Approved by the Meiji University of Integrative Medicine Ethics Committee this study included healthy student volunteers from the acupuncture school of the University (n = 30, aged 18–22 years). Volunteers were informed of the experimental procedures and agreed to participate in the present study by signing a consent form.

The exclusion criteria were that the subjects did not have current injury or pain (bone fracture, bruise and/or sprain of upper arm), consumption of any form of drugs, pregnancy, hemophilia, diabetes mellitus, asthma, weight-training, intense fear of needles and participation in a similar experiment within the past year. All subjects were instructed to avoid any form of exercise during the experiment.

DOMS was induced with standard methods in the non-dominant elbow flexors (biceps brachii) of the subject 1314. Subjects were seated behind an inclined biceps-curl bench so that they could fully flex and extend their elbows. The maximum weight lifted with one voluntary concentric contraction was determined with free weights (a loaded dumbbell) for each subject. This maximum weight was subsequently used to exhaust the elbow flexors. A researcher lifted the weight until the subject's elbow was in a position of full flexion. The subject was instructed to lower the weight as slowly as possible until the elbow reached full extension. The researcher returned the weight to the starting position (full elbow flexion) and the process was repeated for as many times as the subject could control the speed of descent of the weight. The time of exhaustion was taken as the point at which the subject could no longer control the lowering of the weight.

A researcher screened and enrolled subjects. After the participants completed a baseline evaluation, another researcher who was not involved with data collection randomly assigned them to one of the three treatment groups using a computer-generated (SAMPSIZE V2.0, Blackwell Science Ltd, UK), blocked random allocation sequence with a block size of three.

The data are reported as means ± standard deviation (means ± SD). Repeated measures analysis of variance (ANOVA) was used for the primary measure, an inter-group of VAS scores. Repeated measures ANOVA were performed in each group for intra-group differences, followed by pair wise comparisons with Sidak correction to maintain experiment error rate below 5%. Differences between the groups at each time period were analysed with the unpaired t test and Dunnett's test, by which each test performed was judged against a corrected significance level as α/c, where α is significance level (0.05) and c is the number of paired comparisons. SYSTAT 11 was used for the analysis of a between-group (repeated measures ANOVA) and StatView for Windows (version 5.0) was used for the analysis of between the groups (Dunnett's test). P < 0.05 was considered to be statistically significant.

Thirty subjects (13 women, 17 men; aged 18–22 years) were randomly assigned to the groups. No differences were found among the groups in terms of baseline variables and age.

Three subjects in the control group and one subject in the non-TeP group dropped out. The reason for dropout was that they received other treatments (e.g. massage, anti-inflammatory poultice). The analyses were conducted with the data from the 26 patients who completed the study (Figure 1).

Immediately after exercise, subjects in all groups felt warmth and tenderness in the working muscle of the upper arm. The region of tenderness was gradually restricted to the musculotendinous junction, and a rope-like taut band was observed in the tender area after exercise on the first day of the study.

As shown in Table 1, the mean VAS scores tended to decrease from immediately after treatment, although the time courses were different for each group. There were statistically significant differences between the VAS scores of the control group and TeP group immediately (4.2 ± 2.7 vs 0.6 ± 0.9) and three days (2.4 ± 2.3 vs 0.4 ± 0.5) after treatment (P < 0.005 in both cases). There were no significant differences between VAS scores of the control group and non-TeP group. While the non-TeP group and TeP group reported relatively lower pain intensity than the control group, the differences were not statistically significant in ANOVA.

Available experimental evidence does not consistently support acupuncture as a method for pain relief 1516. Past studies were flawed in experimental design, sample size or control. In some randomized controlled trials of acupuncture, control groups were defined as no-treatment controls 17, pricking (without penetration) 18, minimal acupuncture (shallow and weak needling) 19 and mock TENS (without pulse) 2021. Moreover, negative results tended to come from studies using non-tender point acupuncture 10 and positive results from studies using non-acupuncture control groups 9. We used no-treatment controls and non-tender point acupuncture in this study. Non-tender point acupuncture has been proposed as a sham technique 2223, which is problematic due to the existence of diffuse noxious inhibitory controls (DNIC) phenomena. Painful stimulation inhibits pain, and DNIC has been proposed as a physiological basis of acupuncture analgesia 2425. In fact, six subjects complained of dull sensation (known as deqi) during non-tender acupuncture treatment. Therefore, we used non-tender point acupuncture as treatment group in this study.

Previous experience with acupuncture and confidence in acupuncture may influence the measurement of efficacy 2426. Therefore, the limitations of the present study is that the subjects were students of an acupuncture school, who had considerable knowledge of acupuncture and the special sensation of deqi, and who had confidence in acupuncture.

Our results show that tender point acupuncture has analgesic effects. The strength of stimulation may depend on various parameters such as manipulating procedure, size of needle and site of insertion. Tender point insertion of the needle affects sensitized nociceptors, whereas non-tender point insertion does not 2728. Tender points are sites where nociceptors, such as polymodal-type receptors, have been sensitized by various factors 2930. Our data suggest that acupuncture stimulation of tender points may activate sensitized polymodal-type receptors thereby relieving pain. In clinical practice, acupuncture treatment may be effective on myofascial pain syndromes.

Barlas et al. concluded that acupuncture had little effect on the cardinal signs and symptoms of DOMS. We find the claim problematic. Deep pain sensation, such as muscle pain, is very complex 31. A previous study found that the intensity of muscle pain at full flexion and full extension were different 32. Moreover, pressure algometry used in studying muscle pain excites cutaneous receptors and may induce pain sensation in the skin 33. In the present study, we estimated the intensity of muscle pain of the arm at full flexion.