Active Comparator

An active comparator is a control arm in which participants receive an established treatment rather than a placebo, so the effect estimate is a difference between two working drugs rather than a difference from zero. A placebo-controlled trial answers whether the drug does anything; an active-controlled trial answers which of two options does more. It also imposes a requirement placebo control does not, called assay sensitivity: the study must be capable of detecting a difference between the arms if one exists.

The incretin field is unusually rich in head-to-head work. SUSTAIN 7 ran semaglutide against dulaglutide over 40 weeks in type 2 diabetes and found greater glycated haemoglobin and weight reduction with semaglutide. SURMOUNT-5 compared tirzepatide with semaglutide 2.4 mg for obesity and reported roughly 20 percent versus roughly 14 percent mean weight reduction at 72 weeks.

For a prescriber this is usually the decision-relevant design, because the real question is rarely whether to treat but what to treat with. It also carries safety information a placebo arm cannot supply, since an adverse event rate becomes interpretable only beside the rate on the drug it would displace. What it cannot supply, when both arms improve, is any estimate of how much of that improvement either drug is responsible for.

The recurring error is assuming the comparator was chosen fairly. One given below its licensed maximum, by a less effective route, or in a population selected to favour the test drug will flatter the result. The larger version of the same mistake is cross-trial comparison, where percentages lifted from two separate placebo-controlled programmes with different populations, durations and endpoints get presented as though a head-to-head had been run.