American Journal of Epidemiology

Dengue fever is caused by one of the four serotypes of the dengue virus and is transmitted by the urban mosquito Aedes aegypti. In 1993, the city of Charters Towers in the tropical north of Australia experienced an epidemic caused by the dengue 2 virus. A cross-sectional sample of 1,000 people was assessed for determinants of recent symptomatic dengue infection. After exclusion of people with prior exposure to dengue 2, a study group of 797 persons, including 196 patients with recent infection, were evaluated. Stepwise logistic regression analysis identified four determinants of infection: the presence of a case of dengue fever within two residential blocks (odds ratio (OR) = 3.61, 95 % confidence interval (Cl) 2.56-5.10), house screening (OR = 0.60, 95 % Cl 0.40-0.89), the presence of a water tank within two residential blocks (OR = 1.51, 95 % Cl 1.02-2.22), and the use of knockdown insecticide (OR = 1.75, 95 % Cl 1.22-2.51). Classification and Regression Tree analysis identified a group of 152 individuals in whom the prevalence of dengue infection was 50%. These people lived within two blocks of a suspected dengue fever case, did not have house screening, and used knockdown sprays. If dengue had not occurred within two residential blocks, there were no additional factors that significantly influenced the prevalence of dengue fever. Control of dengue epidemics should involve attempts to geographically contain the spread of infection, use of house screening, and the removal of mosquito breeding sites such as water tanks. Am J Epidemiol 1998; 148:1111-16