TREATMENT EFFECT HETEROGENEITY IN RANDOMIZED PLACEBO-CONTROLLED TRIALS OF LAMOTRIGINE FOR ACUTE BIPOLAR DEPRESSION

Abstract

Lamotrigine is used to treat depressive episodes in bipolar disorder, but evidence from clinical trials has been inconsistent. There is some evidence that certain patient subgroups may benefit more than others. The goal of this thesis was to further consider the possibility of heterogeneous treatment effects by conducting a pooled analysis of five randomized, double-blind, placebo-controlled trials of lamotrigine for acute bipolar depression. The results are presented in two manuscripts. In the first study, a prespecified analysis was conducted to determine if patients with melancholic depression were more responsive than nonmelancholic patients. A subgroup with higher scores on depression scale items representing melancholic features had numerically larger treatment effects, but interaction tests were nonsignificant, and a melancholic diagnosis from structured clinical interviews was not associated with larger effects. Furthermore, baseline depression severity was inconsistently associated with response depending on which depression scale was used to define severity. The second study was an exploratory analysis that examined treatment effects on individual depression scale items and then attempted to use these items to create subgroup variables that could identify patients with larger overall treatment effects. On two depression scales, there were larger and statistically significant effects on items representing depressed mood/sadness, decreased interest/anhedonia, fatigue/anergia, and pessimism/guilt. The items with larger effects tended to be more prevalent in the sample at baseline, suggesting a floor effect limited the sensitivity of other items. Patients with higher scores on the mood/sadness and interest/anhedonia items also tended to have larger overall treatment effects compared to the rest of the sample. Taken together, in contrast to previous research, the results did not clearly support the hypothesis that patients with melancholic depression benefit more from lamotrigine compared to placebo, nor did baseline depression severity moderate treatment effects in a straightforward manner. Consistent with previous research, targeted assessments of core depressive symptoms that were more prevalent in the sample at baseline appeared to be more sensitive to change compared to total depression sum scores. More research is needed to establish the precise role of lamotrigine in the treatment of bipolar disorder.