Article Text

other Versions


A feasibility study to examine the role of acupuncture to reduce symptoms of lymphoedema after breast cancer: a randomised controlled trial
  1. Caroline A Smith1,
  2. Marie Pirotta2,
  3. Sharon Kilbreath3
  1. 1National Institute for Complementary Medicine, University of Western Sydney, Penrith, Australia
  2. 2Department of General Practice, University of Melbourne, Melbourne, Australia
  3. 3Faculty of Health Sciences, University of Sydney, Lidcombe, Australia
  1. Correspondence to Dr Caroline Smith, National Institute for Complementary Medicine, University of Western Sydney, Locked Bag1797, Penrith, NSW 2751, Australia; caroline.smith{at}


Objective To determine the feasibility, acceptability and safety of using acupuncture to treat arm lymphoedema in women following treatment for breast cancer.

Methods We conducted a randomised controlled trial of acupuncture compared with treatment as usual. Twenty women with stable unilateral intransient lymphoedema present for at least 6 months were recruited from Sydney, Australia. The women received 12 acupuncture treatments administered to body and arm points on the non-lymphoedematous limb over 8 weeks, twice weekly for 4 weeks then once weekly for 4 weeks. Outcome measures included an assessment of interest to participate in the trial, identification of successful recruitment strategies, the appropriateness of eligibility criteria and compliance with treatment attendance. Clinical outcomes were assessed at baseline and 8 weeks and included extracellular fluid, lymphoedema symptoms, well-being and safety.

Results Acupuncture was an acceptable intervention in women with upper limb lymphoedema. Compliance with the treatment protocol was high, with nine women completing all 12 treatments. Outcome forms were completed by 17 women at 8 weeks. No major adverse occurrences, as defined by the study protocol, were reported from the acupuncture group although one woman found the needling uncomfortable, and no study participant experienced an increase in swelling of >10%. There was no change in extracellular fluid or any patient-reported outcome measurement.

Conclusions Lymphoedema is a persistent symptom experienced by women recovering from breast cancer. Our study suggests that acupuncture may stabilise symptoms and no major safety concerns were identified, so further research is needed.

Trial registration number Australian New Zealand Clinical Trials Registry, ACTRN12612000607875.


Statistics from


Lymphoedema is swelling resulting from the disruption of the lymphatic system which is critical to regulation of fluid homeostasis.1 It is a chronic debilitating condition which diminishes quality of life and currently cannot be cured. In western societies the most common cause of lymphoedema is secondary to treatment for breast cancer.2 The incidence of lymphoedema varies depending upon the diagnostic criteria used, but occurs in at least 20% of women undergoing breast cancer treatment.3 It is associated with an abnormal accumulation of protein-rich fluid in the arm due to impaired lymphatic transport.4 This static protein-rich environment promotes bacterial growth, increasing the risk of infection. The current options for treatment and ongoing management of this chronic debilitating condition are limited. Treatment is focused on reduction of swelling and control of pain,3 and may include gentle exercises, compression garments, bandaging the limb and/or massage. These treatments require therapists with expertise in the management of lymphoedema and can be quite costly.5

Preliminary evidence suggests that women are interested in using acupuncture to manage their lymphoedema.6 However, there is controversy regarding the use of acupuncture for arm lymphoedema. People at risk of lymphoedema are advised to reduce the possibility of introducing infection and exacerbating swelling by avoiding accidental and non-accidental skin puncture in the affected area.7 In cancer policy documents, patients are specifically advised to avoid any needling to the affected area.8

In contrast to the recommendations, women have been experimenting with acupuncture without deleterious consequences.6 These reports are supported by small studies which to date have demonstrated that acupuncture does not exacerbate lymphoedema or cause secondary infection.9 ,10 Moreover, acupuncture may be beneficial, with reports that troublesome sensations within the affected arm decrease; however, changes in arm volume have been mixed. Changes in limb volume have been reported in two studies,9 ,10 with a 30% reduction in size between the affected and unaffected arms in a small study with nine participants.10 A subsequent pre/post study examined the efficacy and safety of acupuncture on upper limb lymphoedema in 37 women.11 The mean difference in interlimb circumference at the point of widest difference was reduced by 0.9 cm (range 0.7–1.1 cm; p<0.005), exceeding the 0.4 cm identified as clinically significant.12 Notably, in 11 women the single interlimb circumference difference was reduced by >30% (95% CI 18% to 52%) following acupuncture. In addition, no serious adverse events and infections were found. Fourteen women reported minor complaints including bruising, pain and tingling. Further evaluation of acupuncture using randomised controlled trials is now warranted. The aim of our study was to assess the feasibility, acceptability and safety of using acupuncture to treat arm lymphoedema in women following treatment for breast cancer.


Trial design

A parallel randomised controlled feasibility study was conducted in which acupuncture was compared with usual care. The intervention lasted for 8 weeks.


Women were eligible to join the trial if they met the following inclusion criteria: stable unilateral intransient lymphoedema present for at least 6 months, no use of intensive therapy (eg, decongestive therapy) within the past 3 months, no infections in the lymphoedematous limb requiring antibiotics within the past 3 months, no recent exacerbation of symptoms that led to change in daily activities, and exceeding the previously determined threshold for lymphoedema with bioelectrical impedance spectroscopy (BIS) for at least one 10 cm segment.13 Women were also required to be aged 18 years or older and available to attend treatment over a 2-month period. Exclusion criteria included currently receiving adjuvant therapy for breast cancer or previously receiving acupuncture for treatment of lymphoedema. Women were recruited using two advertisements in newspapers and through two community support groups, press releases and referrals from health professionals in Western Sydney, Australia.


Ten women were randomised to the acupuncture group and received a diagnosis based on Traditional Chinese Medicine (TCM), primarily based on syndrome patterns.14 Women received 12 treatments administered over 8 weeks, twice weekly for 4 weeks then once weekly for 4 weeks. The acupuncturists were experienced university-trained practitioners who were registered with the Chinese Medicine Board of Australia. The protocol was developed by one of the authors (CAS) in consultation with the literature15 and colleagues. CAS also provided the training to the acupuncturists.

The acupuncture points selected included local and distal points, and a combination of standard points and individualised points based on the individual diagnosis. Practitioners were requested to select three standardised points from the following: (1) CV12 Zhongwan, CV3 Zhongji, CV2 Qugu as these points are traditionally associated with regulation of systematic vessels, increased lymphatic circulation and a reduction in lymphoedema; (2) LI15 Jianyu, TE4 Yangchi, LU5 Chize, LI4 Yangxi on the unaffected side to address heaviness, limitation of movement and pain in the upper limbs; and (3) other general points used to treat oedema in general including ST36 Zusanli, SP9 Yinlingquan and SP6 Sanyinjiao. Other acupuncture points were used based on the individual diagnosis. No needle insertion occurred on the affected arm, otherwise acupuncture was applied bilaterally. Needles were inserted for a minimum of 20 min and to a depth dependent on skin thickness and subcutaneous fatty tissue at the site of the acupuncture point. They were stimulated manually on insertion to elicit the needling sensation de qi, and once more during this treatment session. Log books recorded details of all treatments provided. Single-use disposable stainless steel needles (Serin No 3, 0.20×30 mm) were used.

The control group, who received treatment as usual (n=10), were advised to continue with their regular routine for management of their lymphoedema. Both groups were able to continue with regular use of a compression garment. No additional active treatment was to be sought during their 3 months involvement in the study.

Outcome measures

Feasibility outcomes included an assessment of interest to participate in the trial, identification of appropriate recruitment strategies, the appropriateness of eligibility criteria and compliance with treatment attendance. Clinical outcomes were assessed at baseline and at 8 weeks, with a follow-up planned after a further 3 months if project resourcing permitted.

The increase in extracellular fluid was assessed with BIS and arm circumference measures, both of which are highly reliable validated measures.16 BIS (SFB7 instrument, ImpediMed, Brisbane, Australia) was used to quantify the volume of extracellular fluid in each limb using a standardised protocol previously described.16 ,17 Electrode placement was based on the equipotential principle, thereby capturing the extracellular fluid volume of the entire limb. To measure 10 cm segments, electrodes were spaced 10 cm along the arm commencing at the wrist. The measurement electrodes were moved proximally 10 cm along the arm to measure each of the 10 cm segments. To ensure reliability of the measurements, the location of the electrodes along the arm was marked on the skin with reference to a measuring board. Segment A referred to the distal segment adjacent to the wrist and Segment D was the most proximal segment at the top of the arm. Both right and left limbs were measured to allow calculation of an interlimb impedance ratio. Arm swelling was also assessed using circumferential measurements taken at 10 cm intervals from the ulnar styloid to 40 cm proximally on both arms.16 ,17 These measures were used to derive arm volumes from which the interlimb volume ratios were determined.

Participants’ symptoms were assessed with the validated Lymphoedema Symptom Intensity and Distress Survey-Arm and Trunk (LSIDS-AT).18 Participants identified which symptoms were present and rated each identified symptom in terms of intensity and distress on separate 10-point numerical scales. Lymphoedema symptoms were also measured by a scale validated by Norman19 ,20 which assessed symptoms such as puffiness, indentations and sensations such as tired, thick or heavy. Participant well-being was self-assessed using the Measure Yourself Medical Outcome Profile (MYMOP) scale which asks participants to rate their nominated health concerns on a 6-point Likert scale.21

Acupuncture acceptability was measured by a brief series of questions determining what participants liked and disliked about the intervention and study. The safety of the trial was determined by the number and type of adverse outcomes as measured at each visit by the acupuncturist and self-reports from the study participants. Major adverse events to be reported were an increase in lymphoedema of ≥10% and/or infection of the affected arm.

Study size and statistical analysis

No power analysis was undertaken for this feasibility study. The purpose of this study was not to establish efficacy but to carefully measure the effect of acupuncture in a small sample and to test the safety, feasibility and acceptability of acupuncture for the management of lymphoedema. Data analysis was undertaken blind to the study group. Outcome analysis was undertaken based on intention to treat. Baseline demographic and clinical characteristics were summarised and comparisons of the clinical endpoints were made between groups using non-parametric statistics. Levels of significance were reported at p<0.05.


The research assistant randomised the participants to the acupuncture or usual care group immediately after written informed consent was obtained and baseline data collection completed. The randomisation schedule was computer-generated and concealed in sequentially numbered opaque sealed envelopes by a researcher independent of the study.


Recruitment took place between August 2012 and February 2013. Successful recruitment strategies included paid advertisements in local newspapers and referrals from health professionals. Thirty-seven women made contact with the research centre and 20 women were randomised. Reasons for not participating in the trial included living too far away (n=7), not interested (n=3), unavailable (n=2) or did not meet the entry criteria (n=5) (figure 1). Three women withdrew from the trial, including two women in the control group who withdrew immediately following randomisation. Compliance with the treatment protocol was high, with nine women completing all 12 treatments and one woman withdrawing from active treatment after six treatments (no reason given). Outcome forms were completed by 17 women at 8 weeks. Follow-up data were only collected from 12 women as there was insufficient resourcing for staff to undertake final 12-week measurements for the remaining participants; we therefore do not include 12-week follow-up data in this report.

Figure 1

Recruitment and participant flow through the trial.

The participants were mostly in their late fifties or early sixties (age range 30 years), were in the overweight range (body mass index 28–42 kg/m2) and the majority had retired (table 1). All the women had been previously treated for breast cancer, including lymph node dissection. Fifteen women had received radiotherapy. The participants had typically had breast cancer treatment 3–5 years earlier. By chance, women allocated to the acupuncture group had had symptoms of lymphoedema for a significantly longer period of time than those randomised to the control group (median 24 months in the control group and 87 months in the acupuncture group). The women reported an average of two symptoms associated with lymphoedema, the most common being swelling, pain/aching or shoulder stiffness.

Table 1

Baseline characteristics of participants

Acupuncture was acceptable to all the women who participated in the study, although randomisation to the treatment as usual group was not acceptable to all. Ensuring the trial acupuncture was administered within a reasonable travel distance was identified as influencing the acceptability of the intervention. Women reported that they liked contributing to research to help others, being able to try other options of care and having regular contact with project staff (table 2). The majority of participants reported that there were no features about the study which they disliked. Fourteen women indicated they would definitely take part in the study again and only one woman indicated she would probably not participate. All women would recommend a future study to a friend or relative. Additional written comments made by participants reflect the importance of the study and professional team of researchers and acupuncturists: “I would have liked to have received acupuncture but I am still happy I participated anyway.” “It gave me a sense of hope.” “I liked the concern about my general health.”

Table 2

Participants’ views of the study

Additional data (see online supplementary file 1 table S1) indicated that participants held modest views on the role acupuncture might play in improving their health in general. However, when asked specifically if they thought acupuncture could assist with reducing lymphoedema, their expectations were not high. No major adverse occurrences, as defined by the study protocol, were reported from the acupuncture group although one woman found the needling uncomfortable.

Preliminary data from this study indicate that acupuncture did not worsen the symptoms of women in the acupuncture group (figure 2). The change in whole arm BIS ratio in the acupuncture group ranged from an increase of 0.01 to a decrease of 0.30 whereas, in the control group, the BIS ratio changed from an increase of 0.28 to a decrease of 0.43 (figure 2). No study participant experienced an increase in swelling of >10%. There was no change in quality of life or any patient-reported outcome measurement that assessed symptoms related to their lymphoedema (see online supplementary file 1 tables S2–6).

Figure 2

Indivdual pre and post whole arm interlimb impedance ratios.

Circles represent women who received acupuncture and squares represent women who did not receive acupuncture. Filled symbols represent baseline data and open symbols represent data at 8 weeks. Dashed line represents the whole limb BIS threshold for lymphoedema on the dominant side.


Our feasibility study found that acupuncture was an acceptable intervention in women with upper limb lymphoedema after breast cancer treatment despite community concerns about the safety of using acupuncture needles. The interest shown by women to participate in this trial supports the notion that there remains a gap in the management of lymphoedema symptoms. Recruitment of women within a targeted geographical area was successful, in particular recruitment strategies using newspaper advertising and referral from health professionals. Twenty-seven women were keen to join the study, of whom 74% met the eligibility criteria resulting in a high ratio of women being randomised to the trial. Despite an intensive acupuncture regimen requiring attendance once to twice weekly with use of body and arm points on the non-lymphoedematous limb only, compliance with the treatment protocol was very good and the commitment of study participants to the study was also evident with a high return rate of questionnaires during the intervention. The high compliance with completing weekly outcome forms may have been assisted by the research assistant visiting the women in their own homes to undertake arm measurements. However, follow-up to the 12-week point was more challenging but highlighted realistic resourcing of study personnel needed to undertake these measurements. In this small study there was no evidence that the symptoms of women in the acupuncture group worsened, although there was an imbalance at randomisation with the acupuncture group having more established and longstanding lymphoedema symptoms than the control group; it is possible that less symptom variability is evident in symptoms of longer duration.

Our study identified methodological considerations for the design of future acupuncture studies in this population. Slow recruitment is a common experience for recruitment to trials; this study highlighted the importance of initiating multicentre recruitment from the commencement of the study as well as using social and other multimedia approaches to facilitate recruitment.

A challenge for randomised controlled trials in acupuncture is to identify a comparator for the intervention. We chose ‘usual care’; however, following randomisation, two participants withdrew from the study. For women living with lymphoedema for several years, randomisation to a usual care group is not acceptable to all and use of this group needs careful consideration in future studies. When planning a future trial of acupuncture to treat lymphoedema, the choice of control needs careful thought. Comparator groups can control for the course of the health condition, the time and attention that is given to patients, or the insertion and placement of needles.22 The choice of control will be influenced by whether the research question examines the specific function of points and the need for a sham control to place needles away from true acupuncture points used to treat lymphoedema, or whether skin penetration is being tested and thereby non-penetrating sham needles are required. Trials to investigate the efficacy or effectiveness of acupuncture have a few options including usual care where no acupuncture is provided, sham acupuncture in which non-penetrating needles are used or acupuncture using alternative points. A future design could also consider a study arm where needling of the affected limb is undertaken. For each of these alternatives there are advantages and disadvantages and each must be considered in the light of the primary research question.

The diagnosis of lymphoedema is challenging. A range of thresholds is available to identify patients who do and do not present with lymphoedema, with some normative-based and others commonly used but not based on normative values. We required women in the current study to exceed a normative-based interlimb BIS threshold for a single 10 cm arm segment. Consequently, the women included ranged from those with localised early presentation to women with severe longstanding symptoms. Recent studies suggest that the duration of lymphoedema is not a critical factor in the severity of the disease.23 However, underlying tissue changes do affect the volume of fluid present. With disease progression, the fluid volume decreases and architectural changes including localised adiposity and fibrosis develop (intransient lymphoedema). It is currently unclear which group of women may be best suited to acupuncture treatment. To obtain a more homogeneous group of women with lymphoedema, consideration should also be given to disease progression. One classification system is that of the International Society of Lymphology (ISL) which uses the presence of pitting to categorise the level of severity. Pitting is believed to indicate moveable fluid within the superficial tissues. A positive test is an indentation in the tissue indicating interstitial fluid mobility,24 and a lack of pitting is thought to suggest changes in tissue composition in which fluid has been replaced by fibrotic and adipose tissue.25 ,26 Inclusion of women in whom lymphoedema surpasses normative-based thresholds, either based on circumference or BIS, and in whom pitting is present would help to ensure a more homogeneous group of women than in the current study.

Since implementing this study, a pilot study of acupuncture in the treatment of upper limb lymphoedema has been reported by Cassileth et al.11 This study recruited women with breast cancer who developed lymphoedema within a 5-year period. Our study findings may differ from the results of the pilot study11 due to our recruitment of women with intransient lymphoedema over a longer period of time. Both treatment protocols were similar in terms of acupuncture point selection and treatment frequency, although we needled the unaffected side only. The addition of needling the affected side may be associated with a greater treatment effect. This needs testing in a future clinical trial.


Lymphoedema is a common and persistent symptom experienced by women recovering from breast cancer. Our study suggests that acupuncture may stabilise symptoms and no major safety concerns were identified. An appropriately powered trial to evaluate the effect of acupuncture on lymphoedema and related symptoms is needed.

Summary points

  • A feasibility study of acupuncture in women with lymphoedema was conducted.

  • Acupuncture was acceptable and compliance with treatment and outcome measures was good.

  • Two out of 10 women randomised to usual care dropped out immediately.


We wish to thank Beverley de Valois and Charlotte Patterson for contributing advice to the design of the study and treatment intervention, Professor Paul de Souza for his comments on the paper, all the women who participated in the trial, Sarah Falconbridge and Christina Murphy for their assistance with data collection, the study acupuncturists and referring clinicians and the BCNA for distribution of promotional material to their support group.


View Abstract

Review history and Supplementary material

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors CAS was responsible for the conception and design of the study, interpretation of the analysis and drafted the manuscript. SK and MP contributed to the design of the study, interpretation of the analysis and revision of the manuscript.

  • Funding We wish to acknowledge the funding received from the University of Western Sydney.

  • Competing interests SK is funded by National Breast Cancer Foundation, Australia.

  • Patient consent Obtained.

  • Ethics approval The University of Western Sydney Human Ethics Committee approved the research (H9535).

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.