Yoga has been shown to reduce pain and improve function in populations with chronic low back pain (CLBP) in multiple randomized controlled trials (RCTs) [1,2]. However, results vary, and effects are generally modest [3]. This variation is perhaps due to differences in the specific neurobiological mechanisms through which yoga exerts clinical improvements on pain that are affected by different yoga interventions. These mechanisms, which have been only minimally identified, could potentially be harnessed to achieve more consistent and optimal outcomes from yoga interventions. A study that examined mediators of yoga’s effects on pain lacked strong theoretical bases for assessment of the mechanisms involved [4]. Using a comprehensive theoretical emotion regulation (ER) model of yoga [5], we will conduct the first test of ER as a primary mechanism of yoga’s effects on CLBP. We will also test the role of pain sensitization using somatosensory and gene expression profiles in this model [6] and examine the specific aspects of ER.

ER refers to “extrinsic and intrinsic processes responsible for monitoring, evaluating, and modifying emotional reactions, especially their intensive and temporal features, to accomplish one’s goals” [7]. Emotions strongly influence perceptions of pain intensity and predict disability [8], particularly among individuals with CLBP [9,10]. Interventions that promote adaptive ER skills reduce pain in patients with chronic pain [11]. Increasing evidence demonstrates that consistent yoga practice can promote improved ER [12,13], but research has not yet tested whether the effects of yoga practice on CLBP are due to improvements in ER.

Given the importance of ER in managing chronic pain and yoga’s demonstrated promotion of improved ER, we developed an integrative model of yoga’s effects on CLBP that highlights the key role of ER. Previous work by our team [14,15] and others [16] demonstrated that yoga simultaneously comprises multiple components and that physical activity (poses and transitions between poses), breathwork, and meditation are all important therapeutic elements. Thus, although physical activity interventions alone have demonstrated salutary effects on negative emotion [17] and ER [18,19], yoga’s inclusion of breathwork and meditation along with physical activity may more potently improve ER [5].

Yoga provides myriad opportunities to develop ER skills helpful to people managing chronic pain. Practitioners can learn how to experience sensation without negative evaluative judgments, cultivating nonreactivity. Due to cognitive priming and reinforcement during movement, yoga provides simultaneous input to cognitive, motor, and sensory pathways, which may allow practitioners to consciously influence their interpretation of sensation or pain and manage its affective dimensions. Given the myriad components of yoga that can facilitate developing ER skills, we posit that ER is a key mechanism by which yoga exerts effects on individuals with CLBP. The few studies examining links among yoga, ER, and pain support our model (Figure 1). For example, yoga practitioners were better able to tolerate experimentally induced pain compared to nonpractitioners, with corresponding greater left intrainsular white matter connectivity [20]. Further, the yoga practitioners reported coping with the pain stimulus using (adaptive) ER strategies such as relaxation, nonjudgmental focusing, and reappraisal. Thus, the ability to minimize the affective dimensions of pain may dampen physiological responses to negative emotions such as anxiety, which have been associated with the perception of pain [21,22].

Recent studies document that patients with CLBP demonstrate sensory and functional alterations reflecting peripheral and central nervous system sensitization associated with enhanced pain sensitivity [23-25]. Enhanced pain sensitivity as a mechanism of CLBP includes sensitization of nociceptors and neuronal circuits [26] and increased pain signaling through membrane excitability and synaptic efficacy [27], which are conferred through altered expression of pain sensitivity (and immune-related) genes [6,28]. The somatosensory changes associated with enhanced pain sensitivity in CLBP involve decreased thermal and mechanical pain thresholds at the site of pain and remote areas [25,26].

Preliminary research suggests that negative emotion is an important determinant of pain sensitivity [29,30], predicting up to 23% of the variance in CLBP-related disability [31,32]. For instance, greater pain catastrophizing and anxiety are associated with lower standardized values of heat pain threshold and tolerance in patients with chronic pain [33]. In turn, negative emotions and inability to tolerate stress have been shown to activate inflammatory pathways through the release of proinflammatory cytokines and substance P [34,35]. To date, only 2 small studies have examined whether yoga specifically influences pain sensitivity. However, in a systematic review of 15 RCTs, yoga was associated with decreases in interleukin-1β, interleukin-6, and tumor necrosis factor-α [36]. In addition, specific components of yoga (meditation, relaxation, and breathing) have been shown to alter gene expression profiles and levels of inflammatory mediators [37-39].

As shown in Figure 1, we propose that pain sensitization moderates yoga’s effects on pain, where patients with CLBP with higher sensitization may benefit less [40,41], given that their pain is more centralized [42,43]. This is a fairly novel hypothesis with some supportive evidence [44,45]. We will also explore the possibility that yoga and improved ER abilities might actually improve sensitization and thus decrease pain (Figure 2).

Little research has demonstrated that treatments for pain, including behavioral interventions, modify sensitization once it has occurred [46,47], although a reversal of central sensitization was observed in patients with structural pain–related pathology who underwent corrective surgery [48] as well as patients with migraine and chronic pancreatitis [49,50]. We considered alternative strategies to test these linkages. For example, a systematic review of 1037 individuals with CLBP provided moderate evidence that structural brain differences in specific cortical and subcortical areas of the brain and altered functional connectivity in pain-related areas following painful stimulation differentiate chronic pain from pain-free individuals [51]. While brain imaging provides 1 way to examine changes in pain sensitivity, we assert that identifying specific cellular mechanisms through which ER affects pain sensitivity will generate greater translational capacity for the clinical care of individuals with CLBP.

To be comprehensive, our model includes improved strength and flexibility, given that improvements in these domains are common in yoga intervention trials. In fact, yoga typically shows the effects on strength, flexibility, and other physical outcomes that are as substantial as those of physical therapy or stretching-strengthening exercise groups [52,53]. Sherman et al [4] demonstrated that strength and flexibility may mediate some of yoga’s effects on pain. Overall, accumulating evidence demonstrates moderate but varied effects of yoga for reducing CLBP pain and disability; much remains to be learned about how yoga exerts these effects and why its effects vary. Understanding the proposed mechanism of actions (MOAs) of yoga will provide not only a more solid scientific basis to assure clinicians of the value of yoga to patients with CLBP and other conditions but also important targets for intervention modification and optimization.

To address this gap, we will test the impact of a yoga intervention on ER and pain severity and functioning compared to a general stretching or strengthening intervention. This design allows us to isolate the additional effects of yoga above and beyond more general movement practices, thereby testing ER as a key MOA underlying the clinical effects of yoga on CLBP. Specifically, we will test whether yoga’s effects on CLBP (improved pain severity and interference) are mediated by improved ER, the extent to which effects are related to specific aspects of ER, and the role of pain sensitization as a mediator or moderator of effects.

For the active comparison intervention, we chose to use stretching and strengthening exercises, given that we are interested in determining not only whether yoga influences ER but also the extent to which these effects are due to aspects of yoga above and beyond nonspecific factors (eg, expectancies, time, attention, and effort) [13] as well as improvements in strength and flexibility [4]. The 2 intervention groups will be similar in many aspects, but only yoga will incorporate aspects that we hypothesize will facilitate ER (eg, breathwork, mindfulness, and relaxation). We selected the hypothesized MOA based on our extensive work regarding the therapeutic characteristics of yoga and mechanisms of CLBP [5,14]. Our 3-month and 6-month follow-ups will allow us to examine whether the effects of yoga emerge later or persist. The study is highly innovative in its approach to testing the MOA of yoga for alleviating CLBP. By identifying the specific mechanisms through which yoga may influence pain, these results will facilitate optimizing interventions to improve therapeutic accuracy for prescribing and using yoga for CLBP, a highly prevalent and costly health condition in the United States and worldwide.

This study is a 2-arm parallel group blinded RCT to examine ER as a MOA of yoga in individuals with CLBP (Figure 3). In total, 204 people with CLBP will be randomized to receive yoga or the stretching and strengthening intervention with assessments of pain severity and interference; ER; pain sensitivity; physical strength and flexibility; and quality of life at baseline, 6 weeks (midway through the intervention), 12 weeks (after completion of the intervention), and 3 and 6 months after the completion of the intervention. Both groups will be instructed to continue the use of any other strategies or treatments that they have been using to manage their CLBP throughout the study.

We will enroll 204 men and women aged 18 years and older with nonspecific CLBP who meet eligibility (Textbox 1). Nonspecific CLBP is defined as pain without a specific cause or need for surgical intervention that is anywhere in the region of the low back bound superiorly by the thoracolumbar junction and inferiorly by the lumbosacral junction, which has been present for more than 3 months out of the prior 6 months and is currently rated at a level of ≥2 on the numeric rating scale. Eligibility will be determined by asking the following screening questions [54]: (1) How long has back pain been an ongoing problem for you? (2) How often has low-back pain been an ongoing problem for you over the past 6 months? Responses of greater than 3 months to question 1 and “at least half the days in the past 6 months” to question 2 will define CLBP.

Active and passive recruitment strategies will be used to recruit individuals with CLBP include (1) contacting CLBP participants enrolled in an institutional review board (IRB)–approved pain registry, (2) advertisements on social media (Instagram, Facebook, Twitter, Reddit, and Craigslist), (3) placing printed advertisements in public transportation vehicles and local newspapers, and (4) distributing flyers at local outpatient clinical sites and local universities. We will recruit continuously and start a new cohort every 6 months, so that there will be cohorts in both interventions and follow-up phases throughout the study. This approach will help control for seasonal effects and other potential confounds. All participants will receive a yoga or exercise mat, block, and strap prior to their first session, and US $70 following the baseline, 6- and 12-week postintervention, and 3- and 6-month follow-up assessments with an additional US $50 for those who complete all data collection sessions.

Interested volunteers are asked to complete a web-based screening questionnaire via REDCap (Research Electronic Data Capture; Vanderbilt University), which includes questions to ensure they meet inclusion and exclusion criteria and can participate safely. Volunteers who are deemed eligible to participate are contacted by study staff to schedule their baseline appointment.

The study was approved by the University of Connecticut IRB (#H19-191). Prior to completing any study procedures, all participants provide written informed consent on a form that was approved by the IRB at the University of Connecticut (IRB#H19-191). All study procedures are in accordance with the ethical standards described by the Helsinki Declaration. Study participants are assigned an identification number, allowing all stored records to be deidentified. Deidentified data are kept in locked file room or stored in a password-protected database on a password-protected computer. All participants are monetarily compensated for their participation up to US $400 for completing all study visits. Participants can withdraw from the study at any time.

The assessments of pain intensity, pain interference, ER, and other parameters will be taken at baseline, 6 and 12 weeks, and 3 and 6 months post intervention.

Demographics, class attendance or home practice, expectancies, and substance use will be measured to examine the impact on the primary and secondary outcomes. A brief questionnaire (minimum data set as recommended by the National Institutes of Health Taskforce on Research Standards for CLBP) [54] will assess demographics (age, gender, race or ethnicity, education, and employment status) and medical or psychiatric history (comorbidities, medication usage, and ongoing treatments). Attendance at yoga or stretching sessions will be assessed using attendance data from the session recordings. The Essential Properties of Yoga Questionnaire [72] will be used to rate both interventions (yoga and stretching or strengthening) and measures the extent to which 14 different dimensions are emphasized in a mind-body intervention (eg, physical challenge, breathwork, restorative postures, and mindfulness). The Credibility Expectancy Questionnaire [73], a 6-item self-report instrument, will be used to assess treatment credibility and patient expectancy for improvement [74]. In order to assess participants’ use of tobacco, alcohol, prescription medicines, cannabis, and other nonprescribed substances, the Alcohol, Smoking, and Substance Involvement Screening Test [75] will be administered by study staff at each time point.

The primary aim of this study will test the null hypothesis that the 2 population means are equal between the yoga and stretching and strengthening groups on the primary outcome, pain severity as measured by the BPI-SF, as well as the primary MOA, ER as assessed by the DERS. Power calculations incorporated prior effect sizes from works by others. A proposed sample size of 102 per group, which adjusted for attrition, was calculated to detect statistically significant results with 80% power (estimated d=0.50). In consideration of our secondary aims, we also powered the detection of paths a and b within pain interference mediation models. This was done using simulation in R (R Foundation for Statistical Computing; N=5000). Following simulation methods (N=5000) presented in Fritz and MacKinnon [76], sample size was determined by increasing simulation sample size until 90% of simulations significantly detected both paths a and b with estimated effect size magnitudes of d=0.39 (path a) and d=0.26 (path b) at α=.05. A proposed sample size of 116 was calculated, which was already reached when powering our primary aim.

All analyses will be performed in SPSS (version 29; IBM Corp). A P value of .05 or less will be considered statistically significant. The primary outcome is pain severity (BPI-SF) and the primary MOA is ER (DERS). We will examine baseline demographics and study variables descriptively and test for differences between yoga and stretching and strengthening groups. If assumptions are met, a series of 2-sided t tests, mixed effects models, and path analyses will test our hypotheses. A mediation path analysis will be conducted examining the indirect path of group and change in ER from baseline to 6 weeks (path a), and change in ER from baseline to 6 weeks and pain severity (BPI-SF) at 12 weeks (path b) controlling for baseline pain severity. The magnitude of indirect path b will be assessed controlling for path a and the direct path between group and change in pain severity at 12 weeks (path c) to test if changes in ER predict subsequent changes in pain severity. Identical models will be created with change in ER from baseline to 12 weeks and pain severity at 3- and 6-month follow-ups. The magnitude of indirect path b will be compared between groups to test if the path between ER and pain severity is greater in the yoga group than in the stretching and strengthening group. If 95% CIs of path estimates do not contain 0, they will be considered meaningful. Finally, a series of moderated mediation path analyses will be conducted, examining the measures of baseline pain sensitivity (QST scores and normalized gene expression values), strength and flexibility, pain interference, and quality of life as moderators of the identical mediation model. All analyses will also be adjusted for important covariates including, but not limited to demographics (eg, biological sex, race, and sex), cohort, instructor, and intervention components (Essential Properties of Yoga Questionnaire) and credibility or perception of success (Credibility Expectancy Questionnaire). Sensitivity analyses will examine whether treatment responders and nonresponders differ significantly in baseline ER abilities. Baseline DERS scores will also be examined as a continuous moderator in main study analyses to establish whether some individuals may enjoy greater benefit from this intervention more than others.