CRD Summary: The authors concluded that naltrexone may benefit children with autistic disorder, particularly those with self-injurious behaviour, if other attempted therapies have failed. Given the small sample sizes and unknown quality of the included studies, the reliability of the results is unknown.

CRD Commentary: The inclusion criteria were clearly reported in terms of the participants and intervention, but remained broad for the outcomes. The authors searched two databases and screened the references of included studies. However, it is not clear if non-English papers and unpublished studies were sought, thus potentially introducing language and publication bias. The authors did not state how many reviewers were involved in selecting the studies and extracting the data, thereby introducing the potential for reviewer bias. In addition, the validity of the studies was not formally assessed. While the authors stated that some of the studies were well-designed, it is not possible to determine the overall quality of the included studies. There may also be some overlap in participants between some included studies, thus introducing the potential for bias.

The authors’ decision to describe the results in a narrative synthesis was appropriate given the heterogeneity between the studies. Details of the studies were presented in the tables and text of the review. Given the small sample sizes and unknown quality of the included studies, the reliability of the results is unknown.