Getting it wrong

In a case–control study of 10–12-year-old children with either autism, special educational needs or normal development, Baird et al 1 examined measles antibody responses (plaque reduction neutralisation assay) and the presence of measles virus in peripheral blood mononuclear cells (reverse transcriptase polymerase chain reaction). The study apparently sought to identify autistic children relevant to the original MMR/autism hypothesis, that is, those with both regression and significant gastrointestinal symptoms. The study is severely limited by case definition in the context of the crucial “possible enterocolitis” group. For inclusion in this group the child had to have two or more of the following five current gastrointestinal symptoms (plus past persistent diarrhoea of >14 days’ duration excluding current constipation): Over the last 10 years we have evaluated several thousand children on the autistic …