Using a set of statistical methods and HIV mathematical models applied on nationally representative Demographic and Health Survey data, we characterized HIV serodiscordancy patterns and HIV transmission dynamics in stable couples (SCs) in four countries: Cambodia, the Dominican Republic, Haiti, and India. The majority of SCs affected by HIV were serodiscordant, and about a third of HIV-infected persons had uninfected partners. Overall, nearly two-thirds of HIV infections occurred in individuals in SCs, but only about half of these infections were due to transmissions within serodiscordant couples. The majority of HIV incidence in the population occurred through extra-partner encounters in SCs. There is similarity in HIV epidemiology in SCs between these countries and countries in sub-Saharan Africa, despite the difference in scale of epidemics. It appears that HIV epidemiology in SCs may share similar patterns globally, possibly because it is a natural 'spillover' effect of HIV dynamics in high-risk populations.