The live, attenuated yellow fever (YF) 17D vaccine provides highly effective and durable immunity and is widely used for travelers to and residents of endemic areas of South America and Africa. Neurotropic and viscerotropic serious adverse events associated with these vaccines occur rarely, but YF 17D vaccine-associated viscerotropic disease (YEL-AVD) is notable for its lethality. There appear to be two distinct patterns of risk for YEL-AVD: the first in younger persons, particularly women, with defects in innate immunity, in whom the case-fatality rate is higher; and the second in elderly persons, particularly men with age-related immune senescence and a lower case-fatality rate. From 1990 to the present, the number of cases (n = 31) and deaths (n = 12) from YEL-AVD in travelers has exceeded the reports of YF (n = 6) acquired by natural infection, raising the question whether the risk of vaccination exceeds the benefit in travelers. To provide some guidance on this point, the rate of vaccine-related injury is compared with the rate of naturally acquired disease in a new analysis that estimates the immunologically susceptible denominator population in YF endemic and epidemic areas. For many years, the risk of vaccine-related illness and death was similar to the risk of illness and death from natural infection with YF in South America. Africa posed a substantially higher estimated risk of wild-type YF than vaccine-related injury. Multiple factors should be considered in making decisions about YF vaccination, including specific destination, season of the year, local evidence for YF transmission, likelihood of exposure to vector mosquitoes and individual risk factors for YEL-AVD, with the goal of increasing vaccine coverage for travel to high-risk areas and reducing unnecessary vaccination. Prospects for future, safer vaccines are also described.