Article Text


Acupuncture versus usual care for postoperative nausea and vomiting in children after tonsillectomy/adenoidectomy: a pragmatic, multicentre, double-blinded, randomised trial
  1. Ingrid Liodden1,
  2. Leiv Sandvik2,
  3. Berit Taraldsen Valeberg3,
  4. Einar Borud1,
  5. Arne Johan Norheim1
  1. 1NAFKAM, Institute of Community Medicine, UiT the Arctic University of Norway, Tromsø, Norway
  2. 2Faculty of Odontology, University of Oslo, Oslo, Norway
  3. 3Faculty of Health Service, Oslo and Akershus College University of Applied Science, Oslo, Norway
  1. Correspondence to Dr Ingrid Liodden, NAFKAM, Institute of Community Medicine, UiT the Arctic University of Norway, Tromsø N-9037, Norway; Ingrid{at}


Objectives To investigate the effect of a standardised acupuncture on nausea and vomiting in children after tonsillectomy with or without adenoidectomy when possible placebo effects were precluded.

Methods A pragmatic, multicentre, double-blinded, randomised controlled trial. The study was conducted over 10 months in 2012–2013 at three ambulatory clinics. Two hundred and eighty-two children, age 1–11 years, American Society of Anesthesiologists grade ≤II, were included. To equalise expectancy effects, all parents were told that their child would receive acupuncture. However, children were randomly allocated to perioperative bilateral needling acupuncture at PC6, depth 7 mm, mean time 17 min (SD 5–45) during anaesthesia plus usual care, or to usual care only. The regional ethics committee approved this approach. Primary endpoints were nausea and vomiting 24 h postoperatively.

Results This study did not demonstrate any effect of acupuncture (95% CI) compared with standard care. The overall vomiting in the acupuncture and usual-care groups was 44.2% and 47.9%, respectively. Nausea was experienced by 31.7% in the acupuncture group and by 32.6% in the usual-care group. The test power was acceptable for comparisons of vomiting.

Conclusions The findings suggest that when controlling for possible placebo effects standardised PC6 acupuncture needling during anaesthesia without further stimulation of PC6 is not effective in reducing nausea and vomiting in children after tonsillectomy with or without adenoidectomy. Future studies should investigate acupuncture treatment which balances adequate dose and technique and a feasible, child-friendly acupuncture treatment.

Trial registration number NCT01729052.

Statistics from


Postoperative nausea and vomiting are common complications in children after tonsillectomy and adenoidectomy and remain challenges in our daily practice. Antiemetic drug treatment is not completely effective and can cause unwanted adverse events. A non-pharmacological treatment, such as acupuncture, might therefore be of interest.

Trials on the effect of acupuncture have shown conflicting results and the quality of the methods has been debated. There is also a wide variability in acupuncture applications, such as different point of time, techniques, duration and lack of appropriate placebo controls, making it difficult to evaluate the clinical importance.1 ,2 Meta-analyses indicate that acupuncture may reduce postoperative nausea and vomiting in adults and children.3–5

Opponents of acupuncture contend that acupuncture is not effective beyond a placebo effect.6 ,7 The placebo effect may be defined as “a beneficial effect in a patient following a particular treatment that arises from the patient's expectations concerning the treatment rather than from the treatment itself”.8 Colagiuri9 describes two forms of expectancy: expected efficacy and perceived treatment. The ‘expected efficacy’ is how well a treatment is expected to work. ‘Perceived treatment’ is related to randomised controlled trials constituting patients’ beliefs about their group allocation. The perceived treatment is a possible source of activating or deactivating expectancies about the effect of treatment. A systematic review showed a relationship between patients’ expectancies and acupuncture treatment outcomes.10

The use of sham acupuncture is controversial. It is inclined to have high placebo effects on subjective symptoms11 and may not even be inert, as it evokes peripheral and central neural effects.12

This study was designed to preclude possible placebo effects. First, we assumed that there was a relationship between parental treatment expectancies and treatment outcomes. Second, we assumed that parental beliefs about group allocation could activate or deactivate treatment expectancies. To avoid these biases, which tend to produce placebo effects, we told all parents that their child would receive acupuncture, whereas the children were allocated randomly to one of two groups, only one group receiving acupuncture. We discuss the ethical issues of this procedure in the ‘Ethical considerations’ section. Finally, in order to make sham acupuncture irrelevant, and thus eliminate its effects, we performed acupuncture during anaesthesia.

To our knowledge, this study is the first investigating the effect of acupuncture in children after controlling for the possible placebo effects of parental expectancies.

Material and methods

A pragmatic, multicentre, double-blinded, randomised controlled trial was carried out. Possible placebo effects—that is, effects of sham acupuncture and expectancies, were precluded. We tested the feasibility of the study in a pilot. The protocol remained unchanged after assessment of the pilot study. Trial registry number: NCT01729052.

Participants and settings

We conducted the study at three ambulatory clinics in Norway, specialising in ear, nose and throat surgery. The study was carried out alongside normal practice. Inclusion criteria were children 1–11 years of age, with American Society of Anesthesiologists grade ≤II, admitted for tonsillectomy with or without adenoidectomy. Exclusion criteria were parents requiring an interpreter, rash or infection over the relevant acupuncture points, emesis or antiemetic treatment during the previous 24 h and gastrointestinal illness. Six anaesthetists performed the acupuncture. One of them was a trained acupuncturist; the others were trained by a licensed acupuncturist. Five experienced surgeons conducted the operations.

Sample size

We calculated the sample size by a programme at the web pages of University of California, Department of Epidemiology and Biostatistics, San Francisco, USA (*Power). We based the calculation of sample size on the variable vomiting (yes/no). We expected the prevalence of vomiting to be about 50%. When comparing vomiting in the two study groups, a χ2 test with 5% significance level was used. It may then be shown that in order to have 80% test power, if the true difference in prevalence of vomiting between the groups is at least 20 percentage points,13 at least 198 patients must be included in the study. We expected a dropout rate of about 30% of the included patients and decided to include 282 patients in the study.

Pre-study procedures

The parents received study information in a letter sent beforehand, and on the day of surgery the principal researcher gave information verbally. The principal researcher then enrolled the children in the study if they were eligible according to the inclusion/exclusion criteria and the parents gave written, informed consent. However, if a child showed any resistance to participation by word or action, we respected that.

The children were fasting on the day of surgery, but allowed to drink clear fluids up to 2 h before surgery.


We used computer-generated randomisation with variable block sizes of between two and six. Within each block, a 1:1 ratio allocated children to receive acupuncture plus usual care or usual care only. The anaesthetist carried out the randomisation by drawing envelopes with consecutive numbers. Each new child enrolled in the study received the next consecutive envelope, which was opened after induction of anaesthesia. The information about the group assignment was not accessible to the researchers before the data collection was completed. The Clinical Research Centre, University Hospital of North Norway, prepared the randomisation list.

Anaesthetic management

The anaesthetists used different modes for the administration of anaesthetic agents: (1) induction and maintenance with sevoflurane, (2) induction with sevoflurane and maintenance with propofol and remifentanil, (3) induction and maintenance with propofol and remifentanil. The anaesthesia was induced and maintained with oxygen 30% and air or nitrous oxide and positive pressure ventilation.

A laryngeal mask or uncuffed orotracheal tube secured the airways. All children received Ringer's acetate (at a non-standardised rate), paracetamol (suppository 40 mg/kg or intravenously 15 mg/kg) and dexamethasone intravenously (0.25 mg/kg) or diclofenac suppository (25 mg). The surgeons used dissection suction or diathermy scissors for tonsillectomy.

Intervention and blinding

We used acupuncture point PC6, the point most often studied for reduction of nausea and vomiting. Its popularity may be due to its antiemetic properties and to the easy access on the forearm. After induction of anaesthesia, the anaesthetists needled, at an angle of approximately 30°, at PC6 (Neiguan) bilaterally with Seirin needles no 3 (0.20)×15 mm to a depth of approximately 7 mm and removed the needles before entry to the recovery unit. PC6 is located at the wrist between the tendons of the palmaris longus and flexor carpi radialis, proximal to the palmar crease. There was no manipulation of the acupuncture needles. The needle retention time was set at 15–20 min, depending on the surgery time.

The surgery drape covered the needles (if any) perioperatively. No matter whether the child had received acupuncture or not, the anaesthetists fixed adhesive tape on the acupuncture points after needle removal. Thus, we blinded children, parents and personnel in contact with parents and children postoperatively, such as surgeons, care providers and the investigator/outcome assessor.

Postoperative care

Postoperative pain in the recovery unit was treated with fentanyl intravenously. The use of antiemetics was restricted to persistent nausea and vomiting. Ringer's acetate was continued. Two of the clinics allowed oral intake at 2 h postoperatively and one of the clinics allowed clear fluids on demand. The stay in recovery was for a minimum of 1 h.

Data collection

We collected perioperative information from the anaesthesia and postoperative record. The principal investigator briefed the parents in the assessment of data. The parents assessed retching and vomiting by frequency using a purpose-designed form. Recurrence of retching and vomiting within a period of 2 min was considered as one occurrence. The parents employed a behaviour tool, FLACC-N (Face, Legs, Activity, Cry and Consolability), to measure pain in children <5 years of age and they assessed pain and nausea in children >5 years of age using the Faces Pain Scale and the BARF (Baxter Animated Retching Face) nausea scale. Of interest were the highest scores during 24 h postoperatively. They also reported their evaluation of the children's experience of overall malaise as none, minimal, moderate, great and severe. The principal researcher collected the recorded data by telephone.

Validity and reliability

The Norwegian version of the pain assessment tool FLACC-N had a high inter-rater reliability (0.95) and internal consistency reliability (Cronbach's α 0.96). The correlations between FLACC-N and the Numeric Rating Scale indicated a moderate consistency (0.52).14 The Faces Pain Scale had a strong positive correlation (r=0.93) with a visual analogue scale.15 Likewise, the BARF had a strong correlation (Spearman r=0.93) with the Visual Analogue Scale for nausea.16

Outcome measures

Main outcome measures were nausea and vomiting during 24 h postoperatively. Secondary outcome measures were pain, the use of postoperative analgesics and children's overall experience of malaise.

Ethical considerations

Written informed consent was obtained from all parents. The trial was performed in accordance with the Helsinki declaration.

All parents were told that their child would receive acupuncture, whereas the children were randomised to acupuncture or no acupuncture. This is a major ethical consideration and we had several discussions with the regional ethics committee (REK-nord) about this procedure before they recommended the study. Deception is typically used in psychology and several official bodies have made statements on this issue. Our considerations were in accordance with the National Patient Safety Agency17 and California State University18 stating that deception is justified when (1) it is not possible to obtain useful knowledge without withholding some aspects of the study protocol, (2) the findings may produce knowledge of scientific value, (3) use of deception must not expose the participants to more than minimal psychological or physiological risks, (4) the participants’ willingness to consent is not affected. We informed the parents in the study information letter sent beforehand that some features in the protocol would not be disclosed until the study was concluded.

Children may be afraid of the acupuncture needles; consequently, we performed the acupuncture during anaesthesia.

Data analysis

Retching and vomiting were collapsed into a single variable: vomiting. The continuous variables nausea and pain were converted to dichotomous variables: yes/no. When comparing percentages, the χ2 test was used. A significance level of 5% was applied throughout. Frequencies are presented as percentage with 95% CI. The data were analysed by SPSS V.21.0.


From November 2012 to June 2013, we included 138 children in the acupuncture group and 144 children in the control group. The CONSORT flow diagram of the study is displayed in figure 1.

Figure 1

The CONSORT flow diagram of the study.

Patients’ characteristics, type and duration of surgery/needle retention time are presented in table 1. Types of anaesthesia, use of laryngeal mask or orotracheal tube, surgery techniques and administration of intravenous fluids were evenly distributed in the two groups. One child was readmitted to hospital because of bleeding.

Table 1

Baseline characteristics and perioperative data

This study did not show any significant differences in overall vomiting (0–24 h) or in early (0–6 h) and late (7–24 h) vomiting between the acupuncture and usual-care group. Likewise, the study showed no effect of acupuncture on nausea (in children >5 years). The overall percentages of vomiting in the acupuncture and usual-care groups were 44.2% and 47.9%, respectively. Nausea in children >5 years was experienced by 31.7% in the acupuncture group and by 32.6% in the usual-care group (table 2). Subgroup analysis showed no significant association between gender and vomiting (acupuncture group p=0.441, usual-care group p=0.166) or gender and nausea (acupuncture group p=0.658, usual-care group p=0.901). The range of the 95% CIs for the frequency differences were 23.2% for overall vomiting and 39.4% for nausea.

Table 2

Primary outcomes: vomiting and nausea

The absence of differences between the two groups remained when the anaesthetist who was a trained acupuncturist, performed the acupuncture (p values for vomiting and nausea: 0.679 and 0.348, respectively). There was no difference between the two groups in the children's experience of pain and malaise, or the use of analgesics postoperatively (table 3).

Table 3

Secondary outcomes: pain, use of analgesics postoperatively and children's overall malaise

The parents reported no discomfort or dermatological problems at the acupuncture site. However, they observed minor haematomas in 15 children.


This study did not show any significant difference between needling at PC6 during anaesthesia plus usual care versus usual care only, for postoperative vomiting and nausea after tonsillectomy with or without adenoidectomy in children. The range of the CI for difference in vomiting between the groups was relatively small, showing that the test power of the study was acceptable for comparisons of vomiting.19 There was no difference between the groups in the children's experience of pain and malaise, or the use of analgesics postoperatively.

It has been argued that the possibility of obtaining significant results when comparing two groups is increased when the efficacy variables are measured on a continuous or discrete scale, as compared with a dichotomised scale. The effect variables measuring vomiting, nausea and pain were originally measured on discrete and continuous scales. We have compared and analysed these original variables using a Mann–Whitney U test and the results from these analyses were far from significant.

The lack of evidence of antiemetic effect in this trial is in agreement with a similar double-blind trial by Yentis and Bissonnette20 who concluded that P6 acupuncture during anaesthesia is ineffective compared with non-acupuncture in reducing vomiting after tonsillectomy in children. Shenkman et al21 considered that acupuncture might require an intact nervous system, which again requires the patient to be awake. Accordingly, he conducted a double-blind trial including acupuncture at PC6 during anaesthesia and acupressure preoperatively compared with sham in children scheduled for tonsillectomy. Nevertheless, the combination of acupressure and acupuncture perioperatively was ineffective for emesis in comparison with sham acupuncture.

On the contrary, Somri et al22 concluded in their double-blinded trial that there was a significant difference in postoperative nausea and vomiting (PONV) after dental surgery in children who received acupuncture compared with sham acupuncture during anaesthesia. However, Somri used PC6 and CV13. A combination of points seems to be more effective than with single point.23 Another trial on children scheduled for tonsillectomy and adenoidectomy has demonstrated the effect of acupuncture during anaesthesia and acupressure postoperatively.13 However, in that trial, the control was not sham but usual care and the parents and investigators were not blinded.

Gan et al24 found that transcutaneous electrostimulation at PC6 before and during surgery, in addition to preventing PONV, also appeared to provide analgesia in adults. We have found little support suggesting that the present acupuncture modality is less effective than other modalities. Lee and Fan4 have suggested in their systematic review that invasive (eg, acupuncture, electrical stimulation) and non-invasive (eg, acupressure, transcutaneous electrical stimulation) procedures had similar effects.

Studies comparing acupuncture with sham often fail to demonstrate any differences of effect. It has been suggested that effects after both real and sham acupuncture result from a variety of active components and that sham acupuncture actually acts as an active treatment.11 We avoided this bias by using usual care as control, yet we found no difference between the two groups.

The strength of this study is that the methods allowed us to preclude possible placebo effects. Other strengths are the large sample size, the large follow-up rate and successful blinding; only one parent questioned whether the child had received acupuncture.

This study has some limitations, reducing the internal validity. The variability of anaesthetic management and other medications may lead to different results; volatile substances and opioids may contribute to, and propofol may reduce, emesis. Further, a laryngeal mask and orotracheal tube affect the experience of malaise and vomiting differently. The different postoperative care strategies for the oral intake of fluids may also affect the outcomes. Nevertheless, one of the intentions of a pragmatic trial is to try to answer the question Can this intervention be useful in our clinical practice? One strategy for answering this is to maximise external validity by including patients from as many sources as possible (eg, clinics) and attain flexible interventions and comparison.25 We might have taken these factors into consideration when estimating the sample size; a larger study would have allowed for greater variation between centres. The sample size calculation was based on a difference of 20 percentage points based on results of a previous study.13 In this study, however, we might expect a smaller effect as there was control for the expectancy effect. Recognising these considerations, the study may be underpowered.

The acupuncture procedure was standardised. The only stimulation was at insertion of the needles during anaesthesia. Performing acupuncture during anaesthesia precluded the needle sensation (de qi) and might cause problems in achieving the accuracy of identifying the acupuncture point PC. One may speculate whether this might have produced negligible neurophysiological effect and thus minimal likelihood of benefit. The evidence of acupuncture point specificity is conflicting. A review of studies on acupuncture points and sham points or non-specific acupuncture points confirmed the existence of acupuncture point specificity.26 In contrast, a systematic review of the randomised controlled trials with sham acupuncture controls did not find any acupuncture point specificity.27 Moreover, Somri et al22 found a statistically significant antiemetic effect of standardised acupuncture at two acupuncture points without stimulation during anaesthesia. In our study, the duration of the acupuncture treatment varied according to surgery time. Nevertheless, relatively short needle retention times (1, 5 and 15 min, respectively) have been shown to be effective in children.22 ,28 ,29

Female gender is considered a risk factor for PONV in adults,30 but not in children.31 However, a previous study found that girls were more prone to vomiting than were boys.13 In our study, there was an imbalance of girls/boys in the two groups, as a preponderance of girls received acupuncture. There was no significant association between gender and nausea or vomiting, so gender was not a confounding factor.

Thirty-three of 340 parents declined to participate and nine parents would have needed an interpreter. These numbers are small, suggesting negligible selection bias. Patients often wish to make things turn out well, a notion of ‘eager-to-please’, which may cause information bias. This bias was ruled out in the two groups, as all parents believed that their child received acupuncture treatment. The telephone interview is a possible source of bias, as the principal researcher might subconsciously influence the parents’ responses. In order to minimise this possible flaw, we used a structured interview.

The aim of this pragmatic trial was to investigate the effect of feasible acupuncture treatment in a busy everyday practice. We consider that feasibility involves a standardised acupuncture procedure that does not demand any extra time and resources and is easy to learn and perform by the existing department personnel. Last, but not least, it must be acceptable to the child, and needling while the child was awake was thus not an option. The acupuncture dose in this study might be considered inadequate, biasing the results. Feasibility and child-friendliness should not equate to inadequate acupuncture. To remedy this and optimise a possible acustimulation effect, we have previously performed an open pragmatic study designed using acupuncture during anaesthesia and wristbands postoperatively,13 and found that vomiting was significantly less pronounced in children receiving acupuncture than in children receiving usual care. One may speculate whether these results were due to the additional use of wristbands, or to the open design, allowing the influence of parental expectancy be part of the effect, or due to both. We suggest that future studies should investigate acupuncture treatment balancing adequate dose and technique and a feasible, child-friendly acupuncture treatment. Use of several stimulation points, stimulation of needles and additional acupressure wristbands are characteristics that might increase the effect of acupuncture and are interesting areas for future investigation.


The findings suggest that, when controlling for possible placebo effects, standardised acupuncture needling at PC6 during anaesthesia without further stimulation is not effective in reducing nausea and vomiting in children after tonsillectomy with or without adenoidectomy. Future studies should investigate acupuncture treatment balancing adequate dose and technique and a feasible, child-friendly acupuncture treatment.

Summary points

  • Acupuncture needling at PC6 was compared with standard care for postoperative nausea and vomiting in children after tonsillectomy with or without adenoidectomy.

  • Standardised PC6 acupuncture needling during anaesthesia without stimulation was not effective when controlling for possible placebo effects.

  • Future studies should investigate acupuncture treatment balancing adequate dose and technique and a feasible, child-friendly acupuncture treatment.


The authors thank the staff at the outpatient surgery departments and the anaesthetists for performing the acupuncture. A special thanks to the children and their parents who participated in the study. We collected the data at Nydalen Øre- Nese- Hals, Asker Øre- Nese- Halsklinikk AS and Polikliniske Operasjoner AS.


View Abstract


  • Contributors IL had the main responsibility for the conception and design of the work and the acquisition; she contributed to the data analysis and interpretation and wrote the paper. LS had the main responsibility for the data analyses and interpretation and contributed to the writing. BTV contributed to the design and writing of the paper. EB contributed to the design and data analysis. AJN contributed to the conception and design, analysis, interpretation and the writing of the paper. All contributors revised the work, approved the final version and agreed to be accountable for all aspects of the work.

  • Competing interests None declared.

  • Patient consent Obtained.

  • Ethics approval Regional ethics committee (2012/63/REK nord).

  • 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.