Volume 120, Issue 23 p. 3744-3751
Original Article
Free Access

Electroacupuncture for fatigue, sleep, and psychological distress in breast cancer patients with aromatase inhibitor-related arthralgia: A randomized trial

Jun J. Mao MD, MSCE

Corresponding Author

Jun J. Mao MD, MSCE

Abramson Cancer Center, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania

Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania

Department of Family Medicine and Community Health, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, Pennsylvania

Corresponding author: Jun J. Mao, MD, MSCE, Department of Family Medicine and Community Health, University of Pennsylvania, 227 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104; Fax: (215) 662-3591; [email protected]Search for more papers by this author
John T. Farrar MD, PhD

John T. Farrar MD, PhD

Abramson Cancer Center, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania

Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania

Search for more papers by this author
Deborah Bruner PhD, RN

Deborah Bruner PhD, RN

Emory University School of Nursing, Atlanta, Georgia

Search for more papers by this author
Jarcy Zee BS

Jarcy Zee BS

Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania

Search for more papers by this author
Marjorie Bowman MD, MPA

Marjorie Bowman MD, MPA

Wright State University School of Medicine, Dayton, Ohio

Search for more papers by this author
Christina Seluzicki BA

Christina Seluzicki BA

Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania

Department of Family Medicine and Community Health, Perelman School of Medicine at the University of Pennsylvania, Philadelphia, Pennsylvania

Search for more papers by this author
Angela DeMichele MD, MSCE

Angela DeMichele MD, MSCE

Abramson Cancer Center, Hospital of the University of Pennsylvania, Philadelphia, Pennsylvania

Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania

Search for more papers by this author
Sharon X. Xie PhD

Sharon X. Xie PhD

Department of Biostatistics and Epidemiology, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania, Philadelphia, Pennsylvania

Search for more papers by this author
First published: 30 July 2014
Citations: 105

Abstract

BACKGROUND

Although fatigue, sleep disturbance, depression, and anxiety are associated with pain in breast cancer patients, it is unknown whether acupuncture can decrease these comorbid symptoms in cancer patients with pain. The objective of this study was to evaluate the effect of electroacupuncture (EA) on fatigue, sleep, and psychological distress in breast cancer survivors who experience joint pain related to aromatase inhibitors (AIs).

METHODS

The authors performed a randomized controlled trial of an 8-week course of EA compared with a waitlist control (WLC) group and a sham acupuncture (SA) group in postmenopausal women with breast cancer who self-reported joint pain attributable to AIs. Fatigue, sleep disturbance, anxiety, and depression were measured using the Brief Fatigue Inventory (BFI), the Pittsburgh Sleep Quality Index (PSQI), and the Hospital Anxiety and Depression Scale (HADS). The effects of EA and SA versus WLC on these outcomes were evaluated using mixed-effects models.

RESULTS

Of the 67 randomly assigned patients, baseline pain interference was associated with fatigue (Pearson correlation coefficient [r]=0.75; P < .001), sleep disturbance (r=0.38; P=.0026), and depression (r=0.58; P < .001). Compared with the WLC condition, EA produced significant improvements in fatigue (P=.0095), anxiety (P=.044), and depression (P=.015) and a nonsignificant improvement in sleep disturbance (P=.058) during the 12-week intervention and follow-up period. In contrast, SA did not produce significant reductions in fatigue or anxiety symptoms but did produce a significant improvement in depression compared with the WLC condition (P=.0088).

CONCLUSIONS

Compared with usual care, EA produced significant improvements in fatigue, anxiety, and depression; whereas SA improved only depression in women experiencing AI-related arthralgia. Cancer 2014;120:3744–3751. © 2014 American Cancer Society.

INTRODUCTION

Arthralgia, or joint pain, is a common and bothersome side effect of aromatase inhibitors (AIs)1 experienced by nearly 50% of postmenopausal breast cancer patients who receive AI therapies.2 Among all of the side effects associated with AIs, arthralgia is the most discussed symptom in online, breast cancer-specific message boards, which indicates the intrusive nature of this symptom on patients' daily living.3 Furthermore, the experience of arthralgia results in premature discontinuation of AIs,4 which may negatively impact disease-free and overall breast cancer survival.5

Emerging evidence suggests that acupuncture, a therapy originated from traditional Chinese medicine, may be effective in the management of AI-related arthralgia.6-8 Acupuncture appears to be safe, and few minor and self-limiting side effects (eg, needling pain, local bruising) have been noticed in the trials. Although the definitive efficacy of acupuncture for arthralgia related to AI use requires rigorously conducted trials of larger sample sizes, longer follow-up, and refinement of active and placebo interventions, the initial effect sizes observed in these trials indicate potentially clinically meaningful reduction in pain.6, 8

Although it is not established in the setting of AI therapy, research has demonstrated that pain in patients with breast cancer is often associated with fatigue, sleep disturbance, anxiety, and depression.9-12 Biologically, the clustering of these symptoms with pain may be explained by hypothalamic-pituitary-adrenal (HPA) and sympathetic nervous system (SNS) dysregulation.11 In addition, the dynamic correlation between innate immune inflammatory responses and neuroendocrine pathways may account for both the development and the persistence of these symptoms.13 Together, pain and comorbid nonpain symptoms have a synergistic negative effect on quality of life in breast cancer patients.10 In animal models, acupuncture impacts both the HPA axis14 and the SNS system,15 providing a biologically plausible explanation for acupuncture's effect on not only AI-related arthralgia but also on fatigue, sleep, and psychological distress in these women. We analyzed prespecified, secondary, nonpain symptom outcomes in a recently completed phase 2, randomized controlled trial (RCT) to evaluate the short-term effects and safety of electroacupuncture (EA) for AI-related arthralgia compared with a sham acupuncture (SA) placebo and usual care (waitlist control [WLC]).8 We chose EA because of its clear physiologic effect on the endogenous opioid system (eg, enkaphalin and β-endorphin) and pain reduction demonstrated in animal models.16 For the current article, we evaluated the specific effects of EA and SA compared with usual care WLC on fatigue, sleep, anxiety, and depression.

MATERIALS AND METHODS

We previously published primary outcomes from our clinical trial demonstrating the preliminary efficacy of EA on AI-related pain.8 We summarize the methods here for convenience. This study is registered as a National Clinical Trial (clinicaltrials.gov identifier, NCT01013337).

Study Participants

In brief, we conducted a 3-arm RCT (EA, SA, and WLC) from September 2009 through May 2012 at the Abramson Cancer Center of the Hospital of the University of Pennsylvania, a tertiary care academic medical center in Philadelphia. Eligible patients were women with a history of early stage breast cancer (stages I-III) who were currently receiving an AI (anastrozole, letrozole, or exemestane), had joint pain for at least 3 months that they attributed to their AI, reported a worst pain rating ≥4 on an 11-point numerical rating scale (from 0 to 10) in the preceding week, reported at least 15 days with pain in the preceding 30 days, and provided informed consent. We excluded individuals who had metastatic (stage IV) breast cancer or who had a history of a bleeding disorder. The institutional review board of the University of Pennsylvania approved the study protocol.

Study Design

We randomly assigned participants to treatment groups using computer-generated numbers sealed in opaque envelopes. We used permutated block sizes of 3 or 6 to ensure a 2:1 randomization of acupuncture to WLC. Subsequently, for the acupuncture group, the treating acupuncturist opened a second envelope using computer-generated numbers at the first acupuncture visit to determine whether the participant was to receive EA or SA. We educated all participants on joint pain, staying physically active, and continuing with current medical treatments (including prescription and over-the-counter pain medications) as usual. Patients in the WLC group were allowed to receive 10 real acupuncture treatments after 12 weeks of follow-up. Blinding was evaluated by credibility rating at week 8, as previously reported. Participants indicated that they considered both EA and SA to be credible (pain rating, 4.3 vs 4.0, respectively; P=.54).8

Interventions

Two licensed, nonphysician acupuncturists with 8 years and 20 years of experience, respectively, administered interventions twice weekly for 2 weeks, then weekly for 6 more weeks, for a total of 10 treatments over 8 weeks. The detailed protocol has been previously published.8 In brief, for the EA group, the acupuncturist chose at least 4 local points around the joint with the most pain. In addition, at least 4 distant points were used to address nonpain symptoms (eg, depression/anxiety and fatigue) commonly observed in conjunction with pain. The needles (30 mm or 40 mm and 0.25 gauge; Seirin-America Inc., Weymouth, Mass) were inserted until “De Qi” (sensation of soreness, tingling, etc) was reported by the patient.17 Two pairs of electrodes were connected at the needles adjacent to the painful joint(s) with 2-hertz electrostimulation provided by a transcutaneous electrical nerve stimulation (TENS) unit. The needles were left in place for 30 minutes with brief manipulation at the beginning and the end of therapy. SA was performed using Streitberger nonpenetrating needles at nonacupuncture, nontrigger points at least 5 cm from the joint where pain was perceived to be maximal. The acupuncturists avoided eliciting the “De Qi” sensations by only minimally manipulating the needles apart from their initial contact with the skin. Instead of adding a small electric current to the needles, the acupuncturists turned on the dial of the TENS unit to a different channel so that the participant could observe the light blinking without receiving the electricity. The frequency and duration of treatments between EA and SA groups were identical.

Outcome Measures and Follow-Up

Along with the a priori primary outcome, pain intensity and interference, measured by the Brief Pain Inventory (BPI),18 our a priori secondary outcomes, patient-reported outcomes of fatigue, sleep, and psychological distress, were measured at baseline; at weeks 2, 4, and 8 during treatment; and at week 12 (4 weeks after treatment). We measured fatigue using the summative score of the Brief Fatigue Inventory (BFI). This 9-item instrument was designed to assess 1 construct of fatigue severity in cancer and noncancer populations19 on a numerical rating scale ranging from 0 to 10, with 10 indicating the greatest severity or interference. The score of the scale has been identified as reliable and valid in multiple languages and diverse cancer populations. We measured sleep using the global score on the Pittsburgh Sleep Quality Index (PSQI). The 19-item PSQI instrument produces a global sleep-quality score and 7 specific component scores: quality, latency, duration, disturbance, habitual sleep efficiency, use of sleeping medications, and daytime dysfunction. Global scores range from 0 to 21, with higher scores indicating poor sleep quality and high sleep disturbance.20 Psychological distress was measured using the Hospital Anxiety and Depression Scale (HADS), a 14-item scale with 7 items measuring depression and 7 items measuring anxiety.21 The scale uses various response items. It has been demonstrated that the HADS score is both reliable and valid in diverse groups of cancer patients.22

Statistical Analysis

We based sample size estimations for this study on the primary outcome of pain. To detect an effect size of 1.6 at 80% power and a significance level of .05, we needed 18 participants per arm.8 We used an analysis of variance or the Pearson chi-square test to compare baseline variables among groups. To evaluate whether the nonpain patient-reported outcomes were associated with pain intensity and interference, we calculated the Pearson correlation coefficient (r) among those outcomes at baseline. Because our outcome measures were repeated over time, we assessed differences in changes from baseline to week 12 between EA and WLC and between SA and WLC using separate mixed-effects models.23 We treated time and group as categorical variables and included a random intercept term in the mixed-effects model with adjustment for baseline values. Tests of intention-to-treat differences between the EA and WLC arms and between the SA and WLC arms with respect to changes were based on time-intervention interactions in the mixed-effects models. We calculated P values both for joint tests across all time points and for individual tests at each time point. We checked the goodness of fit of the mixed-effects models using residual plots. We did not observe any evidence for violations of any model assumptions or lack of fit. The impact of missing data on the results was evaluated through sensitivity analyses, including last observation carry forward and analyses of completers only, and none of the results differed from our intent-to-treat analyses. We calculated between-group differences with 95% confidence intervals (CIs) at each time point. Two-sided P values are presented and interpreted at a statistical significance level of .05. Statistical analyses were conducted using STATA (version 12.0; STATA Corporation, College Station, Tex) and SAS (version 9.2l SAS Institute Inc., Cary, NC).

RESULTS

We screened 159 patients and enrolled 76 between September 2009 and May 2012, as previously reported8 (for a Consolidated Standards for Reporting Trials [CONSORT] diagram of the study, see Fig. 1). Of the 76 patients who qualified for baseline evaluation, 9 were further excluded (7 had patient-reported pain levels lower than the inclusion criteria, 1 had severe pain unrelated to AIs, and another did not want to participate), and the 67 eligible participants were randomly assigned to EA, SA, or WLC. Among these participants, 21 (95.4%) in the EA group and 20 (90.5%) in the SA group received all 10 treatments. Of all randomized participants, 4 (6%) were lost to follow-up before week 8, and 8 (12%) were lost to follow-up before week 12.8

Details are in the caption following the image

This Consolidated Standards for Reporting Trials (CONSORT) diagram illustrates screening, randomization, and completion of the 8-week and 12-week evaluations.

Baseline Patient Characteristics

Table 1 provides baseline data for the 67 participants. The mean age of the women enrolled was 59.7 years (range, 41-76 years). There were 48 white women (71.6%) and 16 black women (23.9%). Forty-four patients (66%) were receiving anastrozole at the time of randomization, and, on average, participants had been on an AI for 25.9 months (range, 3-56 months). The mean±standard deviation BPI pain intensity score at baseline was 4.9±1.6, and the pain-related interference score was 3.7±2.2. The mean BFI score was 3.7±2.4, the mean PSQI score was 8.2±3.8, the mean HADS-Anxiety score was 6.0±3.8, and the mean HADS-Depression score was 4.0±3.0. Baseline characteristics were well balanced and did not differ significantly among the 3 groups.

Table 1. Baseline Characteristics of the Study Participants
No. (%)
Variable EA, N=22 SA, N=22 WLC, N=23
Age: Mean±SD, y 57.5±10.1 60.9±6.5 60.6±8.2
Racea
White 13 (59) 17 (77) 18 (78)
Nonwhite 9 (41) 5 (23) 5 (22)
Education
≤High school 2 (9) 3 (14) 5 (22)
≥College 20 (91) 19 (86) 18 (78)
Disease stage
I 11 (50) 11 (50) 11 (48)
II 8 (36) 7 (32) 7 (30)
III 3 (14) 4 (18) 5 (22)
AI
Anastrozole 13 (59) 16 (73) 15 (65)
Letrozole 4 (18) 4 (18) 4 (17)
Exemestane 5 (23) 2 (9) 4 (17)
AI duration: Mean±SD, mo 26.9±17.3 19.5±16.9 31.1±22.1
BPI-Pain: Mean±SD score
Intensity 5.1±1.8 4.7±1.7 4.9±1.3
Interference 3.8±2.6 3.4±2.3 3.9±1.7
BFI-fatigue: Mean±SD score 3.7±2.5 3.5±2.5 3.8±2.1
PSQI-sleep quality: Mean±SD score 8.7±4.4 7.6±4.1 8.3±2.4
HADS-psychological distress: Mean±SD score
Anxiety 6.2±3.6 4.8±3.4 6.9±4.2
Depression 4.1±3.0 3.4±2.3 4.3±3.6
  • Abbreviations: AI, aromatase inhibitor; BFI, Brief Fatigue Inventory; BPI, Brief Pain Invenetory; EA, electroacupuncture; HADS, Hospital Anxiety and Depression Scale; PSQI, Pittsburgh Sleep Quality Index; SA, sham acupuncture; SD, standard deviation; WLC, waitlist control.
  • a Race was reported by the participants.

Baseline Correlation Between Pain and Nonpain Symptoms

At baseline, we observed significant correlations between pain and several nonpain symptoms. Fatigue (r = 0.54; P < .001), sleep disturbances (r = 0.28; P = .025), and depression (r=0.35; P=.0037) were associated with pain intensity. Fatigue (r=0.75; P<.001), sleep disturbances (r=0.38; P=.0026), and depression (r=0.58; P<.001) also were associated with pain-related interference. Anxiety was not associated with either pain intensity (r=0.059; P=.64) or pain interference (r=0.20; P=.099).

Changes in Fatigue, Sleep, and Psychological Distress

Table 2 and Figure 2 provide data on the changes in fatigue, sleep, anxiety, and depression at weeks 4, 8, and 12 compared with baseline among 3 treatment groups.

Table 2. Changes in Symptom Outcomes Among all Participantsa
Mean (95% CI)
Change From Baseline Between-Group Difference
Variable EA SA WLC EA vs WLC Pa SA vs WLC Pa
BFI severity .0095 .18
Wk 4 −0.4 (−1.6 to 0.7) −0.5 (−1.7 to 0.7) −0.1 (−0.8 to 0.6) −0.3 (−1.6 to 1.0) .57 −0.4 (−1.8 to 1.0) .38
Wk 8 −1.4 (−2.7 to −0.1) −0.6 (−1.7 to 0.5) 0.5 (−0.2 to 1.3) −2.0 (−3.4 to −0.5) .0034 −1.2 (−2.5 to 0.1) .046
Wk 12 −1.4 (−2.7 to −0.1) −0.7 (−1.6 to 0.1) 0.2 (−0.8 to 1.2) −1.6 (−3.2 to −0.07) .022 −0.9 (−2.2 to 0.3) .091
PSQI .058 .31
Wk 4 −0.9 (−2.2 to 0.4) −0.1 (−1.1 to 0.9) 1.1 (−0.01 to 2.3) −2.0 (−3.8 to −0.3) .0074 −1.2 (−2.7 to 0.2) .10
Wk 8 −1.4 (−2.9 to 0.2) −0.8 (−2.0 to 0.3) 0.1 (−1.1 to 1.4) −1.5 (−3.5 to 0.4) .087 −0.9 (−2.6 to 0.7) .19
Wk 12 −0.8 (−2.3 to 0.7) −1.2 (−2.5 to 0.1) 0.07 (−1.0 to 1.2) −0.9 (−2.7 to 0.9) .20 −1.2 (−2.9 to 0.5) .13
HADS-Anxiety .044 .91
Wk 4 −0.8 (−1.7 to 0.2) 0.2 (−0.6 to 1.0) 0.6 (−0.5 to 1.7) −1.3 (−2.8 to 0.09) .062 −0.4 (−1.8 to 0.9) .50
Wk 8 −1.1 (−2.4 to 0.2) −0.05 (−0.8 to 0.7) 0.2 (−1.0 to 1.5) −1.3 (−3.1 to 0.5) .075 −0.3 (−1.7 to 1.2) .67
Wk 12 −2.1 (−3.5 to −0.7) −0.3 (−1.0 to 0.5) 0.09 (−1.5 to 1.7) −2.2 (−4.3 to −0.1) .006 −0.3 (−2.1 to 1.4) .59
HADS-Depression .015 .0088
Wk 4 −0.5 (−1.5 to 0.6) −0.5 (−1.4 to 0.4) 1.2 (−0.03 to 2.4) −1.7 (−3.2 to −0.1) .039 −1.7 (−3.2 to −0.2) .022
Wk 8 −1.2 (−2.6 to 0.4) −0.8 (−1.7 to 0.2) 1.2 (0.08 to 2.4) −2.4 (−4.2 to −0.6) .0031 −2.0 (−3.5 to −0.5) .0056
Wk 12 −1.1 (−2.4 to 0.1) −1.3 (−2.3 to −0.4) 0.8 (−0.5 to 2.2) −2.0 (−3.8 to −0.2) .011 −2.2 (−3.8 to −0.5) .0024
  • Abbreviations: BFI, Brief Fatigue Inventory; CI, confidence interval EA, electroacupuncture; SA, sham acupuncture; WLC, waitlist control;entory; PSQI, Pittsburgh Sleep Quality Index; HADS, Hospital Anxiety and Depression Scale.
  • a P values were calculated using a mixed-effects model adjusted for baseline outcome values (bolded P values indicate joint tests of a difference in change of outcome over all times between treatment groups; unbolded P values are for individual tests of a difference in change of outcome at each time point).
Details are in the caption following the image

(a-d) Mean changes in fatigue, sleep, anxiety, and depression are illustrated at 4 weeks, 8 weeks, and 12 weeks after baseline according to treatment group. BFI indicates Brief Fatigue Inventory; EA, electroacupuncture; SA, sham acupuncture; WLC, waitlist control; BFI, Brief Fatigue Inventory; PSQI, Pittsburgh Sleep Quality Index; HADS, Hospital Anxiety and Depression Scale.

Fatigue

In the mixed-effects models, EA produced a significant improvement in the BFI score compared with WLC over time (P=.0095), whereas SA did not produce an improvement (P=.18). Compared with the WLC arm, patients in the EA arm had a greater reduction in the BFI score at week 8 (mean difference in reduction, −2.0 points; 95% CI, −3.4 to −0.5; P=.0034; Cohen d=0.96), and the effect persisted at week 12 (mean difference in reduction, −1.6; 95% CI, −3.2 to −0.07; P=.022; Cohen d=0.86).

Sleep

In the mixed-effects models, compared with WLC, both EA and SA produced nonsignificant improvements in the PSQI score over time (P=.058 and P=.31, respectively). Compared with the WLC arm, patients in the EA arm had a nonsignificant improvement in the PSQI score at week 8 (−1.5; 95% CI, −3.5 to 0.4; P=.087; Cohen d=0.63).

Anxiety

In the mixed-effects models, compared with WLC, EA produced a significant improvement in the HADS-Anxiety score over time (P=.044) whereas SA did not (P=.91). Compared with the WLC arm, patients in the EA arm had nonsignificant improvement in the HADS-Anxiety score at week 8 (−1.3 points; 95% CI, −3.1 to 0.5; P=.075; Cohen d=0.53). The change in HADS-Anxiety score was increased and significant between the EA and WLC arms by week 12 (−2.2; 95% CI, −4.3 to −0.1; P=.006; Cohen d=0.68).

Depression

In the mixed-effects models, both EA (P=.015) and SA (P=.0088) produced a significant improvement in HADS-Depression scores compared with WLC over time. Compared with WLC, EA improved HADS-Depression scores by 2.4 points (95% CI, −4.2 to −0.6; p=.0031; Cohen d=0.75) at week 8. Similarly, SA improved HADS-Depression scores as compared with WLC by 2.0 points (95% CI, −3.5 to −0.5; P=.0056; Cohen d=0.89) at week 8. For depression, the effects of both EA and SA were maintained at week 12.

DISCUSSION

In this RCT, we observed that EA produced significant and clinically relevant improvements in fatigue, anxiety, and depression compared with usual care among breast cancer patients who experienced arthralgia related to AI use. In contrast, SA produced a similar significant benefit for depression but not for fatigue or anxiety. These findings have important implications for pain and symptom management and research in cancer patients.

Our study is consistent with a small but growing body of literature suggesting that acupuncture may be effective for pain,6 fatigue,24 anxiety, and depression.25 Instead of using fixed points, we individualized acupuncture treatment based on a protocol and allowed the acupuncturists not only to treat joint pain but also to address symptoms such as fatigue and psychological distress.8 This tailored approach may account for the significant improvement produced by EA on these nonpain symptoms. Our comparison with usual care provided a clinical context for understanding the relevance of these effects. For example, in this study, the Cohen d for fatigue was 0.96, suggesting a large clinical effect; whereas the Cohen d of 0.68 for anxiety suggested a moderately large effect.26 The Cohen d for sleep was 0.63, indicating a moderately large effect, but this did not quite reach statistical significance. Each of these outcomes is highly correlated and supportive of the efficacy of EA. However, given the small group size and the potential concern about multiple comparisons, these results should be interpreted as preliminary. Additional studies will be needed to confirm the efficacy of EA. Our study also included a 4-week no-treatment follow-up period, which suggests that the effect of acupuncture for these symptoms persisted, at least for the short term. This is clinically important, because continued weekly treatments of acupuncture for a long period can be cost-prohibitive and time-consuming for most patients with breast cancer.

Our study also provides a novel understanding of how fatigue, sleep, and psychological distress relate to pain in patients with clinically meaningful, AI-related arthralgia. We observed that pain-related interference was highly correlated with fatigue (r=0.75) and was moderately correlated with depression (r=0.58) and sleep (r=0.35). These findings extend the previous research in patients with breast cancer who undergo chemotherapy10 or who have metastatic/recurrent disease.10 Our findings suggest that clusters of pain, fatigue, sleep, and psychological distress may exist in patients receiving AIs, requiring further evaluation of the shared biobehavioral pathway of these symptoms as well as the impact of these symptoms on patients.

The neuroendocrine-immune mechanism of pain and behavioral symptoms (eg, fatigue, sleep) was proposed by earlier researchers.13 Thornton et al observed that shared variance of plasma levels of cortisol, adrenocorticotropic hormone, epinephrine, and norepinephrine were associated with shared variance of the pain, depression, and fatigue symptom cluster in breast cancer patients who had advanced disease.11 These findings indicate that both the HPA axis and SNS regulation are part of a common mechanistic pathway.11 More recently, cytokine genetic variations were associated with fatigue and depressive symptoms, providing further evidence for potential neuroendocrine-immune mechanisms.27 It has been demonstrated that acupuncture influences the HPA axis14 and SNS28 in addition to potentially exerting an immune-modulatory effect,15 which offers biologic plausibility regarding why it can simultaneously address common pain and nonpain symptoms in cancer patients. Future clinical trials incorporating appropriate experimental designs and biomarkers may help uncover the mechanisms underlying the effect of acupuncture for these common symptoms.

Several limitations need to be acknowledged. Although our original study was powered to detect a difference in pain between EA and WLC,8 it was not powered to detect a statistically significant difference between EA and SA. In this report, we provide the effect size of both EA and SA compared with WLC in fatigue, sleep, anxiety, and depression; but the study is underpowered to detect differences between EA and SA. Therefore, we did not statistically compare effects between EA and SA but provided the magnitude of change in EA and SA so that future research can be powered. In addition, we evaluated multiple outcomes, raising concerns for multiple comparisons. Although our results should be considered supportive of the efficacy of EA, additional studies will be required to confirm these findings. Furthermore, 12% of our participants were lost to follow-up. We performed sensitivity analyses by analyzing completers only as well as carrying last values forward. These results were consistent with our intent-to-treat analyses. Finally, sham control in our study may not function as a physiologically inert placebo because of tactile stimulation of sensory receptors.

Conclusions

Among patients with breast cancer who experience clinically important joint symptoms that are attributable to AIs, pain is associated significantly with fatigue, sleep, and depression. EA has demonstrated promise as a therapy to address these symptoms in addition to pain. These preliminary findings need to be confirmed in a larger trial that includes longer follow-up. Careful incorporation of behavioral instruments and biomarkers in the trial setting can help elucidate the mechanisms underlying both the symptom cluster and the effect of acupuncture.

FUNDING SUPPORT

This study was supported by the National Institutes of Health (National Center for Complementary and Alternative Medicine [NCCAM] R21 AT004695). Dr. Mao is a recipient of the NCCAM K23 AT004112 award.

CONFLICT OF INTEREST DISCLOSURES

Dr. Farrar has consulted for Pfizer and AstraZeneca. Dr. DeMichele has received research support from Pfizer unrelated to aromatase inhibitors. Dr. Xie has consulted for Roche.