Relationship of Serum Ferritin Levels to Sleep Fragmentation and Periodic Limb Movements of Sleep on Polysomnography in Autism Spectrum Disorders

OBJECTIVE: Although children with autism spectrum disorders experience a range of sleep disturbances, exact mechanisms are not well-characterized. We investigated the association of serum-ferritin to sleep fragmentation and periodic limb movements of sleep using polysomnography in children with autism spectrum disorders. METHODS: We conducted a retrospective chart review of children with autism spectrum disorders followed from 1990 to 2010. Inclusion criteria were availability of polysomnography data and ferritin levels within 12 months of each other. The following variables on polysomnography characterized sleep fragmentation: increased arousal index, alpha intrusions, and reduced sleep efficiency. The data were compared with age- and gender-matched controls. RESULTS: Of 9791 children with autism spectrum disorders identified, 511 had a ferritin level, 377 had polysomnography data, and 53 had both ferritin and polysomnography data. As compared with the controls (86 ng/mL), the median ferritin level was 27 ng/mL in the study autism spectrum disorders population (53 patients) (P <0.01), 27 ng/mL in autism spectrum disorder subjects with periodic limb movements of sleep (25 patients) (P = 0.01), and 24 ng/mL in autism spectrum disorders subjects with sleep fragmentation (21 patients) (P = 0.02). Within the autism spectrum disorders population, median ferritin levels were significantly lower in patients with poor sleep efficiency (7 ng/mL) versus those with normal sleep efficiency (29 ng/mL) (P = 0.01). The prevalence of periodic limb movements of sleep was 47% in autism spectrum disorders compared with 8% in controls (P < 0.01). CONCLUSION: Children with autism spectrum disorders had significantly lower ferritin levels compared with controls. In addition, they experience a higher prevalence of sleep fragmentation, obstructive sleep apnea, and periodic limb movements of sleep than children with ASD and no sleep complaints. Our preliminary observations, which have not been described before, need to be validated in multicenter prospective studies.