Objective. To understand the possible effect that length of time has on disease severity with sequential dengue infections. Methods. Death and hospitalization rates for dengue hemorrhagic fever/dengue shock syndrome (DHF/DSS) per 10 000 secondary dengue-2 infections were compared in the same age group for two dengue-2 (DEN-2) epidemics in Cuba. The first DEN-2 epidemic affected all of Cuba in 1981; the second one, in 1997, impacted only the city of Santiago de Cuba. The sensitizing infection for DHF/DSS for each of the DEN-2 epidemics was dengue-1 (DEN-1) serotype virus, which was transmitted in 1977-1979, that is, 4 years and 20 years before the two DEN-2 epidemics. Using published seroepidemiological data from the cities of Havana and Santiago de Cuba, we estimated the rates at which persons aged 15-39 years old and those 40 years and older were hospitalized or died of DHF/DSS in Havana and in all of Cuba in 1981 and in just Santiago de Cuba in 1997. Results. Among adults 15-39 years old the death rate per 10 000 secondary DEN-2 infections was 38.5 times as high in Santiago de Cuba in 1997 as in Havana in 1981. As a further indication of the increased severity coming with a longer period between the initial DEN-1 infection and the secondary DEN-2 infection, the case fatality rate for that same age group was 4.7 times as high in Santiago in 1997 as it was in Havana in 1981. Conclusion. We found a marked increase in severity with the longer of the two intervals (20 years) between an initial DEN-1 infection and a secondary DEN-2 infection. Such a difference may be due to subtle shifts in causative dengue strains or to changes with the passage of time in the circulating population of human dengue antibodies. These observations have important implications for dengue control, pathogenic mechanisms, and vaccine development.