group received the application of the Eye point (LO5)(10). This point, located in the center of the ear lobe(10), is distant from the points applied in the treatment group and has no relationship with the focus of observation of the present study. Both groups attended five AA sessions, which occurred once a week for one month and a half, with alternation of ears in subsequent sessions. The whole procedure was carried out by a professional specialized in acupuncture, with more than three years of experience in the area. The people allocated in the control group received no orientation and were submitted to no intervention during the evaluation period. The participants were evaluated before the first AA session (initial evaluation), one week after the fifth session (final evaluation), and after a 15-day follow-up 3 Moura CC, Chaves ECL, Chianca TCM, Ruginsk SG, Nogueira DA, Iunes DH www.ee.usp.br/reeusp Rev Esc Enferm USP · 2019;53:e03418 (follow-up evaluation) by the same trained examiners. The following variables were examined to determine the sociodemographic and pain profiles of the sample: age, gender, pain duration and type (constant or recurrent), impact of pain on daily living activities (eating, dressing, and personal hygiene care), mood swings due to pain (to depressive or anxious), emotional changes (fear of movement), physical stamina (tiredness), impossibility to practice physical activities as a consequence of pain, and main causes of pain. Pain intensity, considered a primary outcome, was assessed using the Brief Pain Inventory (BPI)(11). The impact of pain on daily activities(11), the relief provided by the intervention(11), and the pain threshold as measured using digital algometry were classified as secondary outcomes, were. The BPI(11), translated into Portuguese, adapted to the Brazilian culture(12) and with proper psychometric characteristics(13), allows to evaluate pain severity and the level of its impact on common dimensions related to feelings and organic functions using numerical scales ranging from 0 (no pain/does not interfere) to 10 (pain as bad as you can imagine/completely interferes). In addition, the instrument includes a question about the use of medications or non-pharmacological methods for pain relief and the percentage of relief(13). The result is in the form of two final scores related to the mean of the four items that address pain intensity (subscale 1) and the mean of the seven items that focus on the impact of pain on daily activities (subscale 2)(13). In the sample of the present study, Cronbach’s alphas were 0.907 and 0.934 for subscales 1 and 2, respectively, which reveals a significant internal consistency of the instrument. A digital algometer (Kratos®) was used to measure the pain threshold when a mechanical stimulus was applied. During the procedure, patients lay on a stretcher in the prone position. The evaluation followed a standard for points of the cervical (insertion of suboccipital muscles: in the lower trapezius, at the height of the fifth and sixth cervical vertebrae), thoracic (midpoint of the lower trapezius muscle, between the acromion and the seventh cervical vertebra: lower trapezius at the inferior angle of the scapula and at the height of the eighth thoracic vertebra), and lumbar (posterior superior iliac spine: paravertebral muscle, at the height of the fourth and fifth lumbar vertebra and gluteal muscle at the eminence of the sciatic nerve) regions(14). The compression pressure was increased gradually at a speed of 1 kilogram per second. When the sensation of pressure caused by the algometer turned into pain, participants pressed a button, and the examiner ended the stimulus immediately. The mean of the measured values for the dorsal region was calculated for data analysis. Pretest and sample size calculation A pretest was carried out with 15 people(15) to evaluate the feasibility of the intervention proposal, verify the pertinence of the evaluation tools, and estimate the sample size. The intervention protocol was considered adequate by the acupuncturists who evaluated it and suffered no alterations during the pretest. Changes were performed in the evaluation instrument designed to obtain the socioeconomic and pain profiles and in the assessment of the pain threshold after a mechanical stimulus using a digital algometer. Additionally, the application of research instruments was standardized in the form of an interview with the participants. Sample size calculation was run with the programs GPower version 3.1 and BioEstat version 5.0. The statistical power, average effect size, and level of significance were set at 90%, 0.5, and 5%, respectively. The average pain intensity in the previous 24 hours was defined as the main variable, and the result of the calculation was 30 people per group. To prevent sample losses, this calculation was adjusted in 30%(16). Consequently, 110 people who met the inclusion criteria made up the study sample, and 83 completed the study steps, which represents a loss of 27 (24.54%) people (Figure 1). The selected patients were randomly allocated into three parallel study branches: treatment group (n=37), placebo group (n=36), and control group (n=37). Randomization was carried out in four blocks, with approximately 27 people in each block, by a researcher who did not participate in the study as an author using the