This study was performed to test the hypothesis that there are 'hotspots', i.e. geographical heterogeneity, of dengue transmission. Data from two repeat serosurveys in two villages in Vietnam were used to identify incident infections and to relate these to prevalence at baseline and thus assess geographical heterogeneity, i.e. clustering, in dengue transmission. A total of 400 households were surveyed; serological data from 521 children at baseline and from 119 children at follow-up were included in a spatial analysis. Geographical heterogeneity of dengue transmission was explored using a permutation null distribution test. This showed for the first time evidence of clustering of dengue virus transmission at the household level in asymptomatic children. Risk areas could be identified by seroprevalence surveys combined with mapping. Control of dengue virus transmission could be supported by identification and control of hotspots.