Meta-analysing the effectiveness of melatonin for sleep-disturbed individuals with autism spectrum conditions: should Rett syndrome be included?

Abstract

SIR–In a recent article, Rossignol and Frye provided a very complete review of melatonin in pervasive developmental disorders (PDD). In this work, the authors usefully explored the scientific literature about the physiopathology of melatonin in PDD, and reported the first meta-analytic results of melatonin use in sleep-disturbed individuals with PDD. Our purpose in this letter is to discuss the legitimacy of considering Rett syndrome when meta-analysing sleep parameters in autism spectrum disorders (ASD). Rett syndrome proceeds from a progressive encephalopathy of monogenic origin and although its early clinical picture (between 1 and 3y) can include autistic-like manifestations, it is very rarely considered in the specialized community as forming a pertinent part of autism spectrum disorders (ASD). Indeed, in the proceedings of the ‘Autism and other pervasive developmental disorders conference’ Rett syndrome is characterized as a ‘non-ASD that presents with PDD symptoms’; it will be removed from ASD in the forthcoming fifth version of the Diagnostic and Statistical Manual of Mental Disorders, and also probably from the psychiatric section of the International Classification of Diseases, 11th revision. Sleep problems in Rett syndrome generally appear in the latest phase (IV) of the disease, whereas sleep problems in autism generally begin during early childhood. Moreover, sleep problems which are characteristically seen in Rett syndrome are irregular sleep ⁄wake patterns, excessive daytime sleep, and night-time emotional behavior (screaming, crying, and laughing), whereas those seen in autism are settling difficulties, long sleep latency, frequent ⁄ long nocturnal awakenings, and early-morning awakenings. Thus, sleep disturbances in Rett syndrome and autism can be considered different in nature. For these reasons, we believe that autism spectrum conditions and Rett syndrome should be considered separately in sleep studies, and therefore that the data collected in Rett syndrome should not be included in a meta-analysis of melatonin effects on sleep in individuals with ASD. In the research of Rossignol and Frye, this point of controversy could be resolved by verifying the stability of the meta-analytic results when removing the Rett syndrome data. To conclude, in the abstract, the authors wrote that ‘Melatonin administration in ASD is associated with ... better daytime behavior ...’; however, as mentioned later in the article (p 7), only a single study has explored this topic in a placebo-controlled manner, which reports no improvement in daytime behaviour with melatonin compared to placebo. Thus, despite the fact that it is often seen in clinical practice and has been repeatedly reported in open studies, improvement in behaviour in children with ASD with melatonin has not thus far been demonstrated. This argues for further research about the efficiency of melatonin in sleep-disturbed individuals with ASD.