[Antidepressants and their onset of action: a major clinical, methodological and pronostical issue].

Although antidepressant medications are effective in about 50-70% of patients with major depressive disorder (MDD), they have a delayed onset of therapeutic effect. This latency is one of the current major limitations of these medications, in that it prolongs the impairments associated with depression, leaves patients vulnerable to an increased risk of suicide, increases the likelihood that a patient will prematurely discontinue therapy, and increases medical costs associated with severe depression. It is becoming increasingly clear that differences may exist between antidepressants and some evidence suggests that some antidepressant agents may begin to work faster than others. Escitalopram, duloxetine, venlafaxine, and mirtazapine have shown statistically significant differences in some measures of antidepressant action within the first two weeks of treatment, both in placebo-controlled trials and in head-to-head comparisons with other antidepressants. Results of the current review should be regarded with certain important limitations in mind. First, differences in times to onset of antidepressant response have been shown in clinical efficacy studies not specifically designed to detect differences in onset of action (post-hoc analysis). Second, results observed in 'pure' clinical trial samples should not be directly generalized to the real clinical practice since it has been proven in clinical settings that less than one in seven depressed patients would be eligible to participate in antidepressant clinical trials. For instance, depressed patients who are suicidal or who score higher than 30 on the 17-item HAM-D are excluded from antidepressant clinical trials. Third, caution is warranted when applying these findings to clinical populations with more severe depressions with respect to the fact that among clinical populations, severity of depression coincides with comorbidity, including such psychiatric disorders as anxiety disorders, personality disorders and substance abuse. In addition, the magnitudes of the size-effects of antidepressants versus placebo are clearly higher in severely depressed patients. Fourth, specific items on depression rating scales may induce greater antidepressant/placebo differences. For instance, the 17-item HAM-D contains three questions pertaining to sleep. It questions the fact that earlier onset may appear not only via a specific antidepressant effect but also via a non-specific effect on anxiety, sleep, physical pain or other accessory symptoms. Thus, current data do not clearly support claims that one drug reduces the symptoms of depression faster than another, though the existing literature suggests that escitalopram displays some superiority in terms of rapidity of action. Given the potential benefits of early-acting antidepressant treatments, the possibility of superior speed of onset of escitalopram presented here merits further study in adequately designed, prospective clinical trials. A definitive demonstration of early onset of action awaits the results of appropriately designed and powered clinical studies, which may include (1) a prospective definition of early onset of action, (2) more focused assessments of core emotional symptoms and cognitive deficits of depression by using specific and sensitive tools, (3) a data-analytic approach capable of capturing the dynamic nature of symptomatic change (for example, survival analysis), and (4) strategies to minimize biases and heterogeneity of response.