0892 Dim Light Salivary Melatonin Onset in Autistic Children Using an In-Home Protocol

Abstract Introduction Sleep problems and irregular circadian rhythms are common in autistic youth, yet few studies have evaluated dim light melatonin onset (DLMO), the gold-standard measure of circadian phase, in this population. This study assessed the feasibility of an in-home DLMO protocol with autistic youth and examined associations between DLMO timing, sleep patterns, endogenous melatonin production, and child characteristics. Methods Twenty-eight autistic youth ages 8–17 years old completed an in-home DLMO protocol beginning at 4:30pm and ending at 10:00pm. Blackout curtains were installed to maintain ambient light 50 lux across the protocol. Saliva samples were collected every 30 minutes and assayed via radioimmunoassay. Participants wore a Fitbit continuously for one week, and caregivers completed validated questionnaires on sleep and behavior. Results Of the 28 participants, DLMO could be calculated for 12; 5 participants showed chronically low melatonin levels ( 4 pg/mL at every sampling time), and 8 had consistently high melatonin levels (4 pg/mL throughout). Three participants were unable to produce sufficient saliva for assay. For the 12 participants with a calculable DLMO, the mean DLMO was 6:35pm (SD=78 minutes). Later DLMO was associated with older age (r=0.83, p 0.001), lower caregiver-reported sleep anxiety (r=–0.67, p=0.018), later caregiver-reported sleep midpoint on non-school days (r=0.77, p=0.003), later Fitbit-derived sleep midpoint (r=0.66, p=0.026), and greater intraindividual variability in Fitbit-derived sleep midpoint (r=0.69, p=0.020). Differences emerged between endogenous melatonin production groups. Compared to children with normal or high melatonin levels, those with low melatonin at each sampling point demonstrated lower intellectual functioning (F=7.05, p=0.01), more caregiver-reported nighttime awakenings (F=5.68, p=0.03), and more sedentary behavior based on Fitbit data (F=4.16, p=0.05). Conclusion In-home DLMO assessment appears feasible for many autistic children, though substantial variability in endogenous melatonin production may limit the ability to derive circadian phase for a large subset of participants. Consistent with non-autistic youth, later circadian phase correlated with older age and later sleep timing. Differences observed in children with chronically low melatonin levels highlight a potentially meaningful subgroup, but replication is needed. Overall, these findings support continued evaluation of circadian phenotyping as a possible tool for tailoring sleep interventions in autism. Support (if any) R00HD102586)