Design of clinical trials of antidepressants: should a placebo control arm be included?

There is no doubt that available antidepressants are efficacious and effective. Nevertheless, more effective drugs with improved tolerability are needed. With this need in mind, some protagonists claim that future antidepressants should be proved superior to, or at least as effective as, established antidepressants, making placebo control methodologically dispensable in clinical trials. Moreover, the use of placebo control is criticised as unethical because it might result in effective treatment being withheld. There are, however, a number of methodological reasons why placebo control is indispensable for the proof of efficacy of antidepressants. Comparing investigational antidepressants only with standard antidepressants and not placebo yields ambiguous results that are difficult to interpret, be it in superiority or equivalence testing, and this method of assessment requires larger sample sizes than those required with the use of placebo control. Experimental methodology not adhering to the optimal study design is ethically questionable. Restricting the testing of investigational antidepressants only to superiority over standard antidepressants is an obstacle to therapeutic progress in terms of tolerability and the detection of innovative mechanisms of action from which certain subgroups of future patients might benefit. The use of a methodology that requires larger samples for testing of superiority or equivalence is also ethically questionable. In view of the high placebo response rates in trials of antidepressants, placebo treatment does not mean withholding effective treatment. Accepting the necessity of the clinical evaluation of new, potentially ineffective antidepressants implicitly means accepting placebo control as ethically justified. Three- or multi-arm comparisons including placebo and an active reference represent the optimal study design.