Insomnia with objective short sleep duration is associated with a reduced response to active or placebo acupuncture: a secondary analysis of three randomized controlled trials

Studies have shown that insomnia with polysomnography-derived short sleep duration is associated with increased health risks. Preliminary data suggest that this phenotype has a blunted response to cognitive-behavioral therapy and actigraphy-derived sleep variables can be used as a substitute of polysomnography to predict treatment response. This secondary analysis aimed to determine whether insomnia with polysomnography-derived short sleep duration, defined as < 6 h, had a blunted response to active or placebo acupuncture and the use of actigraphy-derived sleep variables as predictors of treatment response. Data from three randomized, placebo-controlled trials of acupuncture for insomnia were analyzed. A reduction of Insomnia Severity Index score ≥ 8 points from baseline to 1-week posttreatment was used to define treatment response. Totally 185 subjects who had laboratory-based 1-night polysomnography for at least 7 h were included; 62 (33.5%) had polysomnography-derived sleep duration < 6 h, while 50 (27.0%) were classified as responders. Non-responders were significantly more likely to have below tertiary education (p = 0.04) and polysomnography-derived short sleep duration (p = 0.02), while baseline sleep-diary and actigraphy-derived total sleep time and sleep efficiency were not significant predictors. Multiple logistic regression showed that polysomnography-derived short sleep duration was the only significant predictor of treatment response (OR 2.23, 95% CI 1.01, 4.91, p = 0.048). Despite the marginal significance level, our findings show that insomnia with polysomnography-derived short sleep duration has a lower response to acupuncture and is a biological phenotype that is worth further investigation. Clinical Trial Registration: ClinicalTrials.gov identifier: NCT00839592, NCT00838994, and NCT01707706.