Impact of Yoga on Functional Outcomes in Breast Cancer Survivors With Aromatase Inhibitor–Associated Arthralgias

Abstract

Arthralgia affects postmenopausal breast cancer survivors (BCSs) receiving aromatase inhibitors (AIs). This study aims to establish the feasibility of studying the impact of yoga on objective functional outcomes, pain, and health-related quality of life (HR-QOL) for AI-associated arthralgia (AIAA). Postmenopausal women with stage I to III breast cancer who reported AIAA were enrolled in a single-arm pilot trial. A yoga program was provided twice a week for 8 weeks. The Functional Reach (FR) and Sit and Reach (SR) were evaluated as primary outcomes. Pain, as measured by the Brief Pain Inventory (BPI), self-reported Patient Specific Functional Scale (PSFS), and Functional Assessment of Cancer Therapy–Breast (FACT-B) were secondary outcomes. Paired t tests were used for analysis, and 90% provided data for assessment at the end of the intervention. Participants experienced significant improvement in balance, as measured by FR, and flexibility, as measured by SR. The PSFS improved from 4.55 to 7.21, and HR-QOL measured by FACT-B also improved; both P < .05. The score for the Pain Severity subscale of the BPI reduced. No adverse events nor development or worsening of lymphedema was observed. In all, 80% of participants adhered to the home program. Preliminary data suggest that yoga may reduce pain and improve balance and flexibility in BCSs with AIAA. A randomized controlled trial is needed to establish the definitive efficacy of yoga for objective functional improvement in BCSs related to AIAA.

Introduction

Arthralgia, or joint pain, affects postmenopausal breast cancer survivors (BCSs) receiving aromatase inhibitors (AIs) and may result in reduced function and long-term well-being. ¹ Joint pain is a major side effect of BCSs receiving AIs ¹ with almost 50% of users reporting AI-associated arthralgia (AIAA). ² AIAA is a prevalent condition given the routine use of AI in postmenopausal BCS. It is likely to increase because AIs are also being investigated in combination with ovarian suppression in premenopausal women with breast cancer or as a preventative agent among healthy postmenopausal women. Although the exact mechanism by which AIs cause arthralgia is unknown, a likely explanation is that depleted estrogen levels decrease the generation of endogenous opioids in the brain, leading to lowered pain threshold. ³ Arthralgia can not only result in impaired function ⁴ but can also lead to premature treatment discontinuation.^5,6 Whereas medications such as NSAIDS may offer some pain relief, extensive use has been associated with various side effects. ⁷ Additionally, many BCSs prefer not to take medications continuously for managing daily symptoms.^8,9

Complementary therapy use among BCSs is quite common.^10,11 Yoga is widely available in the community and has improved patient-reported quality of life (QOL) in diverse groups of BCSs.^12,13 Recent studies also showed the promising effect of yoga in improving musculoskeletal conditions such as lower-back pain, ¹⁴ knee osteoarthritis, ¹⁵ and carpal tunnel syndrome ¹⁶ in noncancer populations. However, the existing literature is limited in understanding the specific effect of yoga on objective functional outcomes in BCSs.

Although research in this area is preliminary, a few human studies have demonstrated that yoga improves QOL in patients with musculoskeletal pain and therefore provides biological plausibility for addressing AI-related arthralgia.^15,17 Additionally, the benefit of yoga for osteoarthritis of the knee suggests that yoga not only decreases pain but also improves function and QOL. ¹⁷ Previous trials of yoga for BCSs, although they showed improved QOL, lacked sufficient objective measures of function.^18-24 To determine the effects of yoga on AI-related arthralgia, we proposed a pilot study to test the impact of yoga for treatment of AIAA among BCSs. The specific aims of this study were to (1) demonstrate the feasibility of recruitment and retention to a yoga trial, (2) determine the safety of the intervention, and (3) explore the effects of yoga on function, pain, and QOL.

Methods

Study Patient Population

We recruited eligible patients from southern New Jersey breast cancer clinics. Several institutional review boards approved the study protocol. Potential participants included English-speaking women aged 18 years or older who were postmenopausal, with postmenopause defined as cessation of menstrual period for more than 1 year; had a history of stage I, II, or III breast cancer; and were currently receiving AIs (anastrozole, letrozole, or exemestane). Participants had to have had joint pain attributable to AI medication and a worst joint pain score of at least three or greater on an 11-point (0-10) numerical rating scale in the preceding week. Exclusion criteria included metastatic breast cancer (stage IV), having completed chemotherapy or radiation therapy less than 4 weeks prior to enrollment, joint pain attributed to inflammatory arthritis (such as rheumatoid arthritis, gout, pseudo gout), having severe pain or noninflammatory arthralgia prior to AI initiation per patient’s history, and surgery or joint injection involving the treatment joint within the last 3 months.

Study Intervention

Iyengar yoga is widely recognized as a form of yoga that has a reproducible format; combines precise postures (asanas), breathing exercises (pranayama), and meditation; and can be used therapeutically to treat a wide variety of conditions. ²⁵ This system of yoga includes a modified approach to performing the asanas (postures) for individuals who have physical challenges, with the use of props and supports, adapting the poses to the age and fitness levels of participants with modifications made on the basis of individual needs. ²⁶ The various styles of yoga interventions for cancer survivors incorporate gentle or restorative postures, typically emphasizing the relaxation component of the yoga practice. This relaxation response is an important component in the Iyengar tradition; however, Iyengar yoga also involves combinations of static and active stretching and isometric and dynamic strengthening to increase strength, flexibility, stability, and balance.^17,27,28

The protocol for this trial was Iyengar inspired, designed according to stationary alignment principles and the isolation demarcation of the body into zones for therapeutic focus (eg, spine, upper extremities, and lower extremities). The protocol diverges slightly from the Iyengar system in that breath work (pranayama) is practiced as a crucial element in the protocol and is used in conjunction with physical postures (asana). In the Iyengar system, pranayama is practiced in isolation. The purpose of physical practice and breath work was to decrease the side effects of treatment (joint and muscle pain, stiffness, and fatigue) and to increase body awareness through the breath. The sense of awareness and empowerment may increase the positive impact of clinical interventions in individuals with side effects from chemotherapy. The Iyengar-inspired protocol was developed through previous yoga pilot studies in knee osteoarthritis^17,29 and reviewed by a panel of experts in physical therapy (PT) and yoga. By connecting breath with movement, this protocol aimed to reacquaint the participant with a sense of wholeness and greater body awareness (Table 1).

Table 1.

Yoga Class Structure and Components

Upper-extremity focus 10 minutes Seated nadi shodana and focusing 10 minutes Upper-back movements—setu bhadasana Child’s pose extended, breath into upper back Gentle twists/arm binds Standing (focusing on opening upper back, chest, and arms) (ii). Virabhadrasana I (arms up variation) (iii). Trikonasana (arms wide variation) (iv). Adhomuka savasana (knees bent) (v). Padottasana (arms behind back) (vi). Ardha uttanasana (half-way up) Seated (focus on upper back, chest, and extremities) (i). Navasana (supported) (ii). Dandasana (iii). Janu sirsasana/pascimottanasana (iv). Wrist work Supine twist (thoracic focus) Savasana

Lower-extremity focus 10 minutes Seated nadi shodana and focusing 10 minutes Lower-back movements—setu bhadasana Child’s pose toes underextended, breath into lower back Standing (focusing on opening lower back and hips) (ii). Virabhadrasana I (arms up variation) (iii). Ajaneasana (knee down lunge) (iv). Lunges with knee at wall on blocks (v). Adhomuka savasana (against wall or straightening legs) (vi). Padottasana (arms under shoulders/twists) (vii). Garudasana (chair balance variations—no arms) (viii). Bhekasana Seated (focus on lower back and hips) (i). Bhujangasana (supported variations) (ii). Bhadha konasana/Janu sirsasana (forward extension) (iii). Ankle and foot rolls Supported backbends with blocks Supine twist (lumbar focus) Savasana

Spine focus 10 minutes Seated nadi shodana and focusing 10 minutes hip and back movements—setu bhadasana Sukasana with twists and side extensions Child’s pose side to side Cat/Cow Adhomuka sanasana—bent knees Standing (ii). Trikonasana I (iii). Parsvakonasana (iv). Padottasana (with extended twists) Seated (i). Dandasana (ii). Pascimottanasana (iii). Praserita Padattonasana (with side extensions) Supported backbend series on block Supta pada gustasana with wall Viparita korani (elevated) Savasana

Home practice options (15-30 minutes) Lower-back movements (ii). Lunges with knees at wall (iii). Lower back twists (iv). Savasana Upper-back movements (i). Wrist work sequence (ii). Arm and shoulder work with blocks (iii). Upper-back twists (iv). Savasana

Participants met twice per week for 8 weeks, and sessions were held in 2 locations in the community (Gilda’s Club and Yoga Nine) and taught by certified yoga instructors. The yoga program had between 5 and 10 participants, lasted 90 minutes, and was structured precisely as shown in Table 1. Our instructors received more than 500 hours of yoga training prior to the initiation of this research and were registered with Yoga Alliance. They ensured that each participant achieved the exact positioning required and that they used the following props correctly: a mat, 2 bolsters, 2 chairs, 2 blankets, 1 eye bandage, belt, 2 blocks, and a 1-lb weight. These props are only assistive devices to ensure proper posture and alignment, and their use depended on an individual level of ability and flexibility. For example, women experiencing greater levels of arthralgia pain required more props because they were not able to achieve the desired posture otherwise. As the participants progressed through the program, they required fewer props to perform the asanas.

Overall, this program is designed to rest the participant in supported postures and promote breath training, facilitating a relaxation response. Pranayama or regulation of breath was taught to the participant while in the upright seated position. As the focus on this study is on postmenopausal women currently experiencing AIAA, the yoga is centered on asanas that benefit the musculoskeletal system (Table 1). Borrowing from studies in yoga and osteoarthritis by Kolasinski et al ¹⁵ and Bukowski et al, ¹⁷ the protocol includes particular asanas that promote musculoskeletal flexibility, strength, and balance.

An abbreviated version of the yoga program was introduced for home practice during week 2 of the structured sessions. Adding the home program teaches participants to become increasingly self-sufficient and may increase the likelihood of a yoga intervention decreasing pain. Additionally, this small amount of daily practice has been used in other mind–body therapy trials.^30,31 The home program asked participants to choose and perform 3 out of the 10 asanas with appropriate awareness of breathing. The home program was performed 3 times a week for a total of 15 minutes on days when yoga sessions did not take place (bolded elements of Table 1 for various asanas). These asanas were chosen because they are simple and require few props. Adherence to the home-based yoga program was measured weekly by phone call and e-mail contact with the participants. Participants recorded their daily practice in a journal and provided insight into their experiences with the yoga classes and home practice.

Data Collection and Outcome Measurement

Primary outcomes

The Functional Reach (FR) and Sit and Reach (SR) were evaluated by trained physical therapists and served as the primary outcomes. Although these tests have not been used in BCSs with AIAA, the FR and SR have been previously used in clinical trials of yoga for osteoarthritis and were sensitive to functional changes over time.

The FR and SR were used to quantify balance and flexibility, respectively. The FR test examines the balance of individuals with respect to the inverted cone and the patient’s limit of stability. Each participant was instructed to flex the upper extremity forward to 90° and to reach forward as far as possible before taking a step. The reach is determined by the total excursion of the third metacarpal from the starting point (with the hand held in a fist) to the point just before balance is lost. The average of 3 measurements is used as the final score.^32,33 The FR has been shown to possess predictive validity for falls.^33,34 Individuals with FR less than 25.4 cm are 8 times more likely to fall than those able to surpass this distance.^33,35 Studies of FR have demonstrated strong reliability and validity for measuring postural control in reaching forward during standing.^36,37 The FR test has shown criterion validity, predictive validity, test-retest reliability, and interobserver reliability for younger and older adults ³⁸ and possesses attributes that can make it a meaningful and accessible piece for balance assessment. ³⁷

The SR test is the most widely used measure of hamstring and lower-back flexibility, examining the maximal reach an individual can make in a seated position.^36,39 This test is included in most health-related fitness studies because of the evidence that maintaining this flexibility may prevent acute and chronic musculoskeletal injuries, lower-back problems, postural deviations, gait limitations, and risk of falling. ⁴⁰ Participants are asked to maintain a long-sit position on the floor while safely reaching forward as far as possible.^33,40 Shoes are removed, feet are flat against the box, and legs are straight. There are established norms for this by the American College of Sports Medicine. ⁴¹

Secondary outcomes

Self-reported Patient Specific Functional Scale (PSFS), Functional Assessment of Cancer Therapy–Breast (FACT-B), and Brief Pain Inventory (BPI) were patient self-reported outcomes used to measure intervention effects. Surveys were administered at baseline and at week 8 by PT research assistants.

The PSFS was used to measure the participant’s perceived level of disability by identifying functional activity limitations specific to the individual.^42-44 Participants were asked to identify in writing up to 5 functional activities that were important to them and in which they had difficulty as a result of their current condition. Participants ranked the level of difficulty for each activity on a 0 (inability to perform the activity) to 10 scale (fully functional).^42,43,45,46 On reevaluation, the participant ranked the same activities again using the 11-point scale.^44,45,47 The PSFS has excellent validity, test-retest reliability, and sensitivity to change,^42,44,45,48 PSFS has also demonstrated more sensitivity than a global rating of change score.^44,49

The BPI was used as a multidimensional pain scale to measure pain severity and interference with function.^42,50 The BPI Pain Intensity Scale includes 4 numeric rating scales scored 0 to 10, which measure current pain, worst pain, least pain, and average pain during the last week. ⁵¹ The BPI Pain Interference Scale has 7 items and also uses a 0 to 10 numeric scale. ⁵⁰ This pain intensity scale has shown high internal reliabilities of 0.78 to 0.97 in cancer patients and has also shown high internal reliability of 0.95 in a sample of arthritis patients.^50,51 The BPI has high internal reliability and validity in adult cancer patients ⁵¹ and has been used to assess cancer pain, noncancer pain, complex regional pain syndrome, and osteoarthritis. ⁵⁰

The FACT-B was used as a disease-specific questionnaire for this BCS population and measures health-related QOL (HR-QOL). It is a 27-item compilation of questions divided into 4 primary QOL domains: Physical Well-Being, Social/Family Well-Being, Emotional Well-Being, and Functional Well-Being. Also, 10 questions are added that are specific to the breast cancer population, yielding a total of 27 questions, and it has good psychometric properties. ⁵²

Statistical Analysis

Data analysis was performed using Statistical Package for the Social Sciences 18.0 (SPSS Inc, Chicago, IL). Normality of data was confirmed through residual plots. Descriptive and inferential statistics were performed. The primary comparison between baseline and 8 weeks was analyzed using a paired t test. All statistical tests were 2-sided. We calculated the sample size based on the assumption that the standard deviation will be larger than or equal to half of the mean. If yoga could have caused 40% improvement in functional measures, we would need 10 participants to detect such a difference using a 2-sided significance level of .05 and a power of 80%. As this is a phase I trial, to demonstrate feasibility and safety, we chose not to adjust for multiple comparisons.

Results

Patient Characteristics

Between February and June 2010, 12 patients were referred from local cancer centers, and 13 responded to various local advertisements. Of this group, 5 did not return screening telephone calls, 6 did not meet eligibility criteria, and 4 had recently stopped AIs. Finally, 10 participants consented and enrolled; the demographics are given in Table 2.

Table 2.

Demographic and Clinical Characteristics of Clinical Trial Participants (N = 10)

Median age, years (minimum-maximum)	58 (50-71)
Race, no. (%)
White	90 (90%)
Black	1 (10%)
Employment, no. (%)
Full-time	4 (40%)
Part-time	2 (20%)
Not currently employed	4 (40%)
Education
High school or less	4 (40%)
College	5 (50%)
Graduate or professional school	1 (10%)
Aromatase inhibitors, no. (%)
Anastrozole	4 (40%)
Letrozole	5 (50%)
Exemestane	1 (10%)
Stage of breast cancer, no. (%)
I	2 (20%)
II	7 (70%)
III	1 (10%)
Joint(s) with most pain
Knee/Hip	2 (20%)
Wrists/hands/elbow	4 (40%)
Ankles/feet	4 (40%)

The median age of the women enrolled was 57 years and ranged from 50 to 71 years. Nine women (90%) were non-Hispanic white, and 1 (10%) was non-Hispanic black; 6 (60%) were currently employed, 6 (60%) had college schooling, and 1 (10%) was retired; 2 (20%) reported that the worst joint pain was in the knees, 4 (40%) in ankles/feet, and 4 (40%) in the wrists/hands/elbows; 1 woman had multijoint diffuse joint pain (10%). In all, 8 participants completed 12 of the 16 sessions, 1 completed 9 treatments (was the caregiver of a severely ill parent), and 1 completed 3 sessions but withdrew because of environmental sensitivities. The scores were determined for all women in the final assessment—with the exception of 1 patient who had missing data at the 8-week measurements—and all outcomes were analyzed.

A total of 80% of participants self-reported adherence to the home-based program of 15 minutes, 3 times per week.

Safety

One individual had right upper-extremity lymphedema prior to yoga class; however, she was able to participate in all classes, and there was no worsening of her condition based on patient self-report and PT clinician exam. This patient received PT treatment for her right upper-extremity lymphedema 1 year prior to the yoga intervention. She wore a compression garment during all classes. Upper-extremity circumferential measurements remained the same before and after yoga sessions.

Improvement in Function

From baseline to the end of intervention, participants had significant improvement in FR (25.36 to 39.19; P = .048) and flexibility measured by SR (22.90 to 30.10; P = .009). The PSFS also improved from 4.55 to 7.21 (P < .05). The most common reported physical activity that participants wanted to improve was walking (60%), followed by lifting (40%), opening jars/windows (40%), stair climbing (30%), household chores (30%), exercising (30%), sleeping (20%), and driving (10%).

Changes in Other Symptoms Related to Pain and QOL

Participants also experienced a significant reduction in pain severity (3.90 to 2.79; P < .05). Although not statistically significant, change in pain interference as measured by the BPI demonstrated a trend toward improvement (2.75 to 1.45; P = .07). The subjective report of QOL measured with the FACT-B significantly improved (89.33 to 106.05; P < .05; Table 3).

Table 3.

Change in Function and Other Symptom Outcomes (N = 10)

Measure	Baseline, Mean (SD)	Week 5, Mean (SD)	P Value a
Functional reach	24.36 (16.37)	39.19 (14.62)	.048
Sit and reach	22.90 (12.98)	30.10 (12.3)	.009
PSFS	4.40 (1.86)	7.21 (2.00)	.004
BPI: Pain severity	3.90 (1.63)	2.79 (1.84)	.016
BPI: Pain interference	2.75 (2.77)	1.45 (1.84)	.070
FACT-B	89.33 (20.18)	106.05 (17.02)	.015

Abbreviations: SD, standard deviation; PSFS, Patient-Specific Functional Scale; BPI, Brief Pain Inventory; FACT-B, Functional Assessment of Cancer Therapy—Breast.

a

P value was evaluated based on a paired t test.

Discussion

In this study, we demonstrated the feasibility and safety of a yoga intervention for AI-related arthralgia as well as preliminary effects on functional outcomes. Yoga appears to substantially improve balance and flexibility, and there are perceived improvements in various functional activities of daily living. Patients reported reduction in joint pain severity and improved HR-QOL. Yoga was well tolerated, and no adverse events occurred during the study.

The FR test of balance was used in one other study using yoga 3 days per week for 24 weeks to reduce hyperkyphosis. This study found no difference between patients undergoing yoga therapy and controls. ⁵³ We measured FR in persons with chronic lower back pain and found trends of improvement after a 6-week yoga intervention. ²⁹ Others have shown that older adults with minimal reach ability are 8 times more likely to fall than those with a reach that is more than 25.4 cm. ³⁵ In our study for women with AIAA, baseline functional measures showed that 60% of the participants had a baseline FR below 20 cm and thus were at elevated risk for fall. After the 8-week yoga intervention, only 20% remained at risk. Other associated comorbidities may affect balance, including postural instability in women treated with chemotherapy. ⁵⁴ Balance and fall risk have a direct correlation to functional outcomes and potential disabilities and warrant further investigation.

A limited number of research studies have shown that yoga may improve strength and flexibility and may help control physiological variables such as blood pressure, respiration and heart rate, and metabolic rate to improve overall exercise capacity.^27,55 One Hatha yoga study specifically addressed flexibility in BCSs, but no differences were seen between an exercise and yoga intervention as both groups improved over time. ¹² Our pilot study found that yoga increased flexibility as measured by the SR test in this cohort of women affected with AIAA, although the lack of a control or comparison group limits the validity of these findings.

The PSFS is intended to complement the findings of generic or condition-specific measures and provides a method for eliciting, measuring, and recording descriptions of patients’ disabilities.^56,57 In our study, we found an improvement of 2 or more points in the PSFS score, which indicates statistically and clinically important changes in patients. ⁴⁴ In all, 60% of the women in our study identified walking as a primary concern through the PSFS, and 60% of participants in this trial also cited knees/ankle/feet as the joint sites with worst pain. As AIAA affects both lower and upper extremities and multiple joints, ² PSFS allowed the women to identify the specific areas of functions they seek to improve.

Positions and relaxation breathing based on yoga have been found to significantly reduce fatigue and pain levels in addition to increasing levels of invigoration, acceptance, and relaxation.^18-20 Yoga (both asanas and meditation) may be particularly beneficial for BCSs with AIAA as it is generally nonselective and could alleviate a wide range of additional cancer-related sequelae, such as psychological adjustment, which could address coping strategies for pain management.

Preliminary evidence suggests that yoga results in modest improvements in cancer-related symptoms and overall HR-QOL^19,21; however, only 3 small studies have examined Iyengar yoga in BCSs.^22-24 As suggested by our pilot trial, yoga not only may improve pain severity and HR-QOL in women with AIAA, but the improvement in functional outcomes may hold particular clinical relevance for this population as they age, in the context of severe estrogen deprivation.

There are several limitations to this study. This single-arm pilot study was conducted to evaluate issues that will be important in the design of a study to examine the effect of yoga on AIAA. The sample size of this pilot study was not intended for an efficacy analysis but rather to obtain an estimate of the effect size and variance necessary to plan a definitive study to test and refine individual components of the yoga protocol for AIAA and measurement tools. The lack of a control group negates the ability to exclude the argument that the findings were caused by placebo or Hawthorne effects or regression to the mean effect. Furthermore, we did not measure long-term follow-up data after 8 weeks to determine the sustainability of the structured sessions and the home-based yoga program, and this is important considering the long-term use of AI and management of musculoskeletal pain. Finally, a number of the activities related to group interaction, such as those that occur in a yoga class, may have played a role in the overall improvement of the participants through nonspecific effects. Future researchers might consider additional control arms that include education or other types of attention control to help elucidate the specific effect of yoga. This type of early-phase yoga trial offers an important opportunity to test recruitment and retention strategies and to refine the yoga protocol. Only by accomplishing this can we appropriately design and power a randomized controlled study of yoga for pain management and measurement of functional outcomes for BCS.

In summary, we have conducted a successful pilot trial of yoga to treat AI-related arthralgia. As AIAA is a relatively new clinical phenomenon, and very few intervention studies have been performed to address this clinical problem, we believe that our effort is an important first step in demonstrating the feasibility of rigorous evaluation of yoga for this indication. Future randomized controlled trials are needed to establish the comparative efficacy of yoga to improve functional outcomes related to AI-related arthralgia, which is a clinical problem affecting hundreds of thousands of BCSs.