Vaginal microbicides, designed to prevent HIV infection in women, are one of the most promising biomedical interventions. Clinical trials of second-generation microbicides have begun; if shown to be effective, they could be licensed within 5–10 years. Because these microbicides contain antiretrovirals (ARVs), they could be highly effective. However, there is concern that, if used by HIV-positive women, ARV resistance may evolve. By analyzing a mathematical model, we find that adherence could have both beneficial and detrimental effects on trial outcomes. Most importantly, we show that planned trial designs could mask resistance risks and therefore enable high-risk microbicides to pass clinical testing. We then parameterize a transmission model using epidemiological, clinical, and behavioral data to predict the consequences of wide-scale usage of high-risk microbicides in a heterosexual population. Surprisingly, we show that reducing a participant's risk of resistance during a trial could lead to unexpectedly high rates of resistance afterward when microbicides are used in public health interventions. We also find that, paradoxically, although microbicides will be used by women to protect themselves against infection, they could provide greater benefit to men. More infections in men than in women will be prevented if there is a high probability that ARVs are systemically absorbed, microbicides are less than ∼50% effective, and/or adherence is less than ∼60%. Men will always benefit more than women in terms of infections prevented per resistant case; but this advantage decreases as the relative fitness of drug-resistant strains increases. Interventions that use ARV-based microbicides could have surprising consequences.