These results should be interpreted in view of several limitations. First, AUD diagnoses came from population records, and the prevalence of AUD was lower compared to estimates from other epidemiological surveys (1, 40). However, there is a high degree of concordance among the registries (23) and the prevalence of ICD AUD in Sweden is similar to that observed here (41). Whether the pattern of genetic and environmental covariance for AUD and divorce generalizes to AUD cases identified through other ascertainment strategies is unknown. Second, because we could not disambiguate the relationship status of individuals who were cohabiting (unmarried) without children they were excluded from our analyses. Third, we included cohabiting couples with children together in our analyses, which may bias the AUD-divorce association. We addressed this possibility in supplementary analyses using strict definitions of marriage and divorce as defined by the Total Population Register. The latent genetic and environmental correlations were very similar to those obtained using the less strict definitions (results available upon request from the first author). Fourth, we limited our analyses to same-sex pairs, which precluded examination
