45 (78%) followed [gte]1 year after the 2nd relapse attained a 3rd remission of whom 16 (46%) relapsed. To facilitate comparisons across outcomes assessed over varying follow-up periods, product-limit (Kaplan-Meier) estimates were calculated. The probability of attaining a 1y remission within 2y of diagnosis was 64% (95% CI 62-66), but 23% (21-25) relapsed within 1y of remission. Within 2y of the first relapse, 75% (69-81) entered a 2nd remission, but 25% (18-32) relapsed again within 1y of attaining the 2nd remission. Finally, 81% (67-95) attained a 3rd remission within 2y of the 2nd relapse, but 34% (17-51) relapsed for a 3rd time within 1y of attaining the 3rd remission. In a Cox model of all possible remission episodes and relapses, those with cryptogenic (Rate Ratio (RR)=0.84, p=0.04) and symptomatic (RR=0.53, p[lt]0.0001) etiology were less likely to enter remission than those with idiopathic etiology. Compared to a first remission, a second remission (RR=1.23, p=0.02) and a third remission (RR=1.48, p=0.03) were achieved more quickly. After adjusting for the effect of tapering medication, symptomatic etiology (RR=1.84, p=0.0001), family history of epilepsy (RR=1.40, p=0.05), and age at onset [gt]12y (RR=2.03, p[lt]0.0001) were associated with an increased risk of relapse. Compared to relapse after a first remission, relapse may be more likely after a third (RR=1.65, p=0.06) but not after a second (RR=1.11, p=0.48) remission.
CONCLUSIONS: One-year remission is common in childhood-onset epilepsy; however, relapses are also common. A repeated remitting-relapsing pattern is not unusual. Etiology, family history, and age at onset explain some of this variation. The current findings suggest that there may be a group that is prone to repeated remission and relapse during the first several years after initial diagnosis. The long-term significance of such a pattern of outcome can only be appreciated with prolonged follow-up.
[Supported by: NIH-NINDS grant RO1-NS 31146]