The purpose of the study was to examine the reported prevalence of mood and anxiety disorders along the adult lifespan in a number of European countries and to consider whether the low prevalence of these disorders often reported by standardised diagnostic instruments in later life was likely to be realistic. Methods Data from the European Union World Mental Health (EU-WMH) project, a cross-national study of 37,289 adults from ten European countries assessed using the Composite International Diagnostic Interview (CIDI) were analysed. A multiple group methodology was used to model all relationships with age, with the countries forming the groups. Analyses included logistic regression, multinomial regression, Poisson regression, latent class analysis and confirmatory factor analysis. This was a much wider examination of the suitability of multiple group methods for cross-national research than has traditionally been the case within psychiatric epidemiology. Sociodemographic variables were included as appropriate. Results In most countries adults over 80 years of age were unlikely to be diagnosed with any l2-month DSM-IV mood or anxiety disorder. The 12-month prevalence of individual disorders, disorder groupings, and latent classes of disorders typically declined with age among older adults in Western Europe and increased with age in Eastern Europe but this was not always the case. The burden associated with double depression increased with age in a number of countries. Older adults in Western Europe were those least likely to endorse any of the lifetime mood and anxiety screening questions. They were also those least likely to receive at least a lifetime diagnosis of major depressive episode or endorse the minimum episode duration requirements assessed at the start of the depression module. There were no significant relationships between the latent construct for major depressive episode, the symptom groups and age in any country. Conclusions Multiple group methods form an attractive strategy for the cross-national study of both observed and latent variables. The prevalence of mood and anxiety disorders in later life as reported by the CIDI does not correspond with other sources. It is debateable whether some of the reporting patterns observed among older adults, such as responses to lifetime screening questions 01' the minimum duration requirements, are likely to be realistic. We recommend that validity studies of the CIDI and sensitivity analyses of the screener specifically take place among older adults in a number of countries, not least within Eastern Europe.