Using maternally reported data to investigate the association between early childhood infection and autism spectrum disorder: the importance of data source

Childhood infections have been found to be associated with autism spectrum disorder (ASD) in previous studies using hospital data or medical records to identify infections. We aimed to replicate these findings using maternal reports of childhood infection.We used the Danish National Birth Cohort consisting of 92 583 live singletons born from 1997 to 2003 in Denmark. ASD diagnoses were retrieved from the Danish Psychiatric Central Register, and a total of 945 children from the cohort were diagnosed with ASD. Data were analysed using Cox proportional hazards regression. We studied the association between ASD and maternal reports of infectious disease in the child from birth to 19 months. Furthermore, we performed secondary analyses using hospital registers to investigate the association between ASD and hospital contact in general as well as hospital contact for various infections.We did not find a general association between maternal reports of infectious illness and ASD. However, hospital contact for all causes was associated with an increased risk for an ASD diagnosis. Danish children with ASD do not appear to have a general pattern of illness from infection in early life, but do have more contact with medical specialists for infections and other indications compared with the general population.[corrected] Hospital data should be used cautiously when studying the co-morbidity of ASD; if the increased rate of hospital contact overall for children with ASD is not considered, then misleading interpretations might be made of observed associations between specific diseases and ASD.