Dr. Fava and colleagues have published an antidepressant trial on adjunctive Brexpiprazole, a novel atypical antipsychotic drug that was developed to treat schizophrenia, in the Journal of Clinical Psychiatry.
The trial consisted of 4 steps:
- The authors carefully selected 50 patients who had not shown improvements with their current antidepressant.
- These patients received 2 more weeks of treatment with their current antidepressant, to provide a baseline measure.
- For the next 6 weeks, patients additionally received Brexpiprazole, with a dosage from 1mg in week 1, 2mg in week 2, and 3mg in subsequent weeks.
- Finally, patients received 4 more weeks of antidepressant treatment without Brexpiprazole.
The authors were specifically interested in anger / irritability – that are common symptoms in depression, e.g. (1) – and analyzed a number of anger and irritability items and scales, as well as depression severity, global functioning, and side effects. They investigated changes from weeks 3 to 8 (the adjunctive Brexpiprazole phase), and then changes from weeks 9 to 13 (Brexpiprazole discontinuation phase).
There was no control / placebo group.
Fava et al. found that all measures of depression severity showed consistent improvement, whereas the analyses of anger or irritability scales were somewhat less conclusive. In the discontinuation phase, some of the symptoms worsened again, which leads the authors to conclude that “adjunctive treatment with brexpiprazole may represent a strategy for patients with MDD and inadequate response to antidepressant treatment”, especially for those with symptoms of irritability.
Like two other clinical trials with major shortcomings I reviewed here recently (1, 2), the main difficulty with this paper (apart from the small sample) is that there was no control group, i.e. a group that received placebo instead of adjunctive Brexpiprazole during weeks 3 to 8.
This means it is unclear whether any of the effects are actually due to Brexpiprazole, that is: what if we had given patients during weeks 3-9 vitamin pills and told them they get a very special and new antidepressant? Since placebo response is known to be very strong in depressed patients, with about 1 of 3 patients improving considerably under placebo (1, 2, 3), I fail to understand how clinical trials are conducted without a placebo group in 2016.
This is especially so when looking at the conflicts of interests and funders of the study – there certainly seem enough possibilities to obtain funding for a placebo group:
One could of course argue that it is unethical to have a control group, e.g. in highly suicidal patients, but the question is what an intervention study that does not have any form of control group actually contributes to science. Additionally, there is the possibility to first give the control group a placebo (while the experimental group receives adjunctive Brexpiprazole), and then afterwards also provide the control group with the medication. This not only allows for a proper comparison, but will also double the sample size of patients who receive the actual treatment.
It is of note that the longitudinal design of this study in which patients first received adjunctive Brexpiprazole which was the discontinued does not change this interpretation, and is entirely consistent with the hypothesis that the placebo effect explains the effects of adjunctive Brexpiprazole.
In sum, I don’t understand why studies without control groups still receive funding.
For the sake of completeness, I should mention that two prior studies have examined adjunctive Brexpiprazole in depressed patients who did not respond adequately to treatment as usual, and there is some evidence that it seems to improve patients’ well-being above that of placebo.
(1) 2mg adjunctive Brexpiprazole, published by Thase et al. 2015a in the Journal of Clinical Psychiatry:
Brexpiprazole (n = 175) reduced mean MADRS total score versus placebo (n = 178) at week 6 in the efficacy population per final protocol (-8.36 vs -5.15, P = .0002). Brexpiprazole improved SDS mean score versus placebo (-1.35 vs -0.89, P = .0349). The most common treatment-related adverse events were weight gain (brexpiprazole, 8.0%; placebo, 3.1%) and akathisia (7.4% vs 1.0%).
MADRS is a depression outcome measure, SDS a measure of adverse side-effect measure.
(2) 1mg and 3mg adjunctive Brexpiprazole, published by Thase et al. 2015b in the Journal of Clinical Psychiatry:
In the efficacy population per final protocol, brexpiprazole 3 mg (n = 213) showed a greater improvement in MADRS total score versus placebo (n = 203; -8.29 vs -6.33; P = .0079), whereas brexpiprazole 1 mg did not (n = 211; -7.64 vs -6.33; P = .0737). The brexpiprazole groups showed comparable improvement in SDS mean score versus placebo (least squares [LS] mean difference: [1 mg] -0.49, P = .0158; [3 mg] -0.48, P = .0191). The most frequent adverse events were akathisia (4.4%, 13.5%, 2.3%), headache (9.3%, 6.1%, 7.7%), and weight increase (6.6%, 5.7%, 0.9%) in brexpiprazole 1-mg, 3-mg, and placebo groups, respectively. Mean changes from baseline in Abnormal Involuntary Movement Scale (LS mean difference = 0.08, P = .0141) and Barnes Akathisia Rating Scale (LS mean difference = 0.17, P = .0001) total scores were significantly greater with brexpiprazole 3 mg versus placebo.