Sleep and Behavior in Children with Autism Spectrum Disorders-Effects of Supplemental Melatonin

Background:

Children with autism spectrum disorders (ASD) have lower levels of melatonin and melatonin metabolites than typically developing children. Retrospective and small open label studies have shown that melatonin promotes sleep in children with ASD, with excellent tolerability and minimal adverse effects.

Objectives:

We are performing a prospective dose-response trial of supplemental melatonin in children with ASD. Our objectives are: (1) To identify optimal dose, tolerability, and adverse effects and (2) To study the impact of supplemental melatonin on sleep and daytime behavior.

Methods:

We included children ages 4-10 years with a clinical diagnosis of ASD, confirmed by the Autism Diagnostic Observation Schedule and Autism Diagnostic Interview-Revised, who took 30 minutes or longer to fall asleep on 3 out of 7 nights per week. Parents completed sleep and behavioral survey forms at the beginning and again at the conclusion of all study procedures. Behavioral measures included the Children’s Sleep Habits Questionnaire (CSHQ), the Repetitive Behavior Scale (RBS), and the Parent Interview for Autism-Clinical Version (PIA-CV). Children wore actigraphy watches (Mini Mitter, Respironics) for 17 weeks to provide an objective measure of sleep latency (time to fall asleep).

After one week of baseline actigraphy was recorded, parents were instructed to give their child inert flavored liquid for two weeks, 30 minutes before bedtime, to acclimate the child to taking a medication at bedtime. Once the acclimation period was completed, melatonin (Natrol ®) dosing was begun at 1 mg. The dose of melatonin was escalated every three weeks, to 3mg, 6mg, and 9 mg, until the child achieved a “satisfactory response,” documented by actigraphy, defined as falling asleep within 30 minutes of bedtime on 5 out of 7 nights per week. The first three weeks and last three weeks of actigraphy were combined in each child to produce a baseline and treatment sleep latency. Baseline and treatment measures were compared using a Wilcoxon signed-ranks test.

Results: To date, 10 children have completed the study. All children tolerated melatonin with no adverse effects. Nine of 10 children tolerated actigraphy and achieved a satisfactory response at relatively low doses—3 children at 1 mg and 5 children at 3 mg with only 2 children requiring 6 mg and none requiring 9 mg.  Sleep latency decreased from 38.7 ± 22.5 minutes (mean ± standard deviation) to 21.8 ± 7.9 minutes (p = 0.039) with treatment. Improvements with treatment were also noted in CSHQ domains of sleep onset delay (p = 0.008) and sleep duration (p = 0.004), RBS domains of compulsive (p = 0.002) and ritualistic (p = 0.004) behavior, and the PIA-CV domain of affective responses (p = 0.02). One child did not respond to melatonin by parent reports and surveys, and did not tolerate actigraphy. She was later diagnosed with bipolar disorder and treated with risperidone, with improved sleep. 

Conclusions:

Low-dose supplemental melatonin is a well-tolerated treatment for insomnia in children with ASD, and appears to impact favorably on both sleep and daytime behavior. Randomized clinical trials appear warranted.