Where there is darkness, let there be light: possible benefit from light exposure to the neck prior to sleep

“Where there is darkness, let there be light” —Attributed to St. Francis of Assisi

In this issue of the Journal of Clinical Sleep Medicine, Kennedy et al¹ report on an innovative device designed to improve the quality of sleep by applying light of specific wavelengths to the necks of participants with generally mild sleep complaints. According to the study design, the investigators recruited 30 adults aged 30 to 60 years with a “sleep complaint” as demonstrated by an Insomnia Severity Index (ISI) score ≥ 8, indicating a sleep disturbance of at least “subthreshold insomnia.”² The majority were women aged 30 to 60 years; individuals were excluded if they had a preexisting sleep disorder (either self-reported or as identified by the Sleep Disorders Symptom Checklist-25 [SDSCL-25]),³ were taking any medication known to affect sleep, had a medical condition that prevented participation, or carried a self-described psychiatric disorder. Participants completed a 2-week baseline period during which they wore the Ōura ring (Ōura Health Oy, Oulu, Finland) (a commercial, finger-worn apparatus designed to detect sleep, wakefulness, and sleep stage)⁴ and completed a sleep diary. Following the baseline period, participants wore the device (described as a neoprene collar outfitted with 48 light emitting diodes [LEDs]) for 25 minutes during the 1-hour period before their chosen bedtime; participants were not required to be sedentary but rather were allowed to pursue “typical bedtime rituals” that included watching television or reading. The 48 light sources emitted a combination of wavelengths spanning the spectrum from red to near infra-red (IR) encompassing 660 nm (red), 740 nm (red to near IR), 810 nm (near IR), and 870 nm (near IR), and were chosen based on previous literature that found maximal biological effects from light in red, near IR, and IR wavelengths; note that this commentary also is limited to data related to these wavelengths. Participants wore a collar every night for 3 weeks, alternating between the active device and a sham device. Sleep was evaluated using the Ōura ring; multiple survey instruments (too numerous to list here) aimed at measuring a wide variety of sleep and mood variables were administered on the first night of device usage and every week thereafter, and outcomes were derived from the data collected during the first night prior to device use compared with those from the final weekly survey at the end of the intervention period. There were 3 a priori–defined primary outcomes: self-reported impression of daytime performance, ISI score, and Ōura ring–derived sleep efficiency; they applied the Holm-Bonferroni correction to assess statistical significance, a more generous methodology compared with the Bonferroni correction.⁵ Nevertheless, the authors reported no statistical difference for any of the primary outcomes attributable to the application of their novel device. With respect to secondary outcomes, the most significant finding was improvement during active treatment compared with sham with respect to the Systematic Assessment for Treatment of Emergent Effects (SAFTEE) survey, which measures self-reported problems falling asleep, trouble thinking or concentrating, and anxiety.⁶ Based primarily on the SAFTEE survey and some within-group analyses, the authors concluded that the use of their low-level light therapy (LLLT) device resulted in a self-reported perception of relaxation and improved sleep that might translate into benefits in sleep and daytime function.

I must admit that, after more than 40 years of practicing academic sleep medicine, my only exposure (pun intended) to light as a therapeutic modality consisted of manipulating circadian rhythm, mood, and cognition by ocular light exposure. As is probably the case with most of us involved in sleep medicine, ocular light exposure is assumed to be the only mediator of these effects and it follows that circadian processes would be involved in some manner or form. Much of the evidence for this assumption derives from light exposure as a therapy for seasonal affective disorder, although proof of this hypothesis has been surprisingly elusive.⁷ In point of fact, other mechanisms for the diverse effects of ocular light exposure may involve serotoninergic or monoaminergic pathways⁸ or even sites other than the suprachiasmatic nucleus of the hypothalamus; for instance, a perihabenular nucleus of the thalamus has been identified as one such site.⁹ The unifying factor, however, is the primacy of ocular light exposure rather than the illumination of other external body sites. One need only recall the widely publicized report in the journal Science that light exposure to the back of the knee could shift circadian rhythm,¹⁰ later debunked by Wright and Czeisler,¹¹ as well as the substantial body of literature comparing light exposure to the eyes vs exposure at other sites to be skeptical of reports countering that principle, with the randomized controlled trial by Koorengevel et al¹² serving as one example.

And yet, there exists an intriguing body of evidence that external light exposure, transduced by other opsins and not involving the eyes, plays a functional role in organisms not much earlier on the evolutionary tree than the genus Homo¹³ and that similar opsins are present in modern humans.¹⁴ The current article brought to my attention the use of light therapy of a fairly low power applied to a variety of external locations with apparent therapeutic effect. Such exposure does not disrupt tissue but rather activates an array of molecular mechanisms beneath the skin that may have salutary effects. A review of the putative actions of red and near-IR light characterized as LLLT, or more generally, photobiomodulation not applied to the eyes may be found in a technical review by de Freitas and Hamblin.¹⁵ These authors posit several possible mechanisms. Cytochrome c oxidase absorbs light in the near-IR region due to its heme and copper centers, leading to photon-induced disassociation of inhibitory nitric oxide (NO) from the enzyme and an increase in electron transport, mitochondrial membrane potential, and ATP production. An alternative hypothesis concerns the activation of light-sensitive ion channels, allowing calcium to enter cells. Whether one or the other or perhaps both explain the phenomenon, photon absorption activates numerous signaling pathways involving reactive oxygen species, cyclic AMP, NO and Ca²⁺, followed by activation of specific transcription factors. Once activated, these transcription factors can lead to increased expression of genes responsible for a broad range of salutary effects. Stem cells and progenitor cells seem to respond robustly to this form of LLLT, while the transcription of specific genes stimulates protein synthesis, cell migration and proliferation, anti-inflammatory signaling, and production of antiapoptotic proteins and antioxidant enzymes. Treatment similar to that described in the article by Kennedy et al has been the subject of meta-analyses and systematic reviews in diverse conditions. The literature is inconsistent with respect to whether lasers or LEDs are used as light sources, but based on the meta-analyses, it is unlikely that the effects of these different light sources differ. For instance, Clijsen et al¹⁶ performed an elegant meta-analysis that revealed an overall salutary response to the use of LLLT in the treatment of pain in adult patients with musculoskeletal disorders. A total of 21 head-to-head comparisons identified 17 that favored LLLT while four comparisons failed to demonstrate beneficial effects. Another well-performed meta-analysis using GRADE (Grading of Recommendations, Assessment, Development, and Evaluations) methodology suggested that LLLT can benefit patients with fibromyalgia, and incorporated studies using both lasers and LEDs.¹⁷ Finally, in a study perhaps more relevant to the practice of sleep medicine, Giménez et al¹⁸ recently reported on a randomized, double-blind, controlled trial that demonstrated a positive effect on well-being and health, although only when administering the highest dose of photobiomodulation (6.5 J·cm⁻²) and only during winter; there was no improvement in estimates of sleep quality. Importantly, circadian-related outputs were not affected by photobiomodulation despite the fact that the light exposure tested necessarily included some ocular illumination. Interestingly, the participants were described as expressing mild sleep-related complaints, presumably a group of individuals with similarities to those reported herein by Kennedy et al.

Ordinarily, a commentary tends to focus on those aspects of a research report incorporating “irrational exuberance” (in the memorable words of Alan Greenspan), meaning overly optimistic interpretations of positive results. Clearly, such an approach is not appropriate in the case of an essentially negative study such as that reported by Kennedy et al. Instead, it is more appropriate to point out whether the hypothesis or methodology might have been improved in some way, leading to positive outcomes for some of the metrics. To their credit, Kennedy et al surely deserve praise, not only for this ambitious study but for their recognition of its weaknesses. Already included in their article is a well-thought-out paragraph on limitations, leaving preciously little for a commentator to add. They admit to the enrollment of a limited number of participants, which almost certainly exposes the study to type II error. They acknowledge the possibility that heat from the LED array may have had a biological effect or been sensed by the participants and thus unblinding them to the active vs sham intervention. I would posit that measuring the temperature under the collar might have eliminated that concern. Since the sham collar was not the active collar with the LEDs turned off, those participants who found a collar to be uncomfortable may have confounded some of the results if there were differences in fit between the two appliances. To those limitations I am tempted to add their use of the Ōura ring for sleep/wake measurements and sleep staging; quantification of sleep time using this device has generally been comparable to that of polysomnography, while detection of wakefulness and sleep stage has been significantly less impressive.⁴ Absent a compelling reason, a medical grade actigraph might have been preferable.

Finally, the sheer number of secondary outcome measurements (I count at least 8, not including the primary outcomes) inevitably raises the accusation that the authors engaged in, at least to some extent, a “fishing expedition.” The more generous term for incorporating a large array of secondary outcomes in a research trial is that such a design is “hypothesis-generating.” However, the investigators already had a hypothesis as represented by the primary outcome measures, and surely concentrating on fewer metrics would have made the study less confusing and time intensive for the participants. Consequently, one might expect that, in order to complete the large number of survey instruments and questionnaires required by the study design, some participants might be tempted to sacrifice accuracy and thoughtfulness in favor of limiting the amount of time devoted to their participation. This trade-off might have been difficult for participants to resist, and one can only speculate what effect this might have had on data quality, and whether fewer secondary outcomes might have led to more positive results.

DISCLOSURE STATEMENT

Dr. Brown co-edits the Sleep and Respiratory Neurobiology section of Current Opinion in Pulmonary Medicine and is co-author of a chapter on positive airway pressure treatment for obstructive sleep apnea in UpToDate. He chairs the Polysomnography Practice Advisory Committee of the New Mexico Medical Board and chairs the New Mexico Advisory Board for Respiratory Care. He chairs the Board of Directors of GAMA-PAC, the political action committee of the Greater Albuquerque Medical Association (GAMA) and serves on the GAMA Board of Trustees and on the Council of the New Mexico Medical Society.