Cligosiban for Premature Ejaculation: Success, Failure, or Insufficiently Tested?

The authors of the 2 papers in the Journal of Sexual Medicine on cligosiban for premature ejaculation (PE)1,2 are to be commended for their candor in disclosing the phase IIb failure of the agent promptly. They discussed the unexpected failure of the second, larger study after stating that “the 2 studies were almost identical, the only notable differences being the number of groups and the study drug formulation.”2 They went on to show that neither factor seemed to relate to the outcomes. However, this conflict of results does have a likely explanation. The authors failed to note another notable difference: the proportion of patients expecting to receive placebo. The proportion receiving placebo in the proof-of-concept (POC) trial was 1/3 (33%); for the phase IIb trial, the proportion was 1/5 (20%). Positive expectation of improvement in a clinical trial is often called “placebo effect” (whether a placebo is actual given or not). It is the bane of interpreting multi-arm trials with subjective or error-prone outcomes, as in sexual medicine. It has been studied in-depth in an older therapeutic area in which outcomes are subjective: psychopharmacology. The proportion of patients to be assigned to placebo—which currently must be disclosed prospectively to subjects before they give informed consent to enroll—has a substantial effect on outcome.3 A mental balance between optimism and pessimism is obviously required for subjects to deal with the possibilities of placement on placebo or active drug in blinded trials. Thus, this balance issue should be most problematic in trials of depression (which is almost always intimately mixed with anxiety), and primary anxiety disorders. The success rate of antidepressant trials with 2 equal-sized groups, 1 of them being placebo, was 60% in a large review of U.S. Food and Drug Administration (FDA)-approved agents; the success rate of antidepressant trials with 3, 4, or 5 equal-sized groups was dramatically lower, 32%.4 The value of using a 1:1 ratio of assignment to verum and placebo has been verified in trials for depression, and extended to migraine, schizophrenia, and Parkinson’s Disease.5