Aim: Several reviews have noted a huge increase in the rate of diagnosed autism spectrum disorders. The main aims of this paper are: 1) to use published empirical findings to consider whether the rise reflects a true increase in incidence, as distinct from the consequences of better ascertainment and a broadening of the diagnostic concept; and 2) to consider how epidemiological data may be used to test hypotheses about possible causal influences, using MMR and thimerosal as examples. Methods: Search of the literature for studies with a large epidemiological base population, systematic standardized screening, a focus on an age group for which diagnostic assessments are reliable and valid, and diagnosis by trained professionals using high-quality research assessments. Also, search of a broader literature to consider the evidence from all epidemiological studies with respect to the hypothesized causal effect of MMR and thimerosal on autism spectrum disorders. Results: The true incidence of autism spectrum disorders is likely to be within the range of 30-60 cases per 10 000, a huge increase over the original estimate 40 years ago of 4 per 10 000. The increase is largely a consequence of improved ascertainment and a considerable broadening of the diagnostic concept. However, a true risk due to some, as yet to be identified, environmental risk factor cannot be ruled out. There is no support for the hypothesis for a role of either MMR or thimerosal in causation, but the evidence on the latter is more limited.

Conclusion: Progress in testing environmental risk hypotheses will require the integration of epidemiological and biological studies.