Article Text


EA versus sham acupuncture and no acupuncture for the control of acute and delayed chemotherapy-induced nausea and vomiting: a pilot study
  1. Chris McKeon1,
  2. Caroline A Smith2,
  3. Kristen Gibbons3,
  4. Janet Hardy4
    1. 1Mater Health Services, Mater Cancer Care Centre, South Brisbane, Queensland, Australia
    2. 2National Institute of Complementary Medicine, University of Western Sydney, Campbelltown, Australia
    3. 3Data Management and Analysis Team, Mater Research Office, South Brisbane, Australia
    4. 4Palliative and Supportive Care, Mater Health Services, South Brisbane, Australia
    1. Correspondence to Chris McKeon, Mater Cancer Care Centre, Mater Health Services, Level 2 Corporate Services Building, Raymond Terrace, South Brisbane, Queensland 4101, Australia; Chris.McKeon{at}


    Objective To assess the feasibility of undertaking a high-quality randomised controlled study to determine whether EA gives better control of delayed chemotherapy-induced nausea and vomiting (CINV) than sham EA or standard antiemetic treatment alone.

    Methods Patients having their first cycle of moderately or highly emetogenic chemotherapy were randomised to EA, sham EA or standard care. EA was given for 30 min on day 1 at the time of chemotherapy and on day 3 using standard acupuncture points bilaterally. Sham EA was given to points adjacent to true EA points. All patients received usual care, comprising antiemetics, according to hospital guidelines. The primary outcomes related to study feasibility, and the clinical outcome measure was the change in Functional Living Index Emesis (FLIE) score captured on days 1 and 7.

    Results 153 participants were screened between April 2009 and May 2011. Eighteen patients did not meet the inclusion criteria, 37 declined to participate and the absence of an acupuncturist or lack of consent from the treating oncologist excluded a further 38 patients; 60 patients were recruited. The FLIE was completed on day 7 by 49 participants; 33 of 40 patients returned on day 3 for treatment. The nausea and vomiting scores were low in all three arms. Adverse events were generally mild and infrequent.

    Conclusions It was feasible to undertake a randomised EA trial on a busy day oncology unit. As few patients experienced nausea with their first cycle of chemotherapy, it was not possible to determine whether EA improves CINV over standard care. An enriched enrolment strategy is indicated for future studies. A simple numerical rating scale may prove a better objective nausea measure than the FLIE.

    Trial registration number ACTRN12609001054202.


    Statistics from


    Nausea and vomiting are rated as highly distressing and debilitating adverse effects identified by patients receiving chemotherapy,1 ,2 and continue to be a major concern for patients despite new and improved antiemetic therapy.3 ,4 w26 Overall, the prevalence of acute and delayed chemotherapy-induced nausea and vomiting (CINV) ranges from 40 to 60% and 40 to 80%, respectively.1 Newly developed antiemetics, such as aprepitant, have increased the efficacy in the control of acute and delayed CINV as compared with standard treatment with 5-HT3 antagonists and a corticosteroid.5 A systematic review of 17 trials of aprepitant for moderate and highly emetogenic chemotherapy showed complete response rates for 72% of participants compared with 54% in the control arm of ondansetron and dexamethasone without aprepitant.5 A recent review6 restated that control of CINV has improved but that patients still experience it, particularly nausea, and further investigation into behavioural interventions for control, such as acupuncture, is needed.

    The latest systematic review of acupuncture7 supported the finding from a Cochrane review by Ezzo et al,8 which identified 11 trials of acupuncture point stimulation for CINV. Acupuncture-point stimulation reduced the proportion of patients experiencing acute vomiting but not the mean number of acute emetic episodes or acute or delayed nausea severity compared with controls. EA reduced the proportion of acute vomiting but MA did not. Unfortunately, most of the EA studies did not use modern antiemetic drugs. The reviews7 ,8 concluded that further studies are needed evaluating the use of acupuncture compared with the concurrent use of modern antiemetics and, evaluating acupuncture treatment protocols using more than one acupuncture point. The review also noted that few of the studies examined the benefit of EA on delayed CINV.

    This study was designed to address the problems identified from both reviews.7 ,8 In addition to the use of modern antiemetics, the inclusion of three study groups allowed examination of the adjunctive effect of EA, compared with usual treatment in controlling CINV, and second would determine if the effects of EA were greater than sham EA.9 w27 This initial investigation was planned as a pilot study to assess the feasibility of recruiting and treating patients within a day oncology unit and to examine the appropriateness of data collection tools with a view to performing a larger definitive trial at a later date.


    Study design

    This was a single-centre, pilot, randomised trial of EA, sham EA and no acupuncture. The study examined recruitment and attrition, suitability of outcome measures and the integrity of blinding in a study aimed at determining whether EA reduced CINV more than sham EA or no acupuncture as measured by a change in the Functional Living Index Emesis (FLIE) tool between days 1 (the day of chemotherapy) and 7. Secondary outcomes included nausea scores, number of vomits, toxicity of acupuncture and FLIE score on days 1, 4 and 7. Mater Health Services human research ethics committee approved the study (approval number 1276A). The trial was registered with ANZCTR (Australian New Zealand Clinical Trials Registry), number ACTRN12609001054202.

    Study population

    The study included patients with cancer, aged ≥18 years, receiving first time moderate to highly emetogenic chemotherapy on non-consecutive days, who had a good understanding of English and the cognitive ability to complete questionnaires. Exclusion criteria were a history of acute, chronic or subacute ileus, platelets <100×109/L, previous acupuncture for nausea and vomiting, pre-existing nausea or vomiting from another medical condition, concurrent radiotherapy, uncontrolled central nervous system involvement by tumour, needle phobias, or insufficient understanding of the trial and requirements. Patients were recruited from the Mater Health Services Adults Day Oncology Unit, Brisbane, Australia. Patients were informed in person and by an information sheet that they would be randomised to one of three arms: standard treatment, standard treatment plus sham acupuncture or standard treatment plus true acupuncture.


    Patients were stratified according to the emetogenic rating (moderate or high) of the chemotherapy.10 w28 The stratified, block randomisation schedule was computer generated and prepared by the Mater research office by a biostatistician not involved in the conduct of the study. The randomisation was concealed using sealed envelopes, numbered consecutively according to patient entry number and was kept on the hospital day unit. The envelopes contained the randomised treatment allocations and were opened by the acupuncturist after each participant's consent.


    A research assistant who remained blinded to the treatment allocation collected the data from the participants. A blinded independent data manager entered the information from the trial diaries into a database. Patients were blind to the allocation of sham or true EA, as described below. The adequacy of blinding was tested at day 7, by obtaining a participant's belief about group allocation, by marking true acupuncture, sham acupuncture or unknown. The acupuncturist was not blind to the treatment allocation.


    The points, timing, duration and frequency of treatments were chosen based on evidence from previous trials, systematic reviews, workflow in the unit and using traditional Chinese medicine textbooks to create a standardised protocol. PC6 (Neiguan) has been widely investigated as the major point for treatment of nausea and vomiting with different causes, such as pregnancy, postoperative and chemotherapy induced.11–13 w29 ST36 (Zusanli) is the other point commonly used for CINV.14 ,15 w30 ST36 is one of the main points used in treating any stomach disorders or patterns.16 Li16 has identified LR3 (Taichong) and LI4 (Hegu) as useful for nausea; these two points may have a calming effect;17 w31 LR3 is also traditionally used for abdominal problems.16

    In the true acupuncture group, EA was applied (Hwato SDZ-III Electronic Acupuncture Treatment Instrument) at ST36, PC6, LR3 and LI4 using sterile surgical stainless steel 0.20 mm×30 mm acupuncture needles (Sensei brand, made in China for Acuneeds Australia) bilaterally, a total of eight needles. Once inserted, the needle was manipulated using a thrusting, twirling movement until a de qi sensation was achieved and the patient felt soreness, fullness, heaviness or local area distention. The depth of needle insertion varied between patients and acupuncture points but was generally to a depth of 0.2–1.5 cm. The negative pole of the electroacupuncture unit was attached to PC6 and the positive pole to ST36 to form one circuit, the negative pole to LI4 and the positive pole to LR3 to form another; the procedure was applied bilaterally. The frequency of stimulation was 10 Hz with the intensity of stimulation adjusted according to the patient's tolerance (maximum 10 mA). Stimulation started 10 min before chemotherapy and continued for a total of 30 min on the first cycle (day 1). Patients were invited to return to the day unit for a second course of EA on day 3, given as on day 1.

    Sham EA was applied at points adjacent to the true acupuncture point, thought to be far enough away to be ‘off channel’ and to negate any antiemetic effect. Each point was defined by the corresponding acupuncture points and measurements—that is, 1 cun (the width of the patients’ thumb knuckle) lateral to PC6, midpoint between ST36 and GB34, 1 cun medial to LI4 and 1 cun lateral to LV3. Sham EA was applied using sterile surgical stainless steel 0.20 mm×30 mm acupuncture needles (Sensei brand, made in China for Acuneeds Australia) at points bilaterally. The needles were inserted superficially, 0.2–0.5 cm depth, they were not manipulated and no de qi sensation was obtained. The circuit was set up as in the treatment arm using a non-functioning EA machine with the amplitude turned up on an unconnected channel, to give the appearance and noise of the actual machine giving a treatment. Sham stimulation as described above was given according to the same regimen as the true EA.

    All acupuncture was administered by CMK, a trained acupuncturist with an advanced diploma of acupuncture, who is a member of the Australian Acupuncture and Chinese Medicine Association with 6 years of clinical practice at the time of the study.

    All participants, including those in the non-acupuncture group, received standard antiemetic medication treatment without acupuncture according to the Mater Health Services hospital cytotoxic prescribing and administration programme of antiemetic protocols.w32 All patients received rescue antiemetics according to protocol, if required. Standard antiemetic treatment at the time was ondansetron 8 mg IV on day 1 and 8 mg orally twice a day on days 2 and 3, and dexamethasone 8 mg IV on day 1 and 8 mg orally daily on days 2 and 3 for moderate emetogenic chemotherapy. Patients receiving highly emetogenic chemotherapy had the same regimen plus aprepitant 125 mg orally on day 1 and 80 mg orally on days 2 and 3.

    Outcome evaluation

    Feasibility outcomes included the rate and ease of recruitment, the number of participants returning on day 3, completeness of trial data and the attrition rate. The success of blinding to intervention and sham treatments was assessed by questioning a participant's belief about their group allocation on day 3 after completion of the intervention. Participant expectations of acupuncture were documented by patients before treatment as ‘complete improvement’, ‘clear improvement’, slight improvement’, ‘no improvement’ or ‘don't know’.

    The primary clinical outcome measurement tool for CINV was the FLIE,18 a validated, nausea and vomiting-specific, patient-centred measure recorded on days 1 (baseline, at time of chemotherapy), 4 and 7. The primary clinical outcome measure was the relative change in FLIE score from day 1 to day 7. The range of total potential FLIE scores was 7–126 (a high score corresponding to a higher quality of life (QoL), or less impact of CINV on daily life). No or minimal impact on daily life was defined as an average FLIE item score of >6 on the seven-point continuous Visual Analogue Scale or a total FLIE score of >108. A patient diary consisting of numerical rating scales (NRS) assessing nausea, number of vomits and rescue medication use was completed daily. Secondary outcomes included change in FLIE score from day 1 to day 4; nausea score change between days 1 and 2 and days 1 and 4; difference in the number of vomits between days 1 and 2 and days 1 and 4 and adverse events. Basic demographics were collected on day 1 together with patient expectations of treatment to identify if there was a correlation between expected benefit and actual benefit. Details of adverse events were collected using modified questions from the National Cancer Institute Common Terminology Criteria for Adverse Events, on days 1 and 7 for all patients and day 3 for those in the treatment arms.

    Statistical analyses

    All statistical analyses were conducted using SPSS for Windows V.15.w33 This was a pilot study conducted to ascertain feasibility and preliminary data which could be used to determine a sample size for a definitive trial. For this pilot study a sample size of 60 was determined to be convenient and able to be funded. Patient sample characteristics and outcome data were summarised using medians and IQRs for continuous data, as data were not normally distributed, and frequencies for nominal data. Statistical comparison between the three study groups was undertaken using the Kruskal–Wallis test, with the significance value reported. Analysis examined the relative change in FLIE score from day 1 to day 4 and 7, the change in nausea score and number of vomits between days 1 and 2 and days 1 and 4.


    Patient recruitment is shown in figure 1. A total of 153 patients were screened and 60 recruited to the trial. Reasons for declining the study included being ‘too busy’ (10 patients), too anxious about chemotherapy (3 patients), not wishing to receive sham EA (4 patients) and the following (1 patient each): varicose veins in needle site, unable to return on day 3 if in a treatment arm, severe bruising, poor previous acupuncture experience, no nausea and bad previous trial experience. Fourteen gave no reason. Thirty-eight were not recruited because the acupuncturist was unavailable or unable to obtain the consent of the treating oncologist. Eighteen patients did not meet the inclusion criteria. Sixty participants were recruited to the trial over a period of 25 months; 20 were randomised to no acupuncture, 19 to sham EA and 21 to true EA. Thirty-three (83%) patients returned for a second course of EA (16 participants (40%)) or sham EA (17 participants (43%)) on day 3. Complete data (FLIE days 1, 4, 7 and daily diary) were obtained from 49 (82%) study participants for the primary clinical outcome measure (12 no acupuncture, 18 sham EA, 19 true EA). Eleven patients were lost to follow-up, with FLIE forms not completed or returned, with no reason given.

    Figure 1

    CONSORT 2010 flow diagram.

    Patient demographics are summarised in table 1. Patient age was similar in all three groups. The majority of patients were female. The most common cancer diagnosis was breast (62%), followed by colorectal (18%). Cancer diagnosis and the emetogenic potential of the chemotherapy received did not differ between groups. Patient blinding to sham acupuncture or EA was successful, in that 71% of participants were blind to their group allocation. Fifty-eight per cent of participants (14/24 participants) were unsure to which group they had been allocated, 42% (10 participants) were incorrect in their response.

    Table 1


    The changes in FLIE scores are presented in table 2. In all three groups, the FLIE scores after chemotherapy were not as low as expected, indicating good control with standard antiemetic therapy. The change in FLIE score from day 1 to day 7 in each group was as follows: standard care −1.02 (19.4), sham acupuncture −1.05 (21.6), true acupuncture −3.36 (20.8) (median (IQR)) with no significant difference between groups (p=0.589). The change in FLIE score from day 1 to day 4 was as follows: standard care −1.98 (20.4), sham −0.54 (21.1) and true −2.76 (13.6) (p=0.852).

    Table 2

    Results of primary and secondary end points

    Side effects from acupuncture were mild: there were 18 reports of mild pain, 4 of moderate pain and 1 of severe needling pain. No participant discontinued because of needling pain. There were 5 reports of localised bruising, 3 of localised skin irritation and 12 reports of exacerbation of CINV (the majority mild). Patient expectation of benefit had no correlation with change in FLIE score at day 7 (0.055, Spearman's coefficient; p=0.762) or day 4 −0.001 (p=0.996).


    This pilot study has highlighted a number of issues that can now be examined in an appropriately powered RCT.

    Recruitment. Initially, the acupuncturist was available for only 2 days a week (this increased to 4 days later in the trial). The lack of a full-time dedicated research officer limited the completeness of data collection and follow-up. Similarly, patients newly diagnosed with cancer were presented with a large volume of information at the start of treatment. Many patients wanted to participate, but asked if they could start at the time of their second cycle. Entry to this study was dependent on obtaining consent from the treating oncologist in ‘real time’. This was a barrier to recruitment. An alternative approach would be considered in a future study—for example, offering clinicians an ‘opt-out’ consent process. In this process, all patients would be considered potentially eligible unless specifically stated to be ineligible by the consultant in charge. Another aspect of recruitment that the study highlighted was the disproportionate ratio of female to male participants. This ratio is similar to that in the population that uses complementary therapies,19 but did not reflect the population treated in the day oncology unit (3:1 female to male). Future studies need to develop strategies to deal with this imbalance.

    Blinding was feasible using this methodology. Over half (58%) of patients in the treatment arms were unsure of their allocation and in the remainder, there was a low correlation between the participant's belief of allocation and their actual allocation.

    Low incidence of CINV. The incidence of CINV after an initial cycle of chemotherapy was not as high as expected making it difficult to show any difference between arms even if an effect of treatment existed. This might, in part, be because the side effects of chemotherapy are cumulative and not experienced by the majority of patients after the first cycle of chemotherapy. Similarly, modern antiemetic regimens have led to much greater CINV control rates.

    Outcome measure tool. The FLIE measures QoL as part of nausea control. Compared with previous studies,18 ,20 our FLIE scores were high, suggesting that CINV did not have a large effect on QoL. Future studies should consider patients after cycle 2 or 3 and target specifically those patients who have experienced CINV using an enriched enrolment design.21 Furthermore, it has been suggested that the QoL aspects of the FLIE may ‘dilute’ the more objective nausea scores.22 To overcome this, we propose using a NRS scoring nausea rather than the FLIE as the primary end-point measure in future studies.

    Sample size. The data from this pilot study predict that a future definitive study would require a sample size of 171 participants. The sample size calculation is based on the primary outcome of the change in ‘worst’ nausea score between day 4 and day 1. It is expected that patients would have a score of 0 or 1 on day 1 in all arms, that the control arm would have a 2-point change in the ‘worst’ nausea score (shown in previous studies to be clinically significant23 ,24 w34), the sham acupuncture group a 1-point change and the acupuncture arm no change. The expected SD of the change in nausea scores is estimated to be 2.8, based on NRS data from patient diaries in this study and information from the literature.23 ,24 w34 Using these assumptions, together with a type 1 error of 0.05 and 80% power, 39 participants in each group are required. Applying an expected attrition rate of approximately 20%, 49 patients in each group are required. To account for the possibility of non-parametric data and therefore a loss of power, a correction of 1.157 gives a total number of 57 per group, and a total sample size of 171 patients.


    A systematic review of acupuncture in cancer care for treatment of CINV indicates a potential benefit, but further research is needed.25 This pilot study has shown that it is feasible to undertake a controlled EA, sham EA and no acupuncture study on a busy day oncology ward in Australia. However, we identified a low incidence of CINV in the first cycle of chemotherapy with no treatment effects found between groups. Future studies may benefit from using an enriched enrolment design. The feasibility study highlighted the need for adequate resourcing to support staff in conducting the trial. Recruitment challenges, especially when dependent on busy medical oncologists for permission to approach patients, were also a factor, which could be overcome if an opt-out consent process was adopted.

    Summary points

    • We tested the feasibility of conducting a clinical trial of acupuncture for chemotherapy nausea within a busy oncology clinic.

    • The pilot study was successful, and problems with recruitment, enrolment and outcome measures were identified that will inform the definitive trial.

    • A study in this setting is feasible.


    The authors would like to acknowledge and thank the Mater Foundation, the Smiling for Smiddy Foundation and Australian Acupuncture and Chinese Medicine Association (AACMA) for funding for the trial. Thanks also to Drs Shannon, Woodward and Lwin for allowing their patients to participate in the trial, Ms Helen Anderson for substantial contribution to acquisition of the data and Ms Corrine Haugstetter for data management.


    View Abstract
    • 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.


    • Collaborators Corrine Haugstetter; Helen Anderson.

    • Contributors CMK, JH: substantial contribution to conception and design of the trial, acquisition, analysis and interpretation of the data, drafting and revision of the work and final approval. CAS: substantial contribution to the design of the trial, analysis and interpretation of the data, revision of the work and final approval. KG: substantial contribution to the analysis and interpretation of the data, drafting and revision of the work and final approval.

    • Funding Funding generously supplied by the Mater Foundation, South Brisbane, the Smiling for Smiddy Foundation and Australian Acupuncture and Chinese Medicine Association (AACMA).

    • Competing interests None declared.

    • Ethics approval Mater Health Service human research ethics committee.

    • 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.