Genetic epidemiological studies of Autism Spectrum Disorders (ASDs) based on twin pairs ascertained from the population and thoroughly assessed to obtain a high degree of diagnostic validity are few. All twin pairs aged 3-14 years in the nationwide Danish Twin Registry were approached. A three-step procedure was used. Five items from the "Child Behaviour Checklist" (CBCL) were used in the first screening phase, while screening in the second phase included the "Social and Communication Questionnaire" and the "Autism Spectrum Screening Questionnaire". The final clinical assessment was based on "gold standard" diagnostic research procedures including diagnostic interview, observation and cognitive examination. Classification was based on DSM-IV-TR criteria. The initial sample included 7,296 same-sexed twin pairs and, after two phases of screening and clinical assessment, the final calculations were based on 36 pairs. The probandwise concordance rate for ASD was 95.2 % in monozygotic (MZ) twins (n = 13 pairs) and 4.3 % in dizygotic (DZ) twins (n = 23 pairs). The high MZ and low DZ concordance rate support a genetic aetiology to ASDs.