Some design issues in trials of microbicides for the prevention of HIV infection

(See the editorial commentary by Padian, on pages 663–5.) Trials for the prevention of human immunodeficiency virus (HIV) infection that evaluate microbicides provide significant design challenges. Three of these design issues deserve more careful consideration. The first issue relates to the benefits of using both blinded and unblinded control groups when the placebo regimen may not be inert and when the effectiveness of an intervention heavily depends on behavioral, as well as biological, factors. The second issue relates to the strength of evidence required for regulatory approval for the marketing of drugs and biologics when only a single pivotal phase 3 clinical trial has provided such evidence. The third issue relates to the appropriate next step after the completion of phase 1 trials, as well as the specific merits of conducting phase 2b screening trials that assess the effects on the same clinical efficacy end point that will be the primary end point in a phase 3 trial. The issues considered in microbicide trials for the prevention of HIV infection are also of importance in many other clinical scenarios. In 2002, 111,000 new cases of HIV infection occurred among adults each day, with the majority of these cases transmitted through heterosexual contact [1]. Although sexual transmission of HIV can be avoided throug