There is a scarcity of adult prevalence studies of OSA outside the Caucasian populations of North America, Europe and Australia, and comparisons have been complicated by methodological differences in sleep study settings, respiratory events definition, measured risk factors and clinical outcomes, and the lack of objective parameters for the measurement of ethnicity itself. Comparing studies with the same methodological design and respiratory events definition, recent large-scale prevalence studies from Hong Kong, Korea and India show similar OSA rates to populations of mainly Caucasian composition. OSA is a very complex disorder determined by several phenotypes such as obesity, craniofacial structure and abnormalities in neuromuscular and ventilatory control. Genetics may partially explain some of the ethnic clustering of these phenotypes, modulated by cultural and environmental factors. The exact contribution of these component phenotypes to overall OSA risk will be determined by their varying prevalence and relative risk conferred across ethnic groups. For lesser degrees of obesity, Asians are at risk for a more severe degree of illness compared with Caucasians. Inter-ethnic studies suggests that African–American ethnicity may also be a significant risk factor for OSA. The increased prevalences of OSA among American Indians and Hispanic adults, and increased severity among Pacific Islanders and Maoris, were mainly explained by increased obesity parameters. Most cephalometric studies have largely been conducted without specific regard to ethnicity and comparisons of findings across studies have been mainly limited by differences in sampling methods and the varying selection and definition of measured cephalometric variables. The limited number of studies with inter-ethnic comparative data suggest cephalometric variables and their degree of contribution to OSA vary across ethnic groups.