There is debate in the current literature regarding the permanence of an Autism Spectrum Disorder (ASD) diagnosis. We undertook a systematic review of the diagnostic stability of ASD to summarise current evidence. A comprehensive search strategy was used to identify studies. Participants were children with ASD. Risk of bias was assessed by examining the sample selected, recruitment method, completeness of follow up, timing of diagnosis and blinding. Twenty three studies assessed diagnostic stability with a total of 1466 participants. Fifty three to 100% of children still had a diagnosis of Autistic Disorder (AD) and 14-100% of children still had a diagnosis of another form of ASD at follow up. There is some evidence that Autistic Disorder is a reasonably stable diagnosis; however a significant minority of children will no longer meet diagnostic criteria after a period of follow up, particularly those diagnosed in the preschool years with cognitive impairment. Other Autism Spectrum Disorders have very variable stability between studies and clinicians when using this diagnosis need inform parents of its instability. This study supports the stricter diagnostic criteria in DSM-V. There is a need for long term, large population cohort studies measuring diagnostic stability. (Contains 3 tables and 2 figures.)