Sudden unexpected death in epilepsy in lamotrigine randomized-controlled trials

Authors


Address correspondence to Robert Leadbetter, GlaxoSmithKline, Mail stop 17.1315C, 5 Moore Drive, Research Triangle Park, NC 27709, U.S.A. E-mail: robert.a.leadbetter@gsk.com

Summary

Purpose:  Nonrandomized studies of the relationship of antiepileptic drugs (AEDs) with sudden unexpected death in epilepsy (SUDEP) may be susceptible to confounding by tonic–clonic seizure frequency, polypharmacy, and other potential risk factors for SUDEP. We evaluated the risk of SUDEP with lamotrigine (LTG) compared to active comparators and placebo in randomized controlled clinical trials conducted by GlaxoSmithKline (GSK) between 1984 and 2009.

Methods:  Among 7,774 subjects in 42 randomized clinical trials, there were 39 all-cause deaths. Ten deaths occurred >2 weeks after discontinuation of study medication and were excluded. Narrative summaries of deaths were independently reviewed by three clinical experts (TT, LH, DF), who were blinded to randomized treatment arm. The risk of definite or probable SUDEP was compared between treatment arms for each trial type (placebo-controlled, active-comparator, crossover), using exact statistical methods.

Key Findings:  Of 29 on-treatment deaths, eight were definite/probable SUDEP, four were possible SUDEP, and 17 were non-SUDEP. The overall, unadjusted rate of definite/probable SUDEP for LTG was 2.2 events per 1,000-patient years (95% confidence interval [95% CI] 0.70–5.4). The odds ratios (OR) for on-treatment, definite/probable SUDEP in LTG arms relative to comparator arms, adjusted for length of exposure and trial, were the following: placebo-controlled, OR 0.22 (95% CI 0.00–3.14; p = 0.26); active-comparator, OR 2.18 (95% CI 0.17–117; p = 0.89); and placebo-controlled cross-over, OR 1.08 (95% CI 0.00–42.2; p = 1.0).

Significance:  There was no statistically significant difference in rate of SUDEP between LTG and comparator groups. However, the CIs were wide and a clinically important effect cannot be excluded.

Ancillary