How acupuncture reduces pain – an overview of its possible modes of action

Introduction

Acupuncture is an ancient healing technique involving the insertion of dry needles into specific sites of the body for the treatment of symptoms and conditions. It is generally believed to have originated in China around 200BC, and has been practised in other Asian countries such as Japan and Korea for thousands of years. Acupuncture was brought to Europe in the early 17th century, but a clash of paradigms meant that it was never accepted into mainstream medicine. The situation has changed dramatically in recent years, with complementary therapy becoming increasingly common in Western societies. Acupuncture is one of the most well-known and accepted of all CAT. In this article we look at the possible mechanisms by which acupuncture is thought to exert its analgesic effects.

Gate control theory

The perception of pain is dependent on the interaction of the various parts of the nervous system as they each process pain signals in their own way. On injury, pain is generated by small perceiving organs (nociceptors) located in the area of the damaged or distorted tissue, causing micro-currents to flow along the peripheral nerves to the spinal cord and therefore up to the brain.

According to the gate control theory of Melzack and Wall, these pain stimuli encounter ‘nerve gates’ in the spinal cord that open or close depending on signals from various parts of the nervous system, including descending signals from the brain.¹ When these ‘nerve gates’ are opened, the pain stimuli pass through more or less easily and pain is therefore experienced, which can in some instances be intense. When the ‘nerve gates’ close, pain stimuli are prevented from reaching the brain and can be reduced in severity or possibly not felt at all.

Pain stimuli reach the brain via A-delta and C fibres. Areas of the dorsal horn of the spinal cord receive input from large-diameter A-beta fibres as well as receiving connections from A-delta and C fibres; these A-beta fibres are non-nociceptive and regulate transmission of pressure and touch signals. The larger diameter of the A-beta fibres means that the signals they carry travel faster than those in A-delta fibres. Like a highway, the ‘nerve gates’ can handle only a limited number of nerve signals at one time. Pain stimuli in A-delta and C fibres travel slowly, while the A-beta nerve signals travel much faster. The faster signals crowd out the slower ones because of the limited capacity of the nerves. This effectively blocks the pain stimuli from reaching the brain and therefore pain is not experienced. It has been postulated that the insertion of needles in acupuncture points stimulates large numbers of A-beta fibres, therefore preventing the brain from perceiving the pain stimulus due to engagement of the synapses by A-beta nerve signals.²,³

The gate control theory was conceived to explain the general mechanisms of pain, and does not fully explain the mechanisms of acupuncture. According to the theory, only pain travelling along peripheral nerves leading to and from the spinal cord, that is from the neck down, would be capable of being blocked. The theory does not explain how acupuncture needles inserted in the face reduce pain during surgery of the abdomen. Man and Chen recently shed some light on this discrepancy with their two-gate theory,⁴ in which they expand on the original gate control theory, suggesting that there may be a second gate in the thalamus of the brain that closes to prevent pain sensations from reaching the cerebral cortex from below or above the spinal cord.

The increased number of impulses coming from A-beta fibres closes the gate in the substantia gelatinosa of the spinal cord, which prevents the passage of painful impulses from the A-delta and C fibres. Acupuncture analgesia works by stimulating the acupuncture points, which affects the A-beta nerve fibres resulting in a steady stream of ‘non-pain’ impulses being transmitted to the substantia gelatinosa causing the gate to close.

Hsiang-Tung Chang⁵ postulates that there could be at least four ‘nerve gates’ that control and modulate pain impulses. This theory holds that in the central nervous system (CNS) the activity of the larger nerve fibres depresses the activity of the smallest, including those that transmit pain, and that this happens at all levels of the CNS, not just in the spinal cord, but also involving the brainstem, thalamus and cerebral cortex. Melzack⁶ also notes that the brainstem reticular formation seems to have a powerful inhibitory effect on the transmission of pain signals within the CNS, indicating the possibility of a ‘nerve gate’ being located in that region. Stimulation through acupuncture needles could activate this inhibitory brainstem system by various pathways, thereby closing the ‘nerve gate’ to pain stimuli.

Transcutaneous electrical nerve stimulation (TENS), although different in mode of action from acupuncture in that needles are not used, was based on the gate control theory. Non-nociceptive fibres are selectively stimulated with electrodes to override the pain stimuli and thereby lessen pain. TENS is widely used in the treatment of chronic pain such as that of chronic low back pain, and is also commonly used in obstetric care, especially labour. However, its actual efficacy is debatable due to the lack of data, and use of TENS is still controversial in many circles. There is some research evidence that TENS reduces pain and improves range of motion in chronic back pain patients,⁷ but the general consensus is that the results are equivocal and inconsistent and that more well-designed studies with a standardised protocol and adequate number of subjects are needed to determine its clinical efficacy.⁸10 The ambiguity in the efficacy of TENS demonstrates the need for more than the gate control theory to explain the mechanism of acupuncture-induced analgesia. The gate control theory does provide the framework for acupuncture analgesia, but it is likely that acupuncture is a multi-modality phenomenon and that other processes are also involved.

Endogenous opiates

Opiate peptides and their receptors are widely distributed at low concentration in the CNS, especially along established nociceptive pathways. Three opiate proteins or peptides are known to be involved in analgesia: enkephalins, beta-endorphins and dynorphins. These opiate peptides produce their physiological effects after binding to the receptor sites, of which only three different types have been identified.

These indigenous opiates are so called because of their similarity in pharmacological action with external opiates such as morphine. They produce similar effects, including the desired analgesia, the only difference being that they are not addictive like morphine. Being produced internally means that the body never produces more of it than is needed, and an endogenous substance is unlikely to cause addiction within the host body. Interestingly, the analgesic effect of endogenous opiates is extremely powerful and many times stronger than that of morphine; it is 200 times higher in the case of endorphins and more than 400 times higher in the case of dynorphin. Beta-endorphin has been administered both within and outside the spinal cord resulting in prolonged analgesia, with few or none of the side-effects commonly associated with external opiates.¹¹,¹²

The role of endogenous opiates and other neurotrans-mitters in acupuncture analgesia was first demonstrated in the 1970s when cerebrospinal fluid (CSF) was removed from animals in which acupuncture analgesia had been induced.¹³ The donor CSF was transfused into a recipient animal, which subsequently developed analgesia too. This proved the existence of substances in the CSF that were capable of inducing analgesia; the challenge however was to prove that it was related to acupuncture in humans, and to determine the substances responsible for the analgesia.

Evidence for opiates

The experiment that threw acupuncture open to scientific research was one in which naloxone – a morphine antagonist – was found capable of reversing the effects of acupuncture analgesia. Mayer et al.¹⁴ carried out one of the first studies on naloxone reversal, with pain threshold being measured by electrical stimulation of dental pulp. Acupuncture analgesia was carried out by manual twirling of needles in the first dorsal interosseous of the hand. It was a double-blind study with one group receiving intravenous naloxone and another receiving just saline. The group on saline achieved analgesia with no problems, whereas the group on naloxone did not achieve analgesia. Unfortunately there was no control group receiving naloxone alone, leaving the possibility that naloxone could have induced hyper-algesia and therefore cancelled the effects of acupuncture analgesia. Subsequent studies have however shown that naloxone administered alone rarely causes hyperalgesia.¹⁵

Another of the early studies on naloxone reversal was done on mice with the use of electroacupuncture.¹⁶ Each group of mice received one of the following treatments: electroacupuncture alone, electroacupuncture plus saline, electroacupuncture plus intravenous naloxone, sham electroacupuncture in a false acupuncture point, naloxone alone, saline alone, and no treatment at all. The results were astoundingly definitive. Naloxone completely blocked the effects of acupuncture analgesia, while sham electroacupuncture did not have any effect. These landmark studies finally showed that acupuncture analgesia was not psychological in nature, and that it had a distinct physiological function that could be reversed by naloxone.

There have been numerous studies carried out since those early papers in which the endorphin antagonists naloxone and naltrexone have been administered to both animals and humans in an attempt to test the naloxone reversal hypothesis. It has been shown that naloxone does prevent acupuncture analgesia when given before the onset of the acupuncture treatment, but has no effect when administered immediately after the acupuncture treatment.¹⁷ This could be because the activation of endorphins by the acupuncture treatment sets off a cascade effect in which endorphin antagonists have no effect. The naloxone reversal hypothesis is therefore somewhat of a misnomer.

Most papers are similar to the two early naloxone reversal studies and have reported naloxone antagonism.¹⁸–20 However there have been studies that show no antagonism with naloxone, including the 1983 study by Chapman et al.,²¹ which was an attempt to follow in the footsteps of Mayer’s earlier experiment. Fourteen subjects who responded to acupuncture were given low-frequency electroacupuncture for 30 mins. Analgesia was measured by electrical stimulation to dental pulp. The subjects then received either 1.2 mg of naloxone or saline, chosen at random. Those who were given naloxone experienced a small reversal in the analgesia, which did not reach statistical significance. It should be noted that the sample size for the above experiment was small and this compromises its reliability and accuracy. It has been shown that the endorphin antagonists work best when given before the treatment, and that they are not that effective in reversing analgesia that has already been initiated. In Chapman’s experiment and other studies that showed no naloxone antagonism, the naloxone was administered only after analgesia had been induced in the subjects. This strongly suggests that the failure of those experiments to elicit antagonism was probably due to the bad timing of the naloxone administration. The number of successful experiments greatly outweighs the number of failures.

There is also criticism of the usage of animal models in some studies, with the concern that the analgesia experienced by the animals might be stress induced and not due to endorphin release.²⁰ This is a valid point as many experimental procedures might prove stressful to animals and provoke a different mechanism to that in humans. More studies need to be done to differentiate between stress-induced and acupuncture-induced analgesia.

Despite the doubts shown in certain quarters about the endogenous endorphin hypothesis, many studies have been carried out that strongly support the hypothesis. Many other endorphin antagonists, in addition to naloxone, have been shown to antagonise acupuncture analgesia.¹⁹,²⁰ It has also been shown that CSF and plasma concentrations of endorphins increase during acupuncture-induced analgesia, indicating a linkage between the two.²²–24 Lesions of or stimulation of the hypothalamic paraventricular nucleus in the midbrain, a site of endorphin production, also terminated acupuncture analgesia.²⁵ The evidence in favour of the endogenous opiate hypothesis is now overwhelming, and indicates that opiates definitely do play a role in acupuncture-induced analgesia.

Neural mechanisms

Current understanding is that acupuncture stimulates nerve fibres in skin and muscle, which then send signals to the spinal cord, resulting in the activation of three centres: spinal cord, midbrain and hypothalamus-pituitary gland, based around the theory of multiple ‘nerve gates’. In the spinal cord, low-frequency electro-acupuncture triggers the release of beta-endorphins and enkephalin, while high-frequency or TENS-like acupuncture increases the release of dynorphin in the spinal cord.²⁶ In the midbrain, the nucleus raphe magnus (NRM) is a key structure in acupuncture analgesia (Figure 1). Electroacupuncture possibly activates the NRM, mediating a negative feedback circuit modulating pain, thus inducing analgesia via descending inhibition.²⁷

Neurotransmitters involved in this process include enkephalin, serotonin and the monoamines.²⁸ The general consensus on the pituitary’s function is that it sends out beta-endorphins into the system to induce analgesia,²⁹ although it is possible that it may exert its action through other neurotransmitters.³⁰ In the hypothalamus, the paraventricular nucleus (PVH) plays an important role in acupuncture analgesia. It has numerous neural connections with spinal cord and midbrain structures involved in acupuncture analgesia such as the NRM. The PVH neurones are capable of syn-thesising neuropeptides that can modulate acupuncture analgesia such as the endorphins, oxytocin and vaso-pressin. A recent study by Yang et al.²⁵ seems to suggest that vasopressin, and not oxytocin and the endorphins, is responsible for the actions of PVH in acupuncture analgesia.

Figure 1. The nucleus raphe magnus receives signals from the spinothalamic tract and is thought to turn on the descending pain modulation circuit. (Taken from http://www.neuroanatomy.wisc. edu/virtualbrain/BrainStem/24PAG.html, accessed September 24, 2007.) ALS, anterolateral system; PAG periaqueductal grey.

The presence of three ‘nerve gates’ indicate a possible reasoning for the workings of acupuncture needling. Insertion of the needle in the tender area would most likely activate all three ‘nerve gates’, therefore producing maximal analgesia, while the insertion of needles at distal points away from the tender region would only activate the higher placed ‘nerve gates’ in the midbrain and hypothalamus and pituitary, thus producing a weaker intensity of analgesia. Placing needles at the tender site and sites distal to the tender region could accentuate the effect of the ‘nerve gates’ and enhance the overall analgesic effect.

Monoamines

Monoamines are thought to play a role in acupuncture analgesia, with a current focus on noradrenaline and serotonin (or 5-hydroxytryptamine, 5HT). Lesions to the NRM, an important store of serotonin cells in the body, result in the impairment of acupuncture analgesia.³¹,³² Several serotonin antagonists such as methy-sergide have been shown to inhibit the action of acupuncture analgesia in rabbits, with the results suggesting that 5-HT₁ (except 5-HT_1A); 5-HT₂ (except 5-HT_2A); and 5-HT₃ receptors are positively involved in electroacupuncture-induced analgesia.³³ The tricyclic antidepressant drug clomipramine blocks serotonin reuptake from nerve terminals, thereby increasing the amount of serotonin in the pre-synaptic cleft; the effect of the drug on acupuncture analgesia was tested in an RCT. Clomipramine was found to increase the analgesic effect of acupuncture during extraction of the third molar teeth.¹⁵

Less attention has been paid to noradrenaline, but there have been studies showing that injections of a noradrenaline antagonist (yohimbine) blocked the actions of acupuncture analgesia.³⁴ Strangely enough, noradrenaline has also been implicated in the exertion of an antagonistic effect on acupuncture analgesia with its actions on the pre-optic area.³⁵ This highlights a definite need to carry out more work on the monoamine hypothesis to determine the exact nature of the serotonin and noradrenaline pathways.

There is much more to learn about the exact role of the monoamines. It is likely that serotonin pathways from the NRM to the other brain structures may help mediate acupuncture analgesia. Serotonin may exert a synergistic effect with noradrenaline to block pain signals in the spinal cord, and other monoamines may be involved.

Cholecystokinin

Cholecystokinin (CCK) is a hormone secreted by the upper gut to stimulate the pancreas, and is also a neuro-transmitter released together with endogenous opiates during acupuncture analgesia. CCK does not however potentiate the effect of acupuncture analgesia, and instead acts as an antagonist to the opiates. CCK administered into the periaqueductal grey matter of rabbits was found to be capable of antagonising opioid analgesia by the activation of CCK receptors.³⁶ Electroacu-puncture analgesic effects were compared between knockout mice lacking a CCK-A receptor gene with normal mice, and the knockout mice showed an enhancement in electroacupuncture-induced analgesic effect of 53% compared to 35% in the normal mice.³⁷ An experiment done on rats with decreased brain CCK levels shows increased responsiveness to peripheral electrical stimulation-induced analgesia.³⁸ The increase in CCK levels after stimulation may limit the action of endogenous opiates, and may be responsible for the balance of nociceptive control.

Oxytocin

Oxytocin is a hormone that stimulates the pregnant uterus and also acts as a neurotransmitter in the brain. It has also been shown to have analgesic function –studies have shown that injection of oxytocin into rats being given electroacupuncture brought about a great increase in pain threshold, while the administration of antioxytocin serum impaired analgesia induction.³⁹,⁴⁰ There is evidence to suggest that oxytocin brings about its effects through a pathway independent of endogenous opiates; more work is currently being undertaken in this area.

Discussion

The growing interest in acupuncture as a viable scientific technique has given rise to numerous studies focusing on elucidating the workings of this ancient healing tool. This has aided the advancements made in the discovery of the neurophysiology of acupuncture. However, many questions still remain, and there are many areas that can still be improved on.

Animal models have been used extensively in many of the studies, and this has given rise to its own set of unique problems. Animals like rats and rabbits could have different mechanisms for pain signalling, which could affect the way acupuncture induces analgesia in them as compared to humans. Pain is a multi-sensory modality, and examining isolated tracts of the nervous system or looking at animal cells might not be able to cover the whole picture of the perception of pain. Human pain thresholds vary significantly, and there is no reason to suppose variability in animals is dissimilar, and they cannot signify their thresholds. Animals have however been used to test a multitude of new drugs, and should still prove relevant and useful if only in a laboratory setting.

Conducting an acupuncture trial can be challenging. Unlike the testing of new drugs, one problem with acupuncture is the lack of a tangible dosage method. There are many variables to consider when it comes to setting an acupuncture dose during the trial – the number of needles to use for each subject, how far to insert the needle and the thickness of the needle. Add to that the choice of manual or electrical acupuncture and then try to standardise the remaining variables such as couches, lighting and attitudes of the therapist!

Controls in acupuncture trials have been another contentious point. It has been difficult to implement good control subjects in acupuncture trials, mainly due to the problem of finding a proper placebo. Early trials did not include controls, which greatly hampered their findings. Sham acupuncture controls used nowadays include non-invasive ones that apply pressure only, without piercing the skin, and invasive ones that actually penetrate the skin but not deep enough to elicit the effects of acupuncture. A novel way of approaching this is to combine both invasive and non-invasive methods to create the placebo control. One of the biggest acupuncture studies ever conducted utilised a placebo control in which real needles were inserted into sham points, while a plastic needle tube was placed at real acupuncture points producing pressure with no insertion.⁴¹ The search is now on for a sham acupuncture control that would be no different in the eyes of the control group from real acupuncture. One of the main problems facing researchers is the similarity in analgesic mechanisms of both acupuncture and placebo: it will be interesting to try and separate the two.

Acupuncture trials need to be planned like drug trials, and attention has to be paid to methodology and statistical accuracy. Sample size has been a problem with acupuncture trials, and a large sample size is required to show up the small differences. The Maryland study in 2004 on osteoarthritis⁴¹ included 570 participants, one of the largest sample sizes in any acupuncture trial, and found that acupuncture was of benefit in osteoarthritis. More rigorous trials like this are needed to determine the efficacy of acupuncture.

Conclusion

The burgeoning popularity of acupuncture has almost compelled its integration into modern healthcare services, mainly by professionalizing and then licensing acupuncture practitioners. The effectiveness of acupuncture in specific clinical conditions has, however, only recently been considered – it is essential that scientific investigations into acupuncture continue in the effort to unearth new discoveries about the mechanism by which acupuncture exerts its effects on the human body.

The traditional Chinese medical philosophy of treating disease by correcting any underlying bodily imbalances in a holistic manner may currently be an anathema to allopathic or reductionist medicine. Therapists’ emphases on communication, establishment of a good rapport and incorporation of patients’ views are seen as the key to successful treatment. Differing cultural expectations, as for instance in attitudes to psychotherapy or to orthodox Western medication, may involve differing placebo effects. In addition, the dynamics of a particular society and the influences of the cultures present in it determine the use and practice of acupuncture.⁴²

Interest in scientific explanations for acupuncture has mirrored increasing awareness and usage by the general public. Treatment outcomes in different contexts have not yet been studied, and popularity is partly driven by widely reported but varying anecdotal reports of benefit. While analysis of psychosocial antecedents of and responses to acupuncture has not yet been thought necessary, it is frequently thought that personal attitudes and cultural milieu can have a significant effect on the well-being and physical condition of the client, and thus their perception of the treatment’s effectiveness.⁴³

Vital to the integration of acupuncture into a health-care setting is to provide evidence of the efficacy, safety and cost of acupuncture so as to ensure that an ordinary person with inadequately controlled pain can benefit from such integration. The usefulness of acupuncture in treating pain associated with osteoarthritis of the knee has been shown, but more clinical trials are needed to prove its usefulness with other conditions. Understanding the biophysicochemical aspects of acupuncture is needed to fully embrace acupuncture analgesia as part of the Western medicine framework. Significant advancements have been made in efforts to elucidate acupuncture analgesia mechanisms, with endogenous opiates, monoamines and oxytocin being implicated. However, more rigorous trials are needed to expand further on these theories. The methodology of acupuncture trials has to be sorted out, with factors like sample size and dosage methods of particular concern. Ultimately, the move to integrate acupuncture into the Western health-care system should be carried out based on clear and scientific evidence, rather than relying on anecdote.

Acknowledgement

We would like to thank Dr Malcolm McCoubrie of the St George’s Division of Mental Health, St George’s University of London, UK, for support and help.

References

1. Melzack R, Wall PD. Pain mechanisms: a new theory. Science 1965; 150: 171–9.
2. Lu GW. Characteristics of afferent fiber innervation on acupuncture points zusanli. Am J Physiol 1983; 245: R606–12.
3. Guowei L, Rongzhao L, Jingqiang X et al. Role of peripheral afferent nerve fiber in acupuncture analgesia elicited by needling point zusanli. Sci Sin 1979; 22: 680–92.
4. Man PL, Chen CH. Mechanism of acupunctural anesthesia. The two-gate control theory. Dis Nerv Syst 1972; 33: 730–5.
5. Hsiang-Tung C. Neurophysiological basis of acupuncture analgesia. Sci Sin 1978; 21: 829–46.
6. Melzack R. Acupuncture and pain mechanisms. Anaesthetist 1976; 25: 204–7.
7. Gadsby JG, Flowerdew MW. Transcutaneous electrical nerve stimulation and acupuncture-like transcuta-neous electrical nerve stimulation for chronic low back pain. Cochrane Database Syst Rev 2000; 2: CD 000210.
8. Khadilkar A, Milne S, Brosseau L et al. Transcutaneous electrical nerve stimulation for the treatment of chronic low back pain: a systematic review. Spine 2005; 30: 2657–66. [Abstract]
9. Sluka KA, Walsh D. Transcutaneous electrical nerve stimulation: basic science mechanisms and clinical effectiveness. J Pain 2003; 4: 109–21. [Abstract]
10. Brosseau L, Judd MG, Marchand S et al. Transcutane-ous electrical nerve stimulation (TENS) for the treatment of rheumatoid arthritis in the hand. Cochrane Database Syst Rev 2003; 3: CD 004377.
11. Oyama T, Fukushi S, Jin T. Epidural beta-endorphin in treatment of pain. Can Anaesth Soc J 1982; 29: 24–6.
12. Oyama T, Jin T, Yamaya R et al. Profound analgesic effects of beta-endorphin in man. Lancet 1980; 1: 122–4.
13. Research Group of Acupuncture Anaesthesia, Peking Medical College. The role of some neurotransmitters of brain in finger-acupuncture analgesia. Sci Sin 1974; 17: 112–30.
14. Mayer DJ, Price DD, Rafii A. Antagonism of acupuncture analgesia in man by the narcotic antagonist naloxone. Brain Res 1977; 121: 368–72. [Abstract]
15. Stux G, Berman B, Pomeranz B. Basics of Acupuncture. 5th edn. Berlin: Springer, 2003.
16. Pomeranz B, Chiu D. Naloxone blocks acupuncture analgesia and causes hyperalgesia: Endorphin is implicated. Life Sci 1976; 19: 1757–62. [Abstract]
17. Pomeranz B, Bibic L. Electroacupuncture suppresses a nociceptive reflex: naltrexone prevents but does not reverse this effect. Brain Res 1988; 452: 227–31. [Abstract]
18. Chao DM, Shen LL, Tjen-A-Looi S et al. Naloxone reverses inhibitory effect of electroacupuncture on sympathetic cardiovascular reflex responses. Am J Physiol 1999; 276: H2127–34.
19. Chen XH, Han JS. Analgesia induced by electroacu-puncture of different frequencies is mediated by different types of opioid receptors: Another cross tolerance study. Behav Brain Res 1992; 47: 143–9. [Abstract]
20. Chen XH, Geller EB. Electrical stimulation at traditional acupuncture sites in periphery produces brain opioid-receptor-mediated antinociceptin in rats. J Pharmacol Exp Ther 1996; 277: 654–60.
21. Chapman CR, Benedetti C, Colpitts YH, Gerlach R. Naloxone fails to reverse pain thresholds elevated by acupuncture: acupuncture analgesia reconsidered. Pain 1983; 16: 13–31. [Abstract]
22. Ho WK, Wen HL. Opioid-like activity in the cerebro-spinal fluid of pain patients treated by electroacupunc-ture. Neuropharmacology 1989; 28: 961–6. [Abstract]
23. Hardebo JE, Ekman R, Eriksson M. Low CSF met-enkephalin levels in cluster headache are elevated by acupuncture. Headache 1989; 29: 494–7. [Abstract]
24. Iguchi Y, Tokuda H, Tamura S et al. Effects of electro-acupuncture on beta-endorphin contents in rats. [In Japanese]. Nippon Yakurigaku Zasshi 1985; 86: 105–14.
25. Yang J, Liu WY, Song CY, Lin BC. Only arginine vaso-pressin, not oxytocin and endogenous opiate peptides, in hypothalamic paraventricular nucleus play a role in acupuncture analgesia in the rat. Brain Res Bull 2006; 68: 453–8. [Abstract]
26. Han JS. Acupuncture and endorphins. Neurosci Lett 2004; 361: 258–61. [Abstract]
27. Liu X, Zhu B, Zhang SX. Relationship between elec-troacupuncture analgesia and descending pain inhibitory mechanism of nucleus raphe magnus. Pain 1986; 24: 383–96. [Abstract]
28. Fan T, Li J, Kong T. Relationship between acupuncture analgesia and neurotransmitters in nucleus raphe magnus. [In Chinese]. Zhen Ci Yan Jiu 1993; 18: 168–71.
29. Takeshige C, Nakamura A, Asamoto S, Arai T. Positive feedback action of pituitary beta-endorphin on acupuncture analgesia afferent pathway. Brain Res Bull 1992; 29: 37–44. [Abstract]
30. Iguchi Y, Ozaki M, Kishioka S et al. The role of the pituitary in the development of electroacupuncture analgesia in rats. [In Japanese]. Nippon Yakurigaku Zasshi 1985; 85: 453–65.
31. Jiang M, Liu X. The lesion of somatosensory area II of cerebral cortex reducing the effects of electroacupunc-ture of ‘zusanli’ on nucleus raphe magnus in rats. [In Chinese]. Zhen Ci Yan Jiu 1994; 19: 4–7.
32. Liu X. The modulation of cerebral cortex and subcortical nuclei on NRM and their role in acupuncture analgesia. [In Chinese]. Zhen Ci Yan Jiu 1996; 21: 4–11.
33. Takagi J, Yonehara N. Serotonin receptor subtypes involved in modulation of electrical acupuncture. Jpn J Pharmacol 1998; 78: 511–14. [Abstract]
34. Cheng RS, Pomeranz B. Monoaminergic mechanism of electroacupuncture analgesia. Brain Res 1981; 215: 77–92. [Abstract]
35. Zhu J, Xia Y, Cao X. Effects of noradrenaline and dopamine in preoptic area on acupuncture analgesia. [In Chinese]. Zhen Ci Yan Jiu 1990; 15: 117–22.
36. Cao W, Zhou ZF, Han JS. Antagonism of morphine analgesia and electroacupuncture analgesia by chole-cystokinin octapeptide (CCK-8) administered in peri-aqueductal gray (PAG) of the rabbits. [In Chinese]. Sheng Li Xue Bao 1989; 41: 388–94.
37. Lee GS, Han JB, Shin MK et al. Enhancement of elec-troacupuncture-induced analgesic effect in cholecysto-kinin-A receptor deficient rats. Brain Res Bull 2003; 62: 161–4. [Abstract]
38. Zhang LX, Li XL, Wang L, Han JS. Rats with decreased brain cholecystokinin levels show increased responsiveness to peripheral electrical stimulation-induced analgesia. Brain Res 1997; 745: 158–64. [Abstract]
39. Song CY, Liu WY, Yang J et al. The role of central oxytocin in electroacupuncture analgesia. [In Chinese]. Sheng Li Xue Bao 1990; 42: 169–74.
40. Liu W, Song C, Yang J et al. Involvement of oxytocin in spinal cord in acupuncture analgesia. [In Chinese]. Zhen Ci Yan Jiu 1990; 15: 24–9.
41. Berman BM, Lao L, Langenberg P et al. Effectiveness of acupuncture as adjunctive therapy in osteoarthritis of the knee: a randomized, controlled trial. Ann Intern Med 2004; 141: 901–10. [Abstract]
42. Birch SJ, Felt RL. Understanding Acupuncture. London: Churchill Livingstone, 1999.
43. Klineman A. Patients and Healers in the Context of Culture. Berkeley, CA: University of California Press, 1980.