Article Text


Effect of bidirectional rotation of an acupuncture needle at LI10 on acupuncture needle sensation and experimentally-induced contact heat pain in healthy human volunteers
  1. Alex Benham1,2,3,
  2. Mark I Johnson1,3
  1. 1Faculty of Health and Social Sciences, Leeds Metropolitan University, Leeds, UK
  2. 2Bradford Institute of Health Research, Bradford Royal Infirmary, Bradford, West Yorkshire, UK
  3. 3Leeds Pallium Research Group, Leeds, UK
  1. Correspondence to Professor Mark I Johnson, Faculty of Health and Social Sciences, Leeds Metropolitan University, Civic Quarter, Leeds LS1 3HE, UK; M.Johnson{at}


Introduction There is insufficient evidence of a relationship between acupuncture needle sensations (de qi) and hypoalgesia. The aim of this study was to investigate the effects of bidirectional needle rotation at LI10 on acupuncture needle sensations and heat pain thresholds.

Methods Twenty-two healthy participants received one acupuncture needle at LI10 with bidirectional rotation of the needle in one experimental session and one acupuncture needle at LI10 with mock rotation in a separate session, in a randomised order. Measurements of heat pain thresholds were taken before needle insertion, during needle retention and 15 min after needle removal. At each measurement time point, participants rated needle sensations using the Massachusetts Acupuncture Sensation Scale (MASS) and a visual analogue scale (VAS) of overall intensity of needle sensation.

Results Bidirectional needle rotation produced significantly higher scores for VAS, MASStotal, MASSpain and MASSsensation compared with mock rotation (all p<0.001). There were significantly higher pain thresholds relative to pre-intervention baseline during (p=0.014) and after (p<0.001) bidirectional needle rotation but not during (p=0.1) or after (p=0.62) mock bidirectional needle rotation. Bidirectional needle rotation increased the pain threshold relative to baseline 15 min after the needles were removed (p=0.009). A significant but low correlation between needle sensation and change in pain threshold after needling was only found when data from mock and rotation interventions were combined.

Conclusions Needle rotation increases the magnitude of hypoalgesia. There is tentative evidence that needle sensation may be associated with the amount of change in pain threshold.

  • Acupuncture
  • Pain
  • Hypoalgesia
  • Acupuncture Needle Sensation
  • De qi
  • Experimental Pain

Statistics from


Manipulation of acupuncture needles is used to increase the intensity of stimulation to improve hypoalgesia. It has been suggested that this may be due to coupling between the needle and sensory receptors and/or their axons, and that this phenomenon may be reflected in acupuncture needle sensations, sometimes referred to as de qi.1–3 Research on the topic is sparse,4 although White et al5 found that the intensity of needle sensation was not correlated with pain relief in patients with osteoarthritis of the hip and knee. Recently, Park et al6 concluded that there was insufficient research to determine whether there was a relationship between treatment effect and acupuncture needle sensations.

Studies investigating the effects of needle manipulation on experimental pain have been conflicting, with studies failing to detect superior effects for manual acupuncture compared with controls.7 Bidirectional rotation of a needle inserted at LI10 has been shown to generate more intense needle sensations than mock bidirectional rotation.4 ,8 Needle sensations were both painful and non-painful. Brain imaging studies have found that painful needle sensations were associated with activation in areas of the brain involved in pain processing and may reflect poor needling technique, whereas non-painful acupuncture needle sensations have been found to deactivate areas of the brain involved in pain processing, suggestive of hypoalgesia.9–13

Kong et al14 reported that manual acupuncture reduced experimentally-induced thermal pain in 11 healthy participants, and that this was related to needle sensations of numbness and soreness but not stabbing, throbbing, tingling, burning, heaviness, fullness or aching. Recently, Choi et al15 reported that deep needling with manipulation significantly increased pressure pain thresholds compared with superficial needling and deep needling. Also, deep needling did not increase pain thresholds compared with superficial needling.

The aim of the present study was to investigate the effects of bidirectional rotation of a needle inserted into ‘deep’ tissue at LI10 on acupuncture needle sensations and experimentally-induced contact heat pain.


A crossover design was used where all participants received both deep acupuncture with bidirectional rotation of the needle in one experimental session and deep needling without stimulation (mock rotation) in the other session. The order of interventions was randomised for each participant to control for order effects.

Recruitment and selection

Twenty-two unpaid healthy human volunteers aged >18 years were recruited from staff and students at Leeds Metropolitan University using announcements in lectures. Individuals expressing an interest in the study were invited to attend a familiarisation session where they were given a participant information pack. During the familiarisation session, volunteers were screened and excluded if they had a chronic medical condition, a fear of needles causing severe anxiety or any skin disorders in the area surrounding the LI10 acupuncture point. They were given an opportunity to experience the experimental pain procedure and to have an acupuncture needle inserted into the LI4 point, which was away from the point used in the experiment (all volunteers took up this offer). Volunteers were then invited to take part in the experiments and were given a minimum of 48 h to consider whether they wished to participate. They were informed that they could withdraw consent at any time and without reason.


Participants took part in two identical experimental sessions. They received acupuncture with bidirectional rotation of the needle in one session and acupuncture with mock bidirectional rotation of the needle in the other. They were told that needling may produce sensations and that these may or may not cause some discomfort. Computerised block randomisation was used to determine the order of the rotated and mock rotated interventions. A washout period of at least 72 h was used between interventions to allow any minor bruising from needling in the first experimental session to subside. If tenderness in the area was reported at the second visit, it was planned that the second experimental session would be postponed to a later date. This did not happen for any participant. Participants lay on a plinth with their head elevated and their non-dominant hand supported on a table to the side of the plinth. A screen was used to conceal the acupuncture procedures. The acupuncture intervention was administered by the principal investigator (AB) who is an acupuncture trained physiotherapist and was not blind to the intervention allocation but was blind to the responses given to all of the assessment measures.

Experimental sessions lasted 30 min and consisted of a baseline measurement of pain threshold followed by a 15 min acupuncture intervention. During acupuncture, needles were inserted and the acupuncturist performed either deep needling with rotation or deep needling with mock needle rotation for 10 min. Measures of acupuncture needle sensation and heat pain threshold were taken in the remaining 5 min with the needle in situ. The needles were then removed and needle sensation and pain threshold measures were taken for a further 15 min (figure 1).

Figure 1

Timeline of experimental procedures at each time point across the experiment. MASS, Massachusetts Acupuncture Sensation Scale; VAS, visual analogue scale.


In both sessions, one single-use sterile disposable acupuncture needle (C&G, UK, 40 mm × 0.25 mm) was inserted at LI10 of the non-dominant forearm using a guide tube.

Bidirectional needle rotation

The rotated intervention consisted of a needle being placed directly into the deep soft tissues of the forearm to a depth of approximately 15–25 mm determined by inserting the 40 mm needle to approximately half of its length. The needle was rotated approximately 360° clockwise followed by 360° anticlockwise at approximately two rotations per second (ie, 2 Hz) for 30 s followed by 30 s of no rotation. This procedure was carried out every minute for 10 min.

Mock bidirectional needle rotation (control)

An identical procedure was used for inserting and removing the needle during the mock rotation intervention, except the investigator did not touch the needle after it had been inserted. Bidirectional rotation was imitated behind a screen placed over the arm out of view of the participant. Mock rotation was performed by rotating an empty guide tube between the thumb and index finger clockwise and anticlockwise at a frequency of 2 Hz for 30 s followed by 30 s of no rotation for a total of 10 min. The guide tube did not touch the participant's skin and the investigator's arm remained behind the screen and out of view of the participant.

Outcome measures

Acupuncture needle sensations

Participants were asked to quantify the overall intensity of acupuncture needle sensation on a 100 mm visual analogue scale (VAS) that stated “During acupuncture, I felt a sensation at the acupuncture point as …” (0 mm=‘not at all’, 100 mm=‘very strong’). In addition, participants completed the Massachusetts Acupuncture Sensation Scale (MASS) which uses 12 100 mm VAS to quantify the intensity of overall needle sensations (MASSTotal) with 10 subscales for non-painful sensations (soreness, aching, deep pressure, heaviness, fullness/distension, tingling, numbness, warmth, cold and throbbing, ‘MASSsensation’) and two painful sensations (dull pain and sharp pain ‘MASSpain’). Participants completed these out of view of the investigator.

Heat pain threshold

A TSA-II Neurosensory Analyser (Medoc Ltd, Israel) was used to administer heat stimuli across onto the skin of the dorsolateral aspect of the forearm. Heat stimuli were administered using a 30×30 mm thermode placed onto the surface of the skin approximately 5 cm distal to the location of the needle and secured for the duration of the experiment via a Velcro strap. The heat pain threshold was determined using the method of ascending limits with temperature starting at 32°C and rising at a rate of 1°C/s until the participant experienced the first pain sensation, at which point the participant pressed a response button which stopped the stimulus. Standardised instructions were read at the start of each set of four pain measurements and the mean values of these were used for analysis.

Data analysis

A sample size was calculated using data from a previous study using contact heat pain which found that 21 participants would be needed to detect a difference of 3°C with a SD of 3.3, an α of 0.05 and a β of 0.8.16 To allow for 5% attrition, a target sample of 22 participants was used.

A within-subject 3×2 factorial repeated measures analysis of variance (ANOVA) was performed on the heat pain threshold. Factors were time (three levels: baseline, during needling (at 15 min) and after needling (at 30 min)) and intervention (two levels: rotation and mock rotation intervention). The order (sequence) of receiving rotation or mock rotation was used as a between-subject factor (two levels: rotation and mock rotation). If Mauchly's test showed sphericity could not be assumed, a Greenhouse–Geisser correction was used. α was set at 0.05, β at 0.2 (ie, power of 0.8) and adjustments were made for multiple comparisons using the Bonferroni correction.


Twenty-two volunteers expressed interest in taking part in the study. All took part in the familiarisation session and completed both experimental sessions (11 men, mean (SD) aged 25.05 (4.32) years).

Intensity of needle sensation

There were significantly higher scores for VAS (p<0.001), MASStotal (p<0.001), MASSpain (p<0.001) and MASSsensation (p<0.001) for the bidirectional needle rotation intervention compared with mock bidirectional rotation (table 1).

Table 1

Mean (SD) scores for each of the four measures of needle sensation for the two interventions

Heat pain threshold

Mauchley's test showed that sphericity of the heat pain measures could not be assumed so Greenhouse–Geisser corrections were used. There were no between-subject effects for the order (sequence) in which the participants received the interventions (F (1, 20) 1.14, p=0.298). Significant effects were detected for time (F (1.52, 30.42) 17.62, p<0.001) but not intervention (F (1, 20) 0.46, p=0.505). Significant effects were detected for time×intervention interaction (F (1.37, 27.30) 6.17, p=0.012).

There were no significant differences in the pain threshold between bidirectional needle rotation and mock bidirectional needle rotation interventions at baseline (p=0.98) or during needling (p=0.830). Fifteen minutes after removal of the needles the pain threshold was higher in the bidirectional needle rotation intervention group than in the mock bidirectional needle rotation intervention group (after needling, p=0.009; table 2). Pairwise comparisons found that the pain threshold relative to pre-intervention baseline was significantly higher during (p=0.014) and after (p<0.001) bidirectional needle rotation but not during (p=0.1) or after (p=0.62) mock bidirectional needle rotation (table 2, figure 2).

Table 2

Post-test statistical comparisons of the change in pain threshold from baseline and the difference between the two groups throughout the experiment

Figure 2

Mean (mm) heat pain threshold for each intervention across all experimental time points.

Relationship between the intensity of needle sensation and hypoalgesia

There were no significant correlations between the pain threshold relative to baseline after needling and the overall intensity of needle sensation (VAS) for bidirectional needle rotation (r=0.41, p=0.857) or mock bidirectional needle rotation (r=−0.41, p=0.85). A significant correlation was found between the change in pain threshold relative to baseline and the intensity of needle sensation when data from both real and mock bidirectional needle rotation were combined (VAS, r=0.36, p=0.014). This relationship can be seen in figure 3.

Figure 3

Scatter plot of the association between needle sensation intensity (mm) and change in pain threshold (°C) at the final experimental time point.


The findings of this study showed that a single acupuncture needle inserted at LI10 with bidirectional rotation increased the heat pain threshold 15 min after removing the needle compared with mock rotation. The mock bidirectional needle rotation intervention controlled for non-specific effects associated with receiving an intervention and enabled us to isolate effects attributable to the bidirectional needle rotation per se. We are therefore confident that the increase in the pain threshold after receiving bidirectional rotation was due to a specific physiological action. Our study failed to detect a correlation between the intensity of needle sensation and the change in pain threshold relative to baseline after needling when the two interventions were analysed separately. Further analysis found that, when both interventions were combined, there was a significant but low correlation between needle sensation intensity and change in pain thresholds from baseline.

In our study, rotation of the needle was attributed to an increase in the heat pain threshold and supports the view that needles should be manipulated to elicit hypoalgesia.1 ,17 Choi et al15 investigated the relationship between acupuncture needle sensation and pressure pain threshold during superficial needling (0.3 cm), deep needling (2 cm) and needling with bidirectional rotation on 53 healthy volunteers. They found that bidirectional rotation produced the largest increase in the pressure pain threshold and the highest ‘total acupuncture sensation’, followed by deep needling and superficial needling. Our study included a mock bidirectional needle rotation intervention in order to isolate effects attributable to the bidirectional needle rotation, and are consistent with Choi et al.15 This adds further evidence that bidirectional rotation has an important role in hypoalgesia.

Divergence of hypoalgesia between real and mock needle rotation increased across the time course of our experiment and was most apparent after needling, suggesting a delay in the development of hypoalgesic effects over and above that seen by mock rotation. This is of interest as the difference from baseline was not significantly different between interventions during needle retention but was sufficient to reach statistical significance 15 min after the needles had been removed (after needling). This delay suggests that a centrally mediated endogenous analgesic mechanism (eg, activation of descending inhibitory pain pathways) may be elicited by needling rather than a local stimulation-specific analgesic mechanism. It would have been interesting to see if the difference in pain thresholds between the real and mock rotation may have increased further if the experiment was continued for a longer period of time.

White et al18 suggested that at least four needles inserted for at least 20 min and eliciting needle sensation is a prerequisite for an adequate dose of acupuncture. The intervention applied in this study was specifically intended to provide fewer needles for less time and with the minimum manipulation required to provide a strong needle sensation so that the individual contribution of needle sensation could be isolated as much as possible. Overall, acupuncture needle sensation intensity was similar between this study (mean (SD) 55.86 (13.88) mm) and that of Choi et al15 who reported approximate intensities of 50 mm. We used one needle that was rotated for 10 min followed by 5 min of no rotation, whereas Choi et al15 administered acupuncture using four needles (SP6, SP9, ST36 and GB39), although the interventions only lasted 5 min. Whether there were differences in the area of distribution of the needle sensations between the studies is not known. An interesting and testable hypothesis for future studies could be that greater numbers of needles generate larger areas of needle sensation; whether this is related to the area of hypoalgesia remains to be investigated.

At present there is insufficient evidence to determine whether acupuncture needle sensation is related to pain relief in clinical practice.19 Recently, Spaeth et al20 found that acupuncture generated stronger needle sensations (de qi) and better clinical outcomes in patients with chronic pain from knee osteoarthritis than sham acupuncture. White et al5 conducted a secondary analysis of data from a single-blind randomised controlled trial and found that the presence and intensity of needle sensations (de qi) did not influence osteoarthritic knee pain. Further clinical studies are therefore warranted to support the increasing number of healthy human studies characterising needle sensation, investigating the physiological basis of the phenomenon and evaluating the relationship with hypoalgesia.

The mean (SD) magnitude of the change after needling in the rotation intervention was 2.43 (1.94)°C, which is less than the 3°C that our study was powered to detect. This suggests that our study may have been underpowered or that the power calculation was insufficient for the intended study design, so the results must be interpreted accordingly. The intended change used in the power calculation was chosen from previous studies that have reported meaningful changes in heat pain thresholds. Differences in pain threshold between real and mock rotation were diverging at the termination of the experiment and therefore a larger magnitude of difference may have been apparent if an experiment of longer duration had been conducted. Methodological challenges in quantifying needle sensation have been described by other investigators, and this may have contributed to redundancy in the higher end of needle sensation scales.5 ,14 It is also possible that needle sensation has no role in therapeutic response or pain relief, and more studies on healthy humans are needed.

The weak association between needle sensation intensity and change in pain threshold after needling suggests that needle sensation intensity may be indicative of hypoalgesic effects. However, the exploratory nature of this analysis means that no firm conclusions can be drawn. This is an area that warrants further research.

In conclusion, bidirectional rotation of an acupuncture needle inserted to a depth of 15–25 mm at LI10 increased heat pain thresholds compared with mock bidirectional rotation, suggesting that needle rotation increases the magnitude of hypoalgesia. A weak positive correlation was detected between needle sensation intensity elicited during bidirectional rotation and the magnitude of change of the pain threshold. Assay sensitivity and inadequate length of post-needling measures may have contributed to this weak relationship. Further studies using more needles that are rotated and left in situ for longer periods of time with extended duration of the measurement phase are planned.

Summary points

  • We investigated the effect on the pain threshold of rotating needles and eliciting a sensation.

  • We compared the effect of rotation and mock rotation.

  • Rotation produced greater sensations and a (delayed) increase in pain threshold.


We wish to thank Dr Osama Tashani for technical assistance and Dr Gillian Phillips for constructive advice.


View Abstract


  • Contributors AB and MIJ designed the experiment and produced the ethics application. AB conducted the experiment and data analysis and both authors contributed to the writing of the manuscript.

  • Competing interests None.

  • Ethics approval Ethics approval was obtained from Leeds Metropolitan University.

  • Patient consent Obtained.

  • 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.