The “Anti-Inflammatory” Properties of CBT-I

While the evidence-base for the effectiveness of cognitive behavioral therapy for insomnia (CBT-I) has accumulated in the past 30 years, some sage observations were present in the medical literature more than 130 years ago:

“Many minor points are worthy of attention in the cure of chronic psychic insomnia. In most cases, whether he sleep badly or well, the patient ought, from day to day, to go to bed at some fixed hour. Healthy sleep tends to occur periodically. Daily bodily exercise, short of great fatigue, must be enjoined. Riding in a carriage is good, walking better, riding on horseback the best of all. A worn and self-worrying man, wrapt up in the absorbing current of self-consciousness, may take exercise in a carriage or upon his legs and still keep up his fretting, but he must come out of himself when he gets into the saddle. […] Very often the surest way of keeping awake is to try hard to get to sleep. We do most things best when we forget ourselves: going to sleep is no exception to the rule.”

—James Sawyer, MD, The Lancet, June 22^nd 1878¹

Although we are not aware of any controlled data showing that horse-riding, specifically, confers benefits to sleep, there is mounting evidence that physical activity levels are associated with restorative sleep and that prescribed exercise may even be an effective single-component therapy for chronic insomnia.² On more familiar territory, Dr. Sawyer intimates that stabilizing bed and rise-times, aligning circadian and homeostatic drives for sleep and wake, tuning down cognitive arousal and obviating sleep effort, may be helpful in the management of insomnia; and indeed these are clear targets of contemporary CBT-I.^3–5

CBT-I is an effective multicomponent therapy, and while the field should not “rest yet,”⁶ we can be content that CBT-I is the treatment of first choice for the improvement of chronic poor sleep.^7,8 One major ongoing challenge for sleep science, however, relates to whether ameliorating insomnia, and optimizing sleep more generally, can improve health outcomes and/ or offset the development of disease.^9,10 Cross-sectional, longitudinal, and epidemiological studies are converging on the view that insomnia is not only associated with ill health but that chronic poor sleep may be involved in the development and maintenance of ill health. Some of the most robust evidence can be found in the mental health domain, particularly for affective disorders,¹¹ but a number of studies in the past decade strongly link insomnia to hypertension, stroke, diabetes, and cardiovascular disease (CVD).^12–14 Whereas it was noted nearly a century ago that insomnia “…when thoroughly engrafted on an individual often lasts months or years, and interferes with its possessor's health and happiness,”¹⁵ surprisingly, we possess little robust evidence that treating insomnia delivers long-term improvements in health¹⁰ or happiness.¹⁶

It is in this context that a recent trial by Irwin and colleagues, appearing in this issue of SLEEP,¹⁷ is particularly noteworthy. They report on a randomized, controlled, three-arm trial of multicomponent CBT-I, Tai Chi Chih (TCC; structured, movement-based meditation, previously shown to impact positively on sleep quality), and sleep seminar (SS) education. Elderly patients aged > 55 years with a diagnosis of primary insomnia, who had been thoroughly screened for sleep disruptive comorbidity, were randomized to one of the three treatment arms. They received weekly, 2-hour therapy sessions (in groups of 7-10), over a 16-week period, led by two psychologists trained in behavioral medicine. Major assessments were conducted at baseline, post-treatment, and 7 months and 16 months follow-up. The primary outcome of interest was insomnia remission, determined by independent and blind clinical assessment. Secondary outcomes included classic measures of self-reported sleep quality (PSQI), sleep diary parameters, polysomnographic sleep continuity variables, and daytime functioning (clinician-rated depressive symptoms, fatigue, and sleepiness). In what may be a “first” for CBT-I investigations, the study team, noting the relationship between insomnia and inflammatory-related conditions (e.g., CVD, hypertension, diabetes), also sampled high-sensitivity C-reactive protein (CRP) at baseline, post-treatment, and 16-month follow-up, to index inflammatory risk. CRP is an annular pentameric hepatocyte protein found in blood plasma, the levels of which rise in response to an acute inflammatory stimulus. The production of CRP is thought to reflect the activity of proinflammatory cytokines, particularly interleukin (IL)-6. While a nonspecific marker for disease, high levels of baseline CRP (> 3.0 mg/L) have been shown to predict future onset of diabetes, hypertension, and CVD. Importantly, elevated CRP levels have been observed after experimental sleep curtailment,¹⁸ and, while still tentative, have been associated with insomnia symptoms.^19–21

The main trial results were that CBT-I outperformed both TCC and SS education with respect to proportion of insomnia remitters, assessed post-intervention (54.2%, CBT-I; 29.7%, TCC; 21.8%, SS). Secondary outcome measures also evidenced treatment effects in favor of CBT-I; for global sleep quality, sleep diary variables (SOL, WASO, SE), fatigue, depressive symptoms, and CRP. The proportion of patients in the CBT-I arm exhibiting CRP > 3.0 mg/L decreased from 20.6% at baseline to 11.6% at 16 months, while the proportion in the SS group increased from 21.7% to 33.3%. This latter finding, with respect to the control arm, may hint at the longitudinal burden of untreated insomnia in the elderly. Categorizing the entire sample into remitters and non-remitters at 16 months revealed CRP levels that were approximately 50% lower in remitters versus non-remitters. Long-term follow-up indicated that CBT-I effects on sleep and functioning were durable, whereas initial TCC improvements (on secondary measures of sleep quality, depression, and fatigue) were not maintained at follow-up. Of note, no treatment effects were observed for PSG variables.

The trial design used by Irwin and colleagues¹⁷ included a number of important methodological strengths, including matched therapist contact time across treatments arms; rigorous participant screening; assessment of subjective and objective sleep; focus on daytime functioning variables; clinically meaningful end-points; long-term follow-up; parity of treatment arm acceptability; and the novel collection of CRP to index inflammatory risk. In contrast to most CBT-I studies—although not atypical of CBT for other psychological disorders—therapist contact time was extensive, involving 32 hours over a 4-month period. This was required, the authors argue, in order to accommodate the time needed to learn TCC skills.

The content of CBT-I evaluated by Irwin is also worth remarking on. Additional modules focused on mood enhancement/emotion regulation and management of daytime activity levels, while sleep restriction therapy (SRT) was omitted (owing to the link between short sleep and inflammation). Some have called for greater attention to emotion regulation in insomnia etiology and management,^22,23 and indeed Dr. Sawyer recognized its importance in 1878.¹ Building on recent trends to dismantle CBT-I and assess the relative efficacy of cognitive and behavioral elements,^24,25 the field should now formally assess whether the addition of emotion regulation strategies and a focus on physical activity (perhaps, also, as a countermeasure to SRT-related sleepiness²⁶) can enhance CBT-I response.

While trials cannot comprehensively inform on mechanisms, it is worth speculating on what may have accounted for the reduction in CRP in the report by Irwin and colleagues.¹⁷ Consistent with several CBT-I investigations, subjective and objective total sleep time did not improve post-treatment or follow-up; ruling out a parsimonious sleep-time explanation. Similarly, self-reported physical activity and BMI—variables known to be associated with CRP levels—remained unchanged across the groups. As the authors note, insomnia has been linked to sympathetic hyperactivation, the consequence of which may be a cascade of processes leading to elevated production of IL-6, and concomitant rise in CRP. However, there is no robust evidence that CBT-I in this trial (or any other) alters sympathetic activity. Future replications of the findings of Irwin et al. and secondary analyses of the current trial dataset, may provide insights into mediating and moderating variables in the CBT-I/CRP relationship. It would be prudent, also, in future studies to sample additional markers of inflammation, cognitive and somatic arousal, and sleep micro-architecture, and, perhaps in the context of larger, more long-term trials, assess incidence of disease.²⁷

The results of Irwin et al.¹⁷ are likely to have an important impact on the insomnia and sleep field. The study provides further robust evidence that CBT-I effectively improves both night and daytime symptoms of insomnia, and, conversely, contributes the first data to suggest that TCC is unlikely to be an effective treatment for insomnia disorder. While the CRP findings require replication and extension, it is encouraging that treatment effects observed here are comparable to previous trials of exercise and weight loss interventions. This may be an important step forward in the public health goal of emphasizing three key pillars of health—diet, exercise, and sleep.²⁸

CITATION

Kyle SD, Spiegelhalder K. The “anti-inflammatory” properties of CBT-I. SLEEP 2014;37(9):1407-1409.

DISCLOSURE STATEMENT

Dr. Kyle has consulted for Sleepio, LTD. Dr. Spiegelhalder has indicated no financial conflicts of interest.