The Gap Between the Randomized Controlled Trial–Based Evidence and Real-World Practice in Switching Strategies of Major Depressive Disorder

Abstract

To the Editor: Switching to an antidepressant in a different pharmacologic class, which is often employed in clinical practice for major depressive disorder (MDD) patients who have not responded sufficiently to an initial antidepressant, usually yields better clinical outcome. However, some randomized controlled trials (RCTs) and RCT-based meta-analyses designed to test the efficacy of switching therapies suggest that switching to another antidepressant may not be superior to continuing the initial antidepressant. 1 Three reasons account for the gap between the conclusions of traditional clinical trials and our clinical experience in the real-world setting.The first-and main-reason for this gap is the use of inappropriate inclusion criteria for nonresponders to initial antidepressant treatment. In a study by Bose et al, 2 depressive patients who did not respond (< 50% Montgomery-Asberg Depression Rating Scale improvement) to 2 weeks of 10 mg/d of escitalopram were randomly assigned to further treatment in a protocol evaluating the efficacy of 20 mg/d escitalopram versus switching to 60 mg/d duloxetine. In clinical practice, it is difficult to reduce MDD symptoms by 50% in 2 weeks with a moderate dose of antidepressant. Patients meeting such criteria for nonresponse are therefore not true nonresponders and could achieve significant improvement in either the antidepressant continuation group or the switching group. Similarly, in Souery and colleagues' study, 3 achieving 50% total score reduction on the Hamilton Depression Rating Scale after 4-week treatment would have been difficult. Additionally, one study 4 did not clearly define the concept of nonresponse, only indicating that some of the patients previously treated with a selective serotonin reuptake inhibitor (SSRI) were switched to duloxetine. Inaccurate definition of nonresponse has therefore impaired the superiority of switching strategies, and a considerable proportion of patients could see clinical benefit if the dose were escalated or therapy time were extended longer until the onset of antidepressant action occurred. The most recent metaanalysis 5 of the antidepressant dose-response relationship has also shown that SSRIs are more efficient at higher doses, so it makes sense that patients experience continuous improvement after dose escalation in broad analysis, which may weaken the clinical benefit of switching strategies.Second, doses of switching antidepressants should be comparable and flexible during the study. One trial compared the efficacy of continuation of venlafaxine 75-375 mg/d with that of switching to fluoxetine 25/50 mg/d. 6 According to Hayasaka and coworkers' findings, 7 fluoxetine 40 mg/d is equivalent to venlafaxine 149.4 mg/d. So, it is no surprise that venlafaxine at a higher dose is more effective than fluoxetine.Third, randomization and blinding were used in most studies cited by this meta-analysis, so we should pay attention to analogous "placebo effects" in such antidepressant switching studies. When nonresponders who had little...