Abstracts

Rationale: It has been reported that the lifetime prevalence of suicidal ideation in people with epilepsy is 25% compared with 13% in those without epilepsy. Patients with epilepsy also have increased lifetime prevalence of mood disorder and anxiety disorder. In 2005, the FDA asked sponsors of 11 anti-epileptic drugs (AEDs) to provide data from placebo (PBO)-controlled clinical trials in order to perform a meta-analysis of a potential association between AEDs and suicidal ideation and behavior. This analysis, consisting of 199 PBO-controlled clinical studies, was reported by the FDA on January 31, 2008 and demonstrated an incidence of suicidal ideation or behavior in 0.43% of patients treated with an AED versus 0.22% of patients treated with PBO. From the FDA’s reported analysis, patients with epilepsy had a higher relative risk for suicidality compared to patients who were given an AED for psychiatric or other conditions. GlaxoSmithKline has conducted an independent pooled analysis of its lamotrigine (LTG) clinical trials for suicidal ideation and behaviors, and the results are reported below. Methods: This pooled analysis comprised data from 35 LTG double-blind, PBO-controlled clinical studies with ≥20 patients/arm (17 neurological: 13 epilepsy and 4 neuropathic pain; 17 psychiatric: 12 bipolar, 3 unipolar and 2 schizophrenia; 1 healthy volunteers). A single blinded assessor at Columbia University independently classified Definitive Suicidal Ideation and Behavior (DSIB) events. Results: Overall, DSIB events occurred in 43/3695 (1.16%) LTG vs 25/2824 (0.89%) PBO patients; Odds Ratio (OR) (95% confidence interval (CI)) = 1.46 (0.89, 2.45), p=0.171. In epilepsy studies, DSIB events occurred in 6/1073 (0.56%) LTG vs 2/805 (0.25%) PBO patients; OR (95% CI) = 2.00 (0.40, 14.96), p=0.478. In bipolar studies, DSIB events occurred in 29/1212 (2.39%) LTG vs 19/1054 (1.80%) PBO patients; OR (95% CI) = 1.31 (0.73, 2.39), p=0.460. In the LTG treated patients, suicidal ideation events occurred more frequently than behavior events in both the epilepsy (4 vs 2) and bipolar (19 vs 10) studies. In the analysis of all indications, there was no difference in time to suicidal ideation or behavior events in patients taking LTG as compared to PBO. Conclusions: The pooled analysis across indications, as well as the analysis of epilepsy and bipolar subgroups, showed the rate of events was numerically, but not statistically significantly, greater for LTG compared with PBO. Suicidal ideation events occurred twice as often as behavior events in both epilepsy and bipolar studies. Definitive interpretation of the results of this analysis is difficult due to the small incidence and absolute numbers of events, the retrospective nature of such an analysis and potential confounding because the events of interest can be symptomatic of both epilepsy and bipolar disorder.