Exploratory analyses of effect modifiers in the antidepressant treatment of major depression: Individual-participant data meta-analysis of 2803 participants in seven placebo-controlled randomized trials

Abstract Background It is clinically important to know who are likely to respond more or less to antidepressants. However, meaningful effect modifiers (variables associated with differential response depending on the treatment) are yet to be identified. Methods We conducted individual participant data (IPD) meta-analysis and meta-regression to explore effect modifiers in placebo-controlled antidepressant trials conducted so far in Japan. Results We obtained access to IPD from seven placebo-controlled trials comparing bupropion, duloxetine, escitalopram, mirtazapine, paroxetine or venlafaxine with placebo in the acute phase treatment of major depression (total n  = 2803). The higher the guilt subscale score of the baseline Hamilton Rating Scale for Depression (HRSD), the greater the difference in reduction in depression severity between the antidepressants and placebo at week 6, while the older the current age or the age at onset, the smaller the difference. At week 8, the guilt subscale score of HRSD and presence of suicidal ideation at baseline predicted greater, and the anhedonia subscale and insomnia subscale scores of HRSD and early response at week 2 predicted smaller, difference in reduction. Limitations Different studies measured different sets of baseline variables and we were able to analyze only a limited set of candidate variables for effect modification. Conclusion Age, age at onset, several HRSD subscales including guilt, anhedonia and insomnia, presence of suicidal ideation at baseline and early response are potential effect modifiers for response to antidepressants in the acute phase antidepressant treatment of major depression. Future trials need to measure these and additional variables in concerted efforts to enable matching of treatments with individual characteristics in depression.