Phase angle between dim light melatonin onset and sleep timing during residential treatment prospectively predicts obsessive-compulsive symptoms

The relation between obsessive-compulsive disorder (OCD) and circadian rhythm disturbance has been increasingly acknowledged in recent years. While prior clinical studies have utilized patients’ self-reported sleep behaviors, there is a need to also explore the measurable, biological aspects of circadian rhythms. The current study has two aims: first, to describe the biological circadian rhythms of individuals with OCD seeking intensive residential treatment, including their relationship with self-reported measures of sleep and OCD symptoms; and second, to examine longitudinal associations between biological circadian rhythms and OCD symptom severity during the course of residential treatment. Adults receiving residential treatment for OCD (n = 23) completed a procedure to measure their dim-light melatonin onset (DLMO) at admission, week two, week four, and discharge from treatment along with a battery of self-report assessments of OCD symptom severity and depression severity. Phase angle between DLMO and the midpoint of self-reported sleep was also calculated as a measure of the alignment between behavioral sleep-wake patterns and biological circadian rhythms at each time point. Cross-sectional correlations between these constructs were assessed and then cross-lagged panel models (CLPM) were fit to these data in order to examine the relation between 1) DLMO and OCD symptom severity across treatment and 2) phase angle of DLMO and midpoint of sleep and OCD symptom severity across treatment. Descriptive statistics indicate that sleep duration and timing were shifting closer toward general population averages across this period of treatment, perhaps due to newly supported bed and wake times in the treatment milieu. There were no significant cross-lagged paths between DLMO and OCD symptom severity during the first weeks of residential treatment. There was a significant cross-lagged path between DLMO phase angle from self-reported sleep midpoint and OCD symptom severity during the first weeks of residential treatment. Specifically, relatively shorter phase angle at admission was associated with less severe OCD symptoms at the second week of treatment; and relatively shorter phase angle at the second week of treatment was associated with more severe OCD symptoms at the fourth week of treatment. This study demonstrated the feasibility of measuring biological circadian rhythms in a residential treatment context and provided initial data demonstrating a longitudinal and dynamic relation between sleep, circadian rhythms, and OCD symptoms. Further study with larger samples is warranted. The non-linear pattern of relations across the course of this study also indicate that consideration of treatment processes and other factors not measured herein will strengthen future studies. Follow-up studies with residential treatment settings that continue salivary melatonin collection after treatment ends and patients return to their daily lives are also possible with this self-administered data collection procedure.