The problems of taste in placebo matching: an evaluation of zinc gluconate for the common cold

Abstract

A recently published randomized, controlled, double-blind, study has reported efficacy of zinc gluconate lozenge treatment of the common cold. Because of a higher rate of side effects in the zinc group and the lack of any confirmatory evidence of blinding efficacy, we are concerned about the possible contribution of bias to the study result. Unflavored zinc gluconate tablets can be easily distinguished from the relatively tasteless, unflavored calcium lactate tablets, which were used as placebo in the cited study, increasing our concern about bias. In order to be able to test zinc lozenges in a second trial of common cold therapy, we developed a matching placebo containing denatonium benzoate (a bitter substance used to prevent thumb sucking in children) and then undertook a formal evaluation of blinding efficacy. This article reports the results of this evaluation, encompassing two randomized, controlled taste trials; the first trial of the placebo matching showed zinc recipients to exhibit a higher rate of side effects and also a significantly higher proportion of subjects believing they were on "active" treatment. Alterations in dosage before the second taste trial yielded comparability of the placebo and zinc lozenges. This study illustrates the pitfalls which may sometimes be encountered in performing and publishing studies in which placebos are either not tested at all or tested in too small a subject population to detect a significant subjective difference between the placebo and the active medication.