Zika and dengue viruses are closely related flaviviruses, with immunological interactions and identical urban, mosquito-borne transmission. Therefore, the recent introduction of Zika virus into the Americas and large-scale exposure of a uniformly previously unexposed population could affect subsequent transmission of dengue virus. This hypothesis had been untested, largely because sufficient epidemiological data were not available from affected locations. We explored this hypothesis in Salvador, the fourth largest city in Brazil (population 2·9 million), where extensive transmission of dengue viruses 1–4 occurred before the introduction and spread of Zika virus in 2015.