The epidemiological factors present in Cuba in 1981, when the dengue haemorrhagic fever/dengue shock syndrome (DHF/DSS) epidemic occurred, were exceptional when compared to those of other countries in the region. Evidence is presented which demonstrates that virulence of the circulating strain is an important element in the analysis of an epidemic. Although the two current hypotheses to explain the occurrence of DHF/DSS epidemics are valid in well defined but different epidemiological situations, neither Halstead's hypothesis of secondary-type infection or Rosen's hypothesis of the role played by the virulence of the circulating strain can explain all cases. An integrated, multifactorial and unifying hypothesis is presented, which could be applied in different epidemiological situations. It is based mainly on an in-depth analysis of the literature and of the Cuban experience.