Article Text

other Versions


Efficacy and safety of auriculopressure for primary care patients with chronic non-specific spinal pain: a multicentre randomised controlled trial
  1. Jorge Vas1,2,
  2. Manuela Modesto1,
  3. Inmaculada Aguilar1,
  4. Camila da Silva Gonçalo3,
  5. Francisco Rivas-Ruiz2,4
  1. 1Pain Treatment Unit, Doña Mercedes Primary Health Care Centre, Dos Hermanas, Spain
  2. 2Carlos III Health Institute, Network of Research in Health Services in Chronic Diseases (REDISSEC), Madrid, Spain
  3. 3Department of Community Health, Jundiaí Medical School, Jundiaí, São Paulo, Brazil
  4. 4Support Research Unit, Costa del Sol Hospital, Marbella, Spain
  1. Correspondence to Dr Jorge Vas, Pain Treatment Unit, Doña Mercedes Primary Health Care Centre, Dos Hermanas 41700, Spain; jorgef.vas.sspa{at}


Background Chronic uncomplicated musculoskeletal pain in the spine (cervical, thoracic or lumbar) is highly prevalent and may severely limit the daily activities of those affected by it.

Objective To evaluate the efficacy and safety of treatment with auriculopressure applied to patients with non-specific spinal pain.

Methods A multicentre randomised controlled trial with two parallel arms (true auriculopressure (TAP) and placebo auriculopressure (PAP)) was performed. The intervention phase lasted 8 weeks and outcomes were measured 1 week after the last intervention (T1) and 6 months after baseline (T2). The primary outcome measure was change in pain intensity according to a 100 mm visual analogue scale (pain VAS) at T1. Secondary outcome measures were the Lattinen index, the McGill Pain Questionnaire and the SF-12 health-related quality of life scale (Spanish version in every case).

Results There were 265 participants (TAP group, n=130; PAP group, n=135). Pain was most frequently located in the upper back (55.1%, n=146), followed by the lower back (25.3%, n=67) and the dorsal area (12.5%, n=33). Nineteen patients (7.2%) reported pain affecting the entire spine. There were statistically significant differences between TAP and PAP in the change in the pain VAS at T1 of 10 mm (95% CI 2.8 to 17.3, p=0.007) and in the change in the pain VAS at T2 of 7.2 mm (95% CI 0.02 to 14.3, p=0.049) in favour of TAP. We also observed a statistically significant difference of 3.4 points in the physical component of the SF-12 in favour of TAP at T2 (95% CI 0.45 to 6.3, p=0.024). No severe adverse effects were detected or reported during treatment.

Conclusions The application of auriculopressure in patients with non-specific spinal pain in primary healthcare is effective and safe, and therefore should be considered for inclusion in the portfolio of primary healthcare services.

Trial registration number ISRCTN01897462.

Statistics from


Non-specific spinal pain (NSP), comprising back and/or neck pain, is responsible for 54–80% of consultations for chronic pain.1 Of the cases of spinal pain treated in primary care, 5% are due to a structural alteration of the spinal column while the remaining 95% are non-specific.2 The lifetime prevalence is 50–70% in Spain.3 Lumbar pain has the highest prevalence, being experienced by an estimated 60–70% of adults at some time.4 Moreover, it tends to be persistent or recurrent.5

Acupuncture and related techniques such as auriculopressure (the application of auricular implants with vaccaria seeds at specific points of the ear) have been used as therapeutic interventions in China for over 2000 years and are becoming ever more widely accepted and applied in the West, especially for pathologies causing pain.6 Stimulation of the pinna is often used, in association with somatic acupuncture, as a technique to treat pain7 as well as sleep disorders, anxiety8 ,9 or disorders of the autonomic nervous system.10 It can also be used alone to treat various musculoskeletal diseases,11–17 with an action probably based on an endorphinergic mechanism.18 However, the relationship between the somatic locations of musculoskeletal pain and the sensitive points in the auricular pavilion remains unclear.19

The multicentre randomised controlled study described in this paper was designed to investigate the efficacy and safety of auricular implants as a sensory stimulus in patients with chronic NSP treated by primary healthcare services.


In this multicentre randomised controlled study with two parallel arms, we compared true auriculopressure (TAP) (applied in accordance with standard practice and individual diagnosis) and placebo auriculopressure (PAP) in the treatment of chronic NSP. The patients were blinded to both interventions and the results were analysed by investigators blinded to the composition of the study groups.

Patients were recruited by physicians at 10 primary healthcare centres belonging to the Public Health System of Andalusia (Southern Health Management Area, Seville, Spain) and were referred to the 15 physicians and nurses who participated in the study.

The patients were at least 18 years old and presented with uncomplicated musculoskeletal spinal pain (cervical, thoracic or lumbar) diagnosed by clinical history and physical examination and had not received prior treatment with auricular implants. The following exclusion criteria were applied: protrusion or prolapse of one or more intervertebral discs with concurrent neurological symptoms; infectious spondylopathy; previous spinal surgery; back pain caused by inflammatory, malignant or autoimmune disease; congenital spinal deformities except for slight degrees of scoliosis or lordosis; vertebral fractures; spinal stenosis; spondylolysis or spondylolisthesis; skin diseases affecting the auricle; allergies to tape; pregnancy; labour litigation by reason of spinal pain; inability to complete the questionnaire or to answer the evaluator's questions.

The sample size was calculated in accordance with the findings of a pilot experiment conducted in the Pain Treatment Unit at the Doña Mercedes Health Centre (Dos Hermanas, Spain) during the period April to December 2006. In this experiment the mean reduction in pain intensity according to a 100 mm visual analogue scale (VAS) from baseline to final outcome was 32.3 mm in the experimental group and 19.8 mm in the control group (SD 31.9 mm). Assuming a 90% power, an α value of 5% and a dropout rate of 30%, a total sample of 400 participants, divided into two equal groups, would be needed to detect this change.

Sampling was performed by consecutive selection in accordance with the above-described inclusion and exclusion criteria over a period of 12 months. The randomised allocation to the two study arms was performed using EpiDat V.3.1 software at the Research Unit of the Costa del Sol Hospital, Málaga following a 1:1 TAP/PAP scheme in blocks of six patients, stratified by physician. Neither the hospitals nor the personnel who participated in the study were involved in the randomisation process, and the sequence of randomisation remained concealed until the conclusion of the study. The study design has been described in detail elsewhere.20

The 15 doctors and nurses who took part in the study received a training course in auriculopressure techniques for the treatment of chronic NSP, with a total of 15 h of theory and 15 h of practical instruction. The personnel at 4 of the 10 participating centres had previously received specific training in acupuncture and related techniques, with an average of 320 classroom hours and over 100 h of clinical practice. All patients were allowed to use medication for symptomatic pain relief if necessary.

The patients in both groups attended on eight occasions (once weekly) for assessment and for the auricular implants to be inserted.

TAP consisted of the application of auricular implants in the form of vaccaria seeds (Vaccaria segetalis Garcke, known in China as Wang buliuxing) as an individualised sensory stimulus attached to the ear pinna by means of flesh-coloured tape. The auricular points were selected in accordance with Chinese maps, and Shenmen (in the triangular fossa) and Thalamus (on the inner wall of the antitragus) were selected as obligatory points, to which other points were added according to the sensitivity of the auricular areas, as determined by a 250 g pressure detector,21 ,22 to a maximum of 10 points per session, in one ear, starting with the most sensitive ear and then alternating weekly. The patients were instructed to exert finger pressure, repeated 10 times, on each ear implant, three times daily. These implants were retained for 7 days until the next consultation, when they were removed by the doctor or nurse.

PAP was performed following the same protocol and under the same conditions as TAP but using adhesive tape to attach inactive black plastic discs, 1.5 mm in diameter and 35 μm thick, simulating the appearance of the auricular implants used in TAP.

Special care was taken to ensure that the time spent with the patients assigned to each of the groups was identical, as was that employed in the pre- and post-session evaluations.

The primary clinical outcome was the change in pain intensity according to the 100 mm VAS (pain VAS)23 at 9 weeks after starting treatment. Secondary clinical outcomes were pain VAS change at 6 months after starting treatment, and changes recorded after 9 weeks and after 6 months on the SF-12 health-related quality of life scale24 and in the Lattinen index (LI), a multidimensional questionnaire that is widely used in Spain for pain assessment. This measure consists of the following pain-related variables: pain intensity, pain frequency, consumption of analgesics, level of activity and nocturnal rest. Each variable was measured on a 5-point Likert scale rated from 0 to 4. The sum of the values of the five variables thus ranged from 0 to 20 points, with a higher score indicating greater severity,25 ,26 and greater consumption of analgesics and non-steroidal anti-inflammatory drugs (whether or not prescribed by the physician). At 3 weeks after starting treatment, treatment expectations and credibility (TEC) were assessed in accordance with the original scale of Borkovec and Nau, with the following four items scored on a numerical scale of 0–10 (0=totally disagree, 10=totally agree):

  • TEC1: Are you confident that this treatment will relieve the pain you feel?

  • TEC2: Does the treatment seem to be a logical one?

  • TEC3: Would you recommend this treatment to a friend or relative suffering the same complaint?

  • TEC4: Do you believe this treatment would be a possible option for dealing with other problems?27

After the conclusion of treatment, patient satisfaction was further assessed in terms of patients’ perceived improvement28 using a 7-point Likert scale and with blinding to treatment options.29

Each patient completed a self-administered questionnaire before starting treatment (T0), 1 week after the end of treatment (T1) and after 6 months (T2). This questionnaire recorded one-dimensional data (pain VAS) and multidimensional data (McGill Pain Questionnaire (MPQ), Spanish version30) on spinal pain and on the present consumption of medication, and also compiled information on disease-related quality of life (SF-12) and days off work due to back pain. In addition, before treatment began, the patients were asked to complete a form providing demographic data. Any adverse effects of the techniques under study and details of any rescue medication taken were also recorded. During the baseline assessment, background information on the characteristics and duration of the back pain was also obtained. In each treatment session the therapist recorded the five variables related to the LI and the auricular points selected for each patient.

A descriptive analysis was performed, including measures of central tendency and dispersion for the quantitative variables and frequency distributions for the qualitative variables. A bivariate analysis was conducted to assess (1) differences in sociodemographic and clinical characteristics and baseline-to-outcome variables between the treatment groups; (2) efficacy, in terms of diverse variables; (3) expectations and credibility, and verification of effective blinding to the treatment options. The χ2 test, with continuity correction for the 2×2 tables, was used to assess the qualitative variables and the Student t test (after verification of normal distribution) was used to evaluate the quantitative variables. Multivariate linear regression was performed, taking as the outcome variable the absolute improvement recorded on the pain VAS at T1 by intention to treat (ITT) including, by the forward stepwise method, all sociodemographic and clinical variables previously addressed. The level of statistical significance was set at p<0.05 and 95% CIs were determined for the differences in the means and β coefficient (β) of regression. The analysis was performed by a statistician who was blinded to the study groups. SPSS V.17 was used for all analyses.


Between June 2008 and February 2010, 358 patients were evaluated, of whom 79 were excluded for various reasons (figure 1). Thus, the initial study population was 279 patients, of whom 14 (5%) were lost at baseline for scheduling reasons (9 in the TAP group and 5 in the PAP group). The sensitivity analysis showed no differences in baseline variables between the subjects lost to the study and the 265 ultimately included.

The study population therefore comprised 265 participants (130 in the TAP group and 135 in the PAP group) with a mean (SD) age of 51.1 (15) years. The majority of the study population were women (n=221, 83.4%). The mean (SD) body mass index (BMI) was 28.3 (5.2); 50.2% (n=133) of the study population had a primary level of education and 41.9% were in employment. The largest proportion of spinal pain (55.1%) was located in the neck, followed by the lower back (25.3%), the thoracic region (12.5%) and the whole spine (7.2%). In 119 patients (44.9%) there was pain in more than one location and 33 (12.5%) reported radiating pain. The patients reported a mean (SD) duration of this pain of 9.2 (8.4) years (table 1).

Table 1

Descriptive analysis of baseline clinical and sociodemographic variables for the two treatment groups

At baseline the mean (SD) pain VAS score was 69.4 (15.8), the mean (SD) physical component of the SF-12 scale was 35.5 (9.5) and the mean (SD) mental component was 43.3 (11.5). Of the items included in the LI, at baseline 47.2% of the subjects frequently suffered pain, 13.2% reported a high or excessive consumption of analgesics, with respect to the level of activity 13.6% of the sample needed help or were totally dependent and 15.8% reported insomnia or the need for sedatives in order to sleep comfortably. The mean (SD) duration of each intervention session was 11.2 (3.5) min and the mean (SD) number of implants per session was 6 (2.1). There were no significant differences in any of the sociodemographic, clinical and outcome variables assessed at baseline between the TAP and PAP groups (table 1).

For the primary outcome variable (ie, the absolute improvement in pain VAS at T1 according to analysis by ITT), we found a mean difference of 10.0 mm between the TAP and PAP groups (95% CI 2.8 to 17.3; p=0.007). This significant difference in favour of TAP persisted at the follow-up examination (T2, ITT), with a mean difference of 7.2 mm (95% CI 0.02 to 14.3; p=0.049) (table 2). There were also significant differences in the change at T2 in the physical component of the SF-12 in favour of TAP, with a difference in the means of 3.4 points (95% CI 0.45 to 6.3; p=0.024) (table 2).

Table 2

Analysis of effectiveness with quantitative primary and secondary outcome variables

With regard to the changes in the LI between T0 and T1, the TAP group showed a greater improvement in all items and the difference was significant with respect to quality of sleep (72.1% in the TAP group vs 60.7% in the PAP group, p=0.027) and patients’ perceived improvement (76.4% in the TAP group vs 57.8% in the PAP group, p=0.002) (table 3).

Table 3

Analysis of effectiveness with Lattinen index and patients’ satisfaction with treatment

In the multivariate linear regression model in which the dependent variable was absolute improvement in pain at T1 by ITT, apart from the treatment group the only variable to be included in the model was BMI. When this variable was held constant, the TAP group had a mean absolute pain improvement score of 10.4 (95% CI 3.3 to 17.6) (table 4).

Table 4

Multivariate model with the dependent variable ‘absolute improvement in pain VAS’ at T1 by ITT

No major adverse effects were reported. In 18 patients (12 in the TAP group and 6 in the PAP group) the implants caused pressure ulcers in the pinna, which healed within 10 days of removal, and eight patients (5 in the TAP group and 3 in the PAP group) reported worsened symptoms. The data on patients’ TEC are shown in table 5. Finally, in the verification of the blinding process, over 70% of the patients in each group thought they were being treated with TAP, and there were no significant differences between the groups (table 5).

Table 5

Assessment of expectations and treatment credibility (ECT) and verification of blinding to treatment option by groups


Primary healthcare requires effective technologies at low cost that are well-accepted by users. Few studies have been made of chronic pain and the use of auriculopressure in primary healthcare services. To the best of our knowledge, the present study is the first to compare true and sham (placebo) auriculopressure. Our main aim was to investigate the efficacy and safety of auricular implants as a means of sensory stimulation in patients with chronic NSP treated in primary healthcare services. The results of our study show that auriculopressure is an effective, safe treatment for reducing pain in the short and medium term in patients with chronic NSP. As well as alleviating pain, this treatment is associated with improved sleep quality in the short term, and an enhanced quality of life with respect to the physical component in the medium term. Equally important is the high level of satisfaction with the auriculopressure process reported by patients. These results are comparable with the findings of other studies.31 ,32

The percentage of women with chronic pain in our sample is consistent with the data reported by other authors,33–35 particularly among users of public healthcare services.36

The association between BMI and the results obtained corroborates the findings of other studies,37 ,38 according to which there is a predisposition among subjects with a high BMI to suffer chronic spinal pain. Such patients should therefore be carefully examined to identify potential inflammatory factors associated with obesity that may promote musculoskeletal pain in the spine.39

Regarding the physical aspect of the patients’ quality of life (PCS-SF12), the improvement achieved in the TAP group was significantly greater than in the PAP group. With regard to the mental component (MCS-SF12), the two groups obtained similar changes. In this context, the question of sham acupuncture and the placebo effect has been widely discussed elsewhere.40 ,41 However, we emphasise that there is a lack of information regarding the extrapolation of this discussion to the area of auricular acupuncture or auriculopressure.

Previous studies of auriculopressure have used control groups with sham auriculopressure applied at non-specific points. We decided to use an inert device at the same active points as in the TAP group and, although the application of manual pressure on the sham elements could have contributed to the improvement obtained by the patients in the PAP group, thus producing some degree of observation bias prejudicing our assessment of the effectiveness of TAP, the approach adopted is considered to be of net benefit to the validity of the results obtained.

We believe the decision to use PAP helped maintain the blinding of the patients, since most of the participants in both groups believed they had been treated with TAP; this circumstance could have been significant to the fact that there were virtually no dropouts during the treatment phase. This finding highlights one of the characteristics of the reliability of this study, to the extent that the non-awareness of group allocation prevented the patients’ behaviour, judgements and responses to the assessment measures from being affected by subjective questions.42 With regard to the sample size originally calculated, a high dropout rate was assumed in order to ensure a sufficient sample. The final sample size obtained does not influence the quality of the values calculated or the power of the hypothesis tests. Fortunately, the conservative dropout rate initially estimated (30%) was not matched by the real-life figure (5%), so the risk of a non-response bias was avoided. It should be noted that the patients included in the pilot study were different from those in the final study presented.

Although one of the measurement instruments selected for evaluating the results obtained was the Spanish version of the MPQ, the fact that most patients failed to complete it properly obliged us to omit this measure from the results report. The MPQ pain rating index consists of 64 pain descriptors broken down into 19 subclasses and, within each class, the descriptors are classified by pain intensity. We believe the reason why this questionnaire was not correctly answered is related to the low educational level of the study population (71.5% of the participants had only completed primary education), taking into account the considerable complexity of the procedure.

NSP had a relatively low impact on sick leave (being cited as a cause by 7.2% of the study population). Our findings differ from others that have reported a higher prevalence of inability to work as the result of chronic NSP43 and the major impact of non-specific musculoskeletal pain as a cause of absenteeism from the workplace.31 ,44 This difference could be accounted for by the moderate degree of pain intensity experienced by the patients in our study population (69.4 mm). This conclusion is corroborated by the results of a European survey in which the patients with the highest levels of labour absenteeism were those who experienced the most severe pain.45

One possible limitation of our study concerns the mechanism by which a non-invasive technique such as auriculopressure is effective. The act of placing auricular implants (whether seeds or flat plastic discs) may trigger the stimulation of the sensitive points when pressure is exerted, albeit at different intensities. Another limitation is the fact that the frequency of auricular implants and the intensity of pressure exerted by the patient at the points in question were not controlled. Although the participants were trained by healthcare personnel as to how the stimulus should be applied to the implants, there was no way to ensure that all these recommendations were followed precisely.

The fact that non-specific pathologies of the neck, back and low back have been grouped under the term ‘spinal pain’ might represent a limitation of this study because, according to some authors, there are differences in the biomechanics and pathology of the different parts of the spinal column.46

Patients’ expectations and the credibility of the technique were higher among the patients in the TAP group, probably because this parameter was assessed during the third treatment session when there were evident differences between the results of the two groups in favour of TAP.

While acknowledging the limitations inherent in this study, we believe that, given the scant previous research into the question of auriculopressure in primary healthcare and related chronic health problems, the present work contributes new understanding in this area.

Finally, we suggest that future studies should be developed to address the efficacy and safety of auriculopressure with respect to other public health problems, particularly those which generate the greatest demand for healthcare.


We conclude that auriculopressure can achieve pain reduction in the short and medium term while improving the quality of life of patients with chronic NSP. The technique is safe and simple and therefore should be considered as a valuable option in primary healthcare for these patients.

Summary points

  • Auriculoacupressure for spinal pain has rarely been tested.

  • We compared 8 weeks of auriculopressure with placebo acupressure for pain.

  • Auriculoacupressure was significantly more effective in reducing pain and had few adverse effects.

  • Treatment with auricular implants should be considered as a treatment option in primary healthcare for patients with chronic non-specific spinal pain.


View Abstract


  • Contributors JV conceived the study, designed the study protocol, sought funding and ethical approval and wrote the manuscript. MM, IA, CdSG and FR-R made a substantial contribution to designing the treatment protocol. MM, IA and CdSG participated in data collection. JV and FRR analysed and interpreted the data. JV drafted the report. MM, IA, CdSG and FRR critically reviewed the report. All authors approved the final version of the manuscript. The corresponding author had final responsibility for the decision to submit for publication.

  • Funding The study was funded by the Spanish Ministry of Health and Consumer Affairs (Carlos III Health Institute, grant ISCII 07/90058) and by the Andalusian Public Health System (grant PI053/2007). The funder played no part in the design or performance of the study, in data analysis or interpretation or in the decision to publish.

  • Disclaimer The funder played no part in the design or performance of the study, in data analysis or interpretation or in the decision to publish.

  • Ethics approval The study protocol was approved by the Clinical Trials Ethics Committee of the Andalusian Regional Government (Decision 10/07)as well as by the corresponding Research Committee at each of the participating centres.

  • 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.