Intended for healthcare professionals

Clinical Review State of the Art Review

Management of chronic pain using complementary and integrative medicine

BMJ 2017; 357 doi: (Published 24 April 2017) Cite this as: BMJ 2017;357:j1284
  1. Lucy Chen, medical doctor and associate professor1 2,
  2. Andreas Michalsen, medical doctor and professor of medicine3 4
  1. 1MGH Center for Translational Pain Research, Pain Management Center of MGH, Department of Anesthesia, Critical Care and Pain Medicine, Massachusetts General Hospital, Boston, MA 02114, USA
  2. 2Harvard Medical School, Boston, MA 02115, USA
  3. 3Institute for Social Medicine, Epidemiology and Health Economics and Immanuel Hospital Berlin, Berlin, Germany
  4. 4Charité - Universitätsmedizin Berlin, Berlin, Germany
  1. Correspondence to: L Chen llchen{at}


Complementary and integrative medicine (CIM) encompasses both Western-style medicine and complementary health approaches as a new combined approach to treat a variety of clinical conditions. Chronic pain is the leading indication for use of CIM, and about 33% of adults and 12% of children in the US have used it in this context. Although advances have been made in treatments for chronic pain, it remains inadequately controlled for many people. Adverse effects and complications of analgesic drugs, such as addiction, kidney failure, and gastrointestinal bleeding, also limit their use. CIM offers a multimodality treatment approach that can tackle the multidimensional nature of pain with fewer or no serious adverse effects. This review focuses on the use of CIM in three conditions with a high incidence of chronic pain: back pain, neck pain, and rheumatoid arthritis. It summarizes research on the mechanisms of action and clinical studies on the efficacy of commonly used CIM modalities such as acupuncture, mind-body system, dietary interventions and fasting, and herbal medicine and nutrients.


Chronic pain, a term that often refers to pain conditions that last more than three months,1 is a common reason for patients to seek medical assistance. Clinically relevant chronic pain conditions include back and neck pain, migraine and other headaches, osteoarthritis, rheumatic arthritis, fibromyalgia, neuropathic pain, and cancer related pain. This review focuses on chronic back and neck pain and rheumatoid arthritis as important examples of chronic pain owing to their high incidence. Although advances have been made in pharmacological and interventional (eg, nerve block) treatments for chronic pain, it remains inadequately controlled for many people. Moreover, side effects and complications of treatment, such as addiction to opioid analgesics, kidney failure, or gastrointestinal bleeding due to long term use of drugs, make the management of chronic pain difficult.

The concept of complementary and integrative medicine (CIM) encompasses both Western-style medicine and complementary health approaches as a new combined approach to treat a variety of clinical conditions. CIM may have a unique role in chronic pain management because the multidimensional nature of the pain experience requires a multimodality treatment approach. Recent advances in basic science and clinical research on CIM have substantially increased patients’ awareness about the potential therapeutic use of CIM.

This review summarizes the evidence from basic science and clinical research on the role of CIM in clinical symptoms (eg, pain) associated with rheumatoid arthritis and chronic neck and back pain.


In the US, about 25.3 million adults have daily chronic pain and 23.4 million adults experience a substantial level of pain.2 The incidence of chronic low back pain, neck pain, and arthritic pain can be as high as 29%, 15.7%, and 28%, respectively, in American adult populations.3 According to the World Health Organization’s 2010 Global Burden of Disease Study estimation, low back pain is among the top 10 clinical conditions that affect all age groups, peaking at ages 35 to 55 years. The lifetime prevalence of low back pain is estimated at 60-70% in several countries.4

According to a 2012 US National Health Interview Survey (NHIS), 33.2% of adults and 11.6% of children (4-17 years old) have used CIM, with the leading indication being chronic pain.5 Women were more likely than men to use CIM,6 and the out-of-pocket spending on CIM totals $30.2bn (£24.4bn; €28.1bn) a year.5

Sources and search criteria

We retrieved the publications included in this review by searching PubMed (1977 to November 2016) using keywords such as “alternative medicine”, “complementary medicine”, “integrative medicine”, “naturopathy”, “controlled trial”, “randomized trial”, “systematic review”, “meta-analysis”, “observational study”, “mechanism ”, “rheumatoid arthritis”, “rheumatic disease”, “back pain”, and “neck pain”. In addition, we used keywords such as “chronic pain”, “back pain”, “neck pain”, “arthritis pain”, “acupuncture”, “analgesia”, and “yoga”. We also searched review articles, meta-analysis, and Cochrane reviews. We included only articles with strong evidence (in animal and human studies). For specific therapies identified as relevant and frequently used, we included the specific keywords in a further search—for example, “herbal medicine”, “phytomedicine”, “herbal”, “mind-body”, “yoga”, “meditation”, “relaxation”, “mindfulness”, “fasting”, “caloric restriction”, and “diet”.

We obtained additional articles by cross referencing search terms with review articles (since 2000) and searching manually through reference lists. We further searched the register of We considered observational studies when a method had a long term use, no controlled trials were available, and the result of the observational study suggested a potential clinical relevance. We further examined observational studies for adverse events. Finally, we considered preclinical and experimental original research and reviews for describing mechanisms of action. We excluded articles published in non-peer reviewed journals, case reports, and small uncontrolled studies.

For the final selection of treatment modalities, we excluded methods that are used in only few countries or regions worldwide or not commonly regarded as CIM. For example, biofeedback is not regarded as a CIM method in central Europe and is only rarely used. Within broader categories such as mind-body therapies or nutrition we focused on those treatment methods that ranked the highest regarding evidence from randomized controlled trials (RCTs).

CIM modalities

CIM includes a variety of practices, which can be divided into four major categories as listed in figure 1.7 On the basis of the NHIS data comparing 2002 and 2007, the overall use of CIM, including acupuncture, deep breathing exercises, massage therapy, meditation, naturopathy, and yoga, among adults has increased from 32.3% to 35.5%.8 In the following sections, we will summarize scientific data for potential mechanisms and clinical evidence about effectiveness and adverse effects on several commonly used CIM modalities such as acupuncture, mind-body system, dietary interventions and fasting, and herbal medicine and nutrients.


Fig 1 Categories of complementary and integrative medicine


Acupuncture has been a substantial component of the healthcare system in Asia for more than 3000 years. Over the past few decades, the popularity of acupuncture has increased in many Western nations. A major reason for patients to seek acupuncture treatment is the low incidence of adverse effects compared with many drugs and medical procedures.

Research data on mechanisms of acupuncture

Much research has focused on the “scientific” basis of acupuncture’s effects. Evidence supports the notion that the central and peripheral nervous system is involved in the process (fig 2), including neurotransmitters and neuromodulators such as endorphins, neurohumoral factors, and other chemical mediators.9


Fig 2 Mechanisms of action of various complementary and integrative medicine methods

Different methods of acupuncture stimulation elicit different mechanisms of pain inhibition. Manual acupuncture exerts a different effect from electroacupuncture on the synthesis, release, and action of neuropeptides such as opioid peptides,10 cholecystokinin-like immunoreactivity,11 and natural killer cells.12 In addition, endorphins and μ-opioid receptors in the mouse brain mediated the analgesic effect induced by 2 Hz but not 100 Hz electroacupuncture, and this effect was blocked by a μ-opioid receptor antagonist.1314

Endogenous factors

Endogenous factors mediate acupuncture analgesia by releasing substances into the cerebrospinal fluid,15 such as nerve growth factor,16 anandamide (an endogenous cannabinoid),17 and γ-aminobutyric acid (GABA, an inhibitory neurotransmitter) but not glutamate (an excitatory neurotransmitter).18 Acupuncture also downregulates phosphorylation of the N-methyl-D-aspartic acid receptor subunit NR2B 19 but upregulates the nitric oxide content.20

Other neurotransmitters involved in the effect of acupuncture include epinephrine (adrenaline), norepinephrine (noradrenaline), dopamine, and 5-hydroxytryptamine.21 Moreover, electroacupuncture inhibited stress induced increases in several neurotransmitters including norepinephrine, dopamine, and corticosterone.22 The frequency of electroacupuncture may play a role in differentiating its effect in various brainstem regions. For instance, the effect of electroacupuncture at 4 Hz was reported to be mediated by endogenous opioids,23 whereas the effect at 2 Hz involved substance P.24 Nonetheless, the effect of electroacupuncture at different frequencies (2, 10, or 100 Hz) can be at least partially blocked by a serotonin receptor antagonist, suggesting that the endogenous serotoninergic function may be a common pathway for the effect of electroacupuncture.25

Brain neural activities

Recent neuroimaging studies have shed light on the involvement of brain neural activities in acupuncture mechanisms. For example, pain signals can activate the periaqueductal gray, thalamus, hypothalamus, somatosensory cortex, and prefrontal cortex regions in humans.26 Acupuncture treatment seems to diminish the increased neural activities after achieving the so called “de-qi” sensation.2728 Moreover, different acupuncture points produce signal changes in specific regions of the central nerve system.29 The neural response to electroacupuncture can also be visualized in the rat primary somatosensory cortex by using an optical imaging system.30

Mechanisms underlying the immunosuppressive effect of acupuncture remain unclear, although studies have shown that acupuncture induces release of neuropeptides from nerve endings, such as calcitonin gene related peptide, substance P, and β-endorphin, and results in changes in endogenous mediators such as cytokines, tumor necrosis factor-α (TNF-α), and vascular endothelial growth factor.31323334 Positron emission tomography using18F-fluorodeoxyglucose has shown regional inflammatory changes after acupuncture in the knee joint of patients with rheumatoid arthritis.35

More recently, functional magnetic resonance imaging (MRI) has shown that the improvement in pain by verum (traditional) acupuncture (compared with sham) is accompanied by restoration of the balance in the connectivity of several brain regions implicated in executive control and descending pain modulation, prevention of cortical thinning, and/or altered pain related attention and memory.363738Table 1 includes animal and human studies as examples of research in this area.

Table 1

Possible mechanisms of acupuncture

View this table:

Clinical evidence for acupuncture

Systemic reviews and meta-analyses have examined the effect of acupuncture on a variety of pain conditions. Most suggest that acupuncture is significantly better than both sham acupuncture and standard care for some but not all types of chronic pain. In some cases, the effect of acupuncture can last up to six to 12 months. Table 2 includes several meta-analysis and systemic reviews related to back pain, neck pain, and arthritis pain.3940414243444546474849505152

Table 2

Effect of acupuncture on chronic back, neck, and rheumatoid arthritis pain

View this table:

Rheumatoid arthritis

Rheumatoid arthritis is a common form of inflammatory joint disease. Patients with rheumatoid arthritis often need a prolonged course of treatment with drugs such as non-steroidal anti-inflammatory drugs (NSAIDs) and immunomodulatory agents (disease modifying anti-rheumatic drugs (DMARDs)) to treat pain, swelling, joint stiffness, and loss of joint function. However, these treatments are often associated with significant side effects and risks. The prevalence of use of acupuncture, medicinal herbs, and other CIM approaches ranges from 28% to 90%, and the results are promising.53545556 on the basis of several reviews and data analyses highlighted in the following sections, acupuncture can improve several rheumatoid arthritis related symptoms and signs including pain, joint mobility, fatigue, depression, and sleeplessness, which leads to improvements in patients’ lifestyle and emotional wellbeing.57585960

Several RCTs have shown that acupuncture is as effective as drugs with no major adverse events in improving disease activity score (DAS28) components including pain, swollen joint count, health related quality of life (SF-36), rheumatoid factor, erythrocyte sedimentation rate, and C reactive protein.616263 In addition, the onset of the effects of auricular electroacupuncture was earlier during the second treatment week, and the most significant improvement occurred at the end of a seven week treatment period. In this RCT, 44 rheumatoid arthritis patients were randomized to an electroacupuncture group or an autogenic training group to receive treatment once a week for six weeks. Pain intensity and the disease activity score 28 (DAS28) were the primary outcomes. Use of pain medication, pain disability index, clinical global impression, and pro-inflammatory cytokine concentrations were secondary outcomes. Both groups showed a significant improvement at the end of the treatment and at a three month follow-up compared with baseline (P<0.05). The electroacupuncture group reported significantly less pain (P=0.040) and overall improvement (P=0.035) than the autogenic group. The erythrocyte sedimentation rate was also significantly reduced (P=0.01).64

An RCT of 146 patients with rheumatoid arthritis compared electroacupuncture plus meloxicam, sulfasalazine, and methotrexate with drug treatment only and found that the combination produced a better therapeutic effect than the drugs only group (effectiveness rate was 79.7% v 51.4%; P<0.05).65 These findings are consistent with those of another RCT of 80 patients with chronic pain due to rheumatoid arthritis, in which the effect of electroacupuncture, traditional Chinese acupuncture, or sham acupuncture was compared. All patients received 20 sessions of treatment over a period of 10 weeks. The results showed a reduction in the number of tender joints and physician’s global score following electroacupuncture and traditional Chinese acupuncture but not sham acupuncture.66

A practical question about the effect of acupuncture on rheumatoid arthritis symptoms is whether different methods of acupuncture have different effects on clinical outcomes.

Electroacupuncture—An RCT, 63 patients with rheumatoid arthritis compared electroacupuncture using continuous wave for a total of 45 minutes against manual acupuncture at the same acupoints for the same duration. Both groups received the treatment once every other day for 20 days as one course. After three courses, compared with traditional acupuncture, electroacupuncture significantly modulated peripheral (blood, joint fluid) interleukin (IL-1, IL-4, IL-6, and IL-10) concentrations.67 An RCT of 90 patients with rheumatoid arthritis compared the combination of electroacupuncture with Chinese herb iontophoresis plus drug treatment (meloxicam, salazosulfamide, methotrexate) against drugs only. After one month, the combination significantly improved duration of morning stiffness, average grasp strength of both hands, and joint tenderness and swelling (P<0.01). In addition, several laboratory indicators of rheumatoid arthritis were significantly decreased in the treatment group (P<0.01).68 However, an RCT comparing acupuncture needle twisting with electroacupuncture frequency in 33 patients with rheumatoid arthritis found no differences in their analgesic effects.69 In another RCT, 50 patients with rheumatoid arthritis were randomized to a treatment group with the needle sticking method or a control group with routine filiform needle therapy. The acupuncture needle sticking method reduced the number of pressure pain joints after two therapeutic courses (P<0.01).70

Moxibustion refers to burning dried herbs mugwort (moxa) on acupuncture points with or without using acupuncture needles, which is increasingly used for rheumatic conditions.71 A meta-analysis of eight RCTs (n=631) found that moxibustion reduced the rate of symptoms compared with conventional drug therapy (relative risk 1.13, 95% confidence interval 1.02 to 1.26; P=0.02). Meta-analysis of an additional six RCTs also suggested that moxibustion plus drug treatment reduced the rate of symptoms compared with conventional drug treatment alone (n=433; relative risk 1.25, 1.09 to 1.43; P=0.02).72

Bee venom acupuncture falls into the category of herbal acupuncture, which combines the effect of bioactive compounds isolated from bee venom with acupuncture stimulation. Several meta-analyses found limited evidence assessing its effectiveness. Pain was lowered significantly more with bee venom acupuncture than saline acupuncture (n=112; weighted mean difference on 100 mm visual analog scale 14.0, 95% confidence interval 9.5 to 18.6, mm; P<0.001). However, the limited number, low quality, and small size of the RCTs make drawing conclusions difficult.737475

Summary—Although several studies described above have indicated a positive role for acupuncture in the treatment of rheumatoid arthritis, other studies have failed to show positive outcomes.76777879 A survey study of 350 patients in rheumatology clinics with face-to-face structured interviews on self perceived efficacy of different types of complementary alternative medicine found that satisfaction was lowest among complementary medicine users with rheumatoid arthritis, vasculitis, or connective tissue diseases.80 This result seems to be consistent with the findings reported in several review articles.818283 The discrepancy among studies may be related to methodological factors such as the type of acupuncture (acupuncture v electroacupuncture), site of intervention, and sample size differences.8485 Overall, the effectiveness of acupuncture in the treatment of rheumatoid arthritis needs to be further determined in large RCTs.8687

Chronic low back pain

Chronic low back pain (CLBP) and chronic neck pain (CNP) are common clinical pain conditions that are associated with high medical financial cost and/or loss of productivity. Acupuncture is one of the most frequently used tool CIM treatments for CLBP and CNP. Acupuncture has a low rate of adverse effects (0.13-0.14%) and is cost effective as assessed using the WHO standard of cost per quality adjusted life year.88 A large RCT in 11 630 patients with CLBP randomized 1549 to an acupuncture group and 1544 to a control group, and 8537 were included in a non-randomized acupuncture group. All patients were allowed to continue routine medical care during the study. At three months, back function was improved from 12.1 (standard error 0.4) to 74.5 (0.4) points in the acupuncture group and from by 2.7 (0.4) to 65.1 (0.4) points among controls (difference 9.4, 95% confidence interval 8.3 to 10.5, points; P<0.001).88 The clinical practice guideline provided by the American College of Physicians and the American Pain Society recommends acupuncture as a non-drug therapy for CLBP patients.8990

Several studies support the use of acupuncture for the treatment of CLBP. An RCT in 131 patients with CLBP compared acupuncture, sham acupuncture, and a control group who received neither. Each group received 20 sessions of treatment plus active physiotherapy for 12 weeks. Acupuncture plus physical therapy was superior to physical therapy alone in reducing pain intensity (P<0.001), pain related disability (P<0.001), and psychological distress (P=0.020). The same trial compared acupuncture with sham acupuncture and found that acupuncture significantly reduced psychological stress (P=0.040). However, the effect of acupuncture was not detectable at the nine month follow-up.91 In another RCT, 143 patients were randomized into receiving acupuncture twice a week for three weeks (n=74) versus control (n=69). All patients participated in a standardized 21 day inpatient rehabilitation program. CLBP patients treated with acupuncture reported significant improvement in pain, emotion, functioning, and work days.92 A multicenter RCT randomized 130 patients to individualized real acupuncture treatments or sham acupuncture treatments twice a week for six weeks. It found that, compared with sham acupuncture, real acupuncture treatment significantly improved the pain intensity VAS score, Oswestry Disability Index, Beck Depression Inventory, and SF-36 (P<0.05). The effect remained at a three month follow-up.93 Two large RCTs (n=1162 and n=3093) reported that the benefit of acupuncture treatment lasted up to six months.889495

In addition to traditional acupuncture, several modified acupuncture methods have been used in the treatment of CLBP, including scalp acupuncture, laser acupuncture, auricular acupuncture, Hegu acupuncture, motion style acupuncture, and electroacupuncture of the spinal nerve root.96979899100101 Acupuncture, both manual and electroacupuncture, showed significant effects on pain reduction compared with a control group in an RCT (n=50) of acupuncture once a week for eight weeks. The decrease in pain intensity remained at three and six month follow-up assessments (P<0.05).102

Other considerations in acupuncture therapy include the method of acupuncture needling, selection of acupuncture points, duration of treatment, and manual versus electric stimulations. These factors could substantially influence clinical outcomes. An RCT (n=75) assessed the effects of different durations of acupuncture. Treatment (0, 15, 30, and 45 minutes) was delivered in a random sequence in each session over 11 weeks. Both 30 and 45 minute sessions were more effective than a 15 minute session. However, the 45 minute session did not further improve the outcome compared with the 30 minute session. Similarly, 30 minutes and 45 minutes of electroacupuncture using percutaneously placed needles produced similar improvements in pain, physical activity, quality of sleep, and analgesic requirement, and both treatment durations were better than 0 (no treatment) or 15 minute durations.103

An analysis of 53 studies found that selection of acupuncture points is also critical, as each set of acupuncture points reflects acupuncture point networks necessary for the treatment of CLBP.104 A meta-analysis of 13 RCTs with 2678 people found that acupuncture significantly reduced pain and improved disability status and quality of life, although the influence from non-specific needling cannot be ruled out.105 Two more systemic reviews (seven and 25 RCTs) found that acupuncture reduced pain score and improved functional status compared with sham acupuncture or usual care.106107

Current evidence, including recent data from the meta-analyses, reviews, and RCTs described above, suggests that acupuncture may be a good option for treatment of chronic back pain. However, the validity of current acupuncture studies needs to be interpreted in the context of study design and other methodological considerations.

Chronic neck and shoulder pain

In two meta-analyses of 10-14 clinical trials, moderate evidence showed that acupuncture was more effective for pain relief than were sham controls or inactive sham treatments (pooled standardized mean differences −0.37, 95% confidence interval −0.61 to −0.12),108109 and acupuncture treatment also seems to be cost effective in these settings.110

In a multicenter trial of more than three months with 14 161 patients, of whom nearly 4000 were randomized into an acupuncture group plus routine care and the remainder to a control group (routine care), significant improvement in neck pain and disability was observed in the acupuncture group.111 Compared with massage therapy, two RCTs (n=177) showed that acupuncture treatment (five sessions over three weeks) significantly improved pain and range of motion (P=0.052; confidence interval 16.5 to 31.9), and the effect was greater in a subgroup of patients with neck pain for more than five years.112113 An RCT (n=135) with eight treatments over four weeks found no significant difference between the acupuncture group and sham control group.114 Another RCT (n=115) compared the effect of drugs, acupuncture, and chiropractic manipulation. Acupuncture reduced visual analog scale (VAS) score significantly more than chiropractic manipulation (P<0.01), although chiropractic manipulation showed better improvement (P<0.01) in other endpoints (neck disability index, SF-36, Oswestry Back Pain Disability Index).115

Acupuncture reduced chronic neck and shoulder pain for at least three years with a concomitant improvement in depression, anxiety, sleep quality, pain related activity impairment, and quality of life in two small RCTs (n=24 in both studies) with 10 acupuncture treatments over a four week period.116117 An RCT (n=46) found that the combination of acupuncture with physical therapy was superior to acupuncture (one or two sessions a week for 10 weeks) or physical therapy alone for improvement in neck and shoulder pain for up to six months. Although all three groups showed significant improvement in pain (P<0.001), the physical therapy plus acupuncture group was superior to physical therapy (P<0.05) or acupuncture alone (P<0.01).118

Adverse effects of acupuncture therapy

The incidence of complications from acupuncture treatment are significantly lower than with many medical treatments.119 The most commonly reported complications are bruising or bleeding at the needle insertion site and a transient vasovagal response. Other rare complications include infection, dermatitis, and broken needle fragments. In two large scale surveys including 34 407 and 31 822 acupuncture treatments, no serious adverse events were reported that caused unexpected hospital stays, permanent disablement, or death. Minor adverse events included nausea, pain, bruising, fainting, aggravation of symptoms, and psychological and emotional reactions, and the rate of these events was 0.13-0.14%. Most adverse events can be prevented by not leaving patients unattended during treatment, ensuring that needles are removed after treatment, and avoiding areas of cellulitis.120121 Although rare (6/97 733 patients in a meta-analysis), serious complications may occur with acupuncture, including pneumothorax, hemothorax, internal organ puncture, and pericardial effusion.122 Septic arthritis also has been reported as a rare complication of acupuncture.123

Mind-body medicine

According to a definition of mind-body medicine by the National Institutes of Health,124 mind and body practices focus on the interactions among the brain, mind, body, and behavior, with the intent to use the mind to affect physical functioning and promote health. In clinical practice, the essential and most frequently used part of mind-body medicine focuses on techniques such as meditation, yoga, or tai chi. These and other techniques share the elicitation of the physiologically defined relaxation response and intend to counterbalance the detrimental effect of increased uncontrolled emotion or stress related to disease or pain.125

Besides pure meditative or concentrative techniques including various traditional forms of meditation, mindfulness, autogenic training, breathing techniques, and imagery; meditative movement therapies such as yoga, qigong, and tai chi; and some further modern somatic therapeutic techniques such as progressive muscle relaxation, the Feldenkrais method and the Alexander technique are often used.126 In the long term, most of these techniques can be applied as a self care therapy.127

Generally, the concept of mindfulness has an increasing role in mind-body medicine. Mindfulness practice and training involve the cultivation of a non-judging mental attention toward unwanted thoughts, feelings, or disease related bodily experiences by means of mindfulness meditation and non-judging body observation.128 Most studies used the standardized eight week program of mindfulness based stress reduction (MBSR), which focuses on the practice of mindfulness meditation.129 Mindfulness may help to ameliorate both psychological and physical symptoms in patients with various chronic disease states. Besides mindfulness meditation, other techniques such as transcendental meditation, zen meditation, or jyoti meditation have been used and investigated for treating pain in clinical research. A large number of studies have examined yoga in chronic pain conditions. Yoga, in its original concept within its Indian tradition, comprises meditation, breathing exercises (pranayama), lifestyle advice, and body positions (asanas).130 In the Western version and in most clinical studies, the focus is on the practice of asanas that incorporates muscle strengthening and stretching combined with relaxation and breathing control.

Research data on mechanisms of mind-body practices

The relaxation response leads to a variety of physiologic beneficial effects that may enhance pain relief through reduced sympathetic activity, decreased muscular tension, modulated pain awareness, and an increased release of endogenous opioids.131132 Neuroimaging studies have described neurobiologic effects of mind-body techniques on brain neural activities and pain signaling relevant for chronic pain conditions (fig 2).132

Most research has used meditation techniques such as mindfulness meditation, transcendental meditation, or other yogic meditation techniques. Some studies show that people who meditate have thicker cortices in frontal regions, including the prefrontal cortex, the cingulate cortex, and the insula.133 Meditation modulates the contextual evaluation of pain and lowers pain sensitivity,134135136137 but it is likely to do so dynamically over time and experience.138

Functional MRIs of the response to thermally induced pain applied outside the meditation period found that long term meditators (transcendental meditation) showed 40-50% less brain response to pain (which is paralleled by less subjective pain on a VAS) than controls. After the controls started to practice meditation for five months, their pain response decreased as well.139 Movement based contemplative practices lead to similar changes, as they also act via the relaxation response and some of the mechanisms that apply to meditation. In addition, muscle strengthening, muscle relaxation, stretching (eg, yoga), coordination (eg, tai chi and qigong), and improved awareness of inappropriate musculoskeletal patterns of use and their correction (eg, Alexander technique, Feldenkrais, yoga) may contribute to pain relief.

Clinical evidence for mind-body medicine

Rheumatoid arthritis

MBSR—Only few trials have investigated the effects of mind and body practices in rheumatoid arthritis. Three RCTs found that mindfulness meditation alleviated psychological distress and improved wellbeing in patients with rheumatoid arthritis.140141142 The first RCT in 63 patients compared the MBSR program with a waiting list control and found improvement in psychological distress and wellbeing with MBSR but no significant difference in disease activity after two and six months.141 Adverse events were not reported. The second RCT in 51 patients with moderate disease activity found that the MBSR program significantly reduced the disease activity score (DAS28) and pain scores after two, four, and six months compared with people on a waiting list. No mean differences for outcomes and no values for pain scores were described, and adverse events were not reported.140 The third RCT included 73 patients with different inflammatory joint diseases and compared an intensive mindfulness based group intervention against routine care plus a CD instructing mindfulness based home exercise.142 After 12 months, the MBSR group showed significant benefits on various outcomes of emotional wellbeing, psychological distress, and fatigue but only a slight non-significant effect on pain (mean difference on VAS −0.6, 95% confidence interval −1.28 to 0.02). Adverse events were not reported. Although increasing evidence links the practice of mindfulness techniques to improved immune function,143 larger studies with higher methodological quality are needed to examine long term effects in rheumatic disease.

Yoga—Two studies have examined the effect of yoga in rheumatoid arthritis.144145 One small randomized study of 30 young women with confirmed rheumatoid arthritis compared a six week Iyengar yoga intervention that was positively tested in a pilot study,146 including yoga postures and relaxation for 90 minutes twice weekly, with usual care.144 The study found some significant beneficial effect on pain acceptance, pain disability, and selected items of quality of life but not on pain intensity. Mean differences were not reported. No adverse events related to yoga were reported. A second randomized trial with low methodological quality (according to GRADE criteria) included 80 patients with rheumatoid arthritis and used a multimodal yoga intervention with postures, breathing technique, meditation, and lifestyle advice for 90 minutes six times a week for seven weeks. Compared with usual care, yoga significantly reduced pain intensity.145 However, owing to the high risk of bias in both studies, only a weak recommendation for yoga in rheumatoid arthritis can be given at this time.147

Chronic low back pain

MBSR—A systematic review on MBSR in chronic low back pain published in 2014 found three RCTs that included 117 patients.148 It found inconclusive evidence for pain reduction through mindfulness meditation but moderate evidence of improved pain acceptance. However, two of the three studies investigated low back pain in older adults (>65 years), which limits generalizability to the commonly younger population affected by chronic back pain. The largest study so far investigating MBSR in low back pain was published in 2016 and included 342 patients (mean age 49 years) who were randomized to the MBSR program, time equivalent cognitive behavioral therapy (CBT), or usual care. MBSR and CBT resulted in significantly greater improvement in back pain and functional limitations at 26 and 52 weeks compared with usual care.149 Clinically meaningful improvement in disability and function was achieved in 60.5% with MBSR, 57.7% with CBT, and 44.1% with usual care (relative risk for MBSR versus usual care 1.37, 95% confidence interval 1.06 to 1.77). These benefits were achieved even though only about 50% of patients attended at least six of the eight treatment sessions.

Yoga—Yoga is one of the most popular CIM approaches in the management of back pain. Twelve RCTs and six systematic reviews of yoga for non-specific back pain have been published.150 The studies used mainly hatha yoga, iyengar yoga, or vini yoga and 12 week treatment periods. Overall, strong evidence exists that yoga significantly reduces pain and improves back related disability, with effect sizes mostly varying from 0.4 to 0.7 in the short term but only moderate evidence for long term effects. A more recent systematic review on yoga and chronic back pain analyzed nine trials with 810 participants that compared yoga with non-exercise and found a moderate improvement in back related function at three or four months (standardized mean difference (SMD) −0.40, 95% confidence interval −0.66 to −0.14) and slight improvement for pain on the 100 mm VAS (mean difference −4.55, −7.04 to −2.06) after seven months. When yoga was compared with exercise (four trials, 394 patients), there was no difference in function at three to 12 months but evidence for reduced pain (mean difference on 100 mm VAS −20.40, −25.48 to −15.23). Adverse effects of yoga and exercise were comparable.151 Thus yoga may have a role in the management of back pain for people who are willing to practice it, but longer term outcomes need to be evaluated.

Tai chi and qigong—Studies indicate a large treatment effect of tai chi for fibromyalgia,152 and limited evidence supports a role for tai chi or qigong in the treatment of chronic back pain. In one of two available RCTs, 160 patients with a mean age of 44 years were randomized to 18 sessions of tai chi over 10 weeks or to a waiting list control. Tai chi practice improved function (mean difference 2.6, 1.1 to 3.7) and reduced pain intensity (−1.3., −0.7 to −1.9) to a mild extent. There were no relevant adverse events.153 The design of the second RCT was unclear, so it is not discussed here.154 A subsequent meta-analysis analyzing the two available studies on tai chi in back pain found an SMD of −0.84 (−1.27 to −0.42) compared with no treatment.155 A non-inferiority RCT in 128 patients with chronic back pain comparing a three month intervention of 12 sessions of qigong or exercise found comparable pain relief in the 12 month study period (VAS group difference at three months –1.8 (–9.4 to 5.8).156 A recent RCT of 176 older adults (mean age 73 years) with chronic back pain assessed pain intensity by the Functional Rating Index after three months. Mean pain intensity after three months was 1.71 (1.54 to 1.89) for yoga, 1.67 (1.45 to 1.89) for qigong, and 1.89 (1.67 to 2.11) for no intervention with no significant difference between groups.157 Thus, tai chi/qigong may be more effective in younger patients with back pain. Neither tai chi nor qigong have been related to relevant adverse events in studies so far.

Alexander technique—The effectiveness of Alexander technique in low back pain was examined in a large randomized study with 579 patients. After randomization, patients were allocated to six or 24 Alexander technique lessons, massage, or usual care. In addition, in each group half of the patients received an exercise prescription. Exercise plus Alexander technique remained effective after one year, with 24 lessons of Alexander therapy having the most pronounced effect on function (mean difference −3.40, −4.76 to −2.03) and pain (von Korff pain −1.30, −1.93 to −0.67) compared with massage and usual care controls. Six lessons of Alexander technique followed by exercise were nearly as effective as 24 lessons, suggesting that a combined approach is useful and cost effective.158 There were no serious adverse events with Alexander technique or exercise.

Chronic neck pain

Neck pain is associated with high levels of perceived stress,159 so mind and body practices would be expected to be a useful treatment approach. However, only a few studies have evaluated meditation in chronic neck pain.

Meditation—A randomized study of focused meditation in 64 patients with chronic neck pain showed reduced pain (mean difference −13.2, −2.1 to −24.2) and pain related discomfort but no effect on disability after eight weeks compared with self care exercise. No adverse effects were associated with meditation.160 These results are in line with findings from neuroimaging showing that meditation can modify pain perception but may not have somatic effects on the cause of pain.139

Yoga—The effectiveness of yoga in chronic neck pain was examined in two randomized high quality trials and with 77 and 51 patients comparing yoga with self care exercise over nine weeks. Both studies found significant pain relief with yoga compared with self care exercise (mean difference −20.1, −30.0 to −10.1, and −13.9, −26.2 to −1.4).161162 There were no relevant adverse events in either study. Thus, yoga seems to be a viable option in chronic neck pain, although studies with longer term outcomes are needed.

Tai chi—An RCT in 2016 compared 12 weekly sessions of 75 to 90 minutes of tai chi or neck exercises with a waiting list control in 114 patients with chronic neck pain. After 12 weeks, tai chi participants experienced less pain than the wait list group (mean difference VAS −10.5, −20.3 to −0.9). Group differences were also found for functional disability and quality of life compared with the waiting list group, but no differences were found for tai chi compared with neck exercises.163 Tai chi might be considered a suitable alternative to standard exercise or yoga for patients with a preference for tai chi.

Qigong—In contrast, findings for qigong in chronic neck pain are inconsistent.164 A three armed trial in 121 older patients found no significant differences between a qigong group and a waiting list control group (VAS mean difference −11, −24.0 to 2.1, mm) or between the qigong group and the exercise therapy group (−2.5, −15.4 to 10.3, mm).165 In a further three armed randomized trial in 123 mostly middle aged women, qigong but not exercise was slightly superior to no intervention with regard to reduction of pain and disability after six months (VAS mean difference −14 mm, −23.1 to −5.4 mm).166 Thus, the existing evidence is not convincing to support the regular use of qigong in chronic neck pain.

Alexander technique and Feldenkrais method—A large RCT compared the effectiveness of Alexander technique, acupuncture, or usual care in 517 patients with chronic neck pain.167 Ten acupuncture sessions or 14 Alexander technique lessons both led to comparable significant reductions, with between group reductions in the Northwick Park Pain Questionnaire score of 3.92 (0.97 to 6.7) percentage points for acupuncture and 3.79 (0.91 to 6.66) percentage points for Alexander lessons at 12 months compared with usual care. However, a smaller RCT found that Alexander technique was not superior to local heat application in treating chronic non-specific neck pain.168 To date, only one poor quality study, which included 61 patients with visual impairment, has investigated the Feldenkrais method. Patients undergoing the Feldenkrais method reported less pain after 12 weeks and one year compared with no treatment. Mean differences and adverse events were not reported.169

Adverse effects of mind-body therapy

Adverse effects of mind-body practices were documented in numerous trials. Collectively, these trials found that meditation, yoga, tai chi, qigong, and Alexander technique for chronic pain conditions are relatively safe. Most of the adverse events related to mind-body practices are self limited and mild. Findings from a recent review of published studies indicate that yoga is as safe as common exercise.170 However, the safety of these techniques outside the context of clinical trials is unknown. Meditation may be rarely associated with short term symptoms of anxiety in the initial phase of practice.171

Dietary interventions and fasting

Research data on mechanisms of diet and fasting

Disease specific dietary recommendations have an important role in the clinical practice of CIM in rheumatic diseases. Patients with rheumatoid arthritis often report associations between disease activity, pain, and diet.172173 A plant based diet may decrease body inflammation and pain by reducing concentrations of exogenous arachidonic acid, which is claimed to decrease production of arachidonic acid derived pro-inflammatory eicosanoids such as leukotriene B4 and prostaglandin E2.174 Animal derived proteins such as methionine and cysteine might further induce body inflammation through distinct mechanisms,175 and a plant based diet thus may beneficially affect rheumatic inflammation. Furthermore, a plant based diet may decrease inflammation by modulating the intestinal microbiota and by the increased intake of antioxidants that are naturally occurring in plants.176 Moreover, fasting, intermittent fasting, and caloric restriction have shown pronounced anti-inflammatory effects in experimental research.175177 Fasting induced mood enhancement is a well known phenomenon,178 and the psychological effect together with reported mechanisms such as increased central serotonin availability and enhanced endogenous opioid release as well as peripheral antinociceptive effects may add to total pain relief (fig 2).179180 Furthermore, fasting downregulates the NLRP3 inflammasome and decreases TNF-α.181

Clinical evidence for diet and fasting in rheumatoid arthritis

Mediterranean and plant based diets

The Mediterranean diet, which is high in plant foods, olive oil, and nuts and low in red meat, has been investigated in rheumatoid arthritis in two clinical trials. The first was an RCT in 56 patients with rheumatoid arthritis comparing a Mediterranean diet with patients’ normal diet. After 12 weeks, pain intensity was significantly reduced in the Mediterranean diet group (mean difference −14.0, −23.6 to −4.4). The Mediterranean diet group also had a significantly larger reduction in disease activity.182 A larger non-randomized controlled trial included 130 female patients with rheumatoid arthritis. After six months, the median pain remained unchanged in the Mediterranean diet group whereas the ordinary diet group had an increase from median 55 to 63 on the pain VAS (P=0.049). Quality of life and global wellbeing were significantly improved with the Mediterranean diet. Mean differences of effects were not reported.183 Neither study reported adverse effects.

The effect of a low fat strict vegan diet has been tested only in a small uncontrolled study in 24 patients with mild to moderate rheumatoid arthritis. After four weeks, an improvement in rheumatic symptoms and a reduction in pain from a mean VAS of 49 (SD 20) to 34 20 were found, and the diet was not associated with adverse events.184

The Mediterranean diet is well established in the secondary prevention of coronary disease and the metabolic syndrome. As patients with rheumatoid arthritis have an increased cardiovascular risk, a Mediterranean diet can clearly be recommended


Empirical and observational data indicate a role for fasting therapy in the management of rheumatoid arthritis.185 A systematic review identified four randomized trials of fasting in rheumatoid arthritis and reported an effect size for pain relief of d=0.58 (P<0.01).186 The largest and highest quality RCT included in the meta-analysis, in 53 patients with rheumatoid arthritis, compared an experimental group who fasted for seven to 10 days followed by a vegan gluten-free and sugar-free diet for three months and thereafter a lacto-vegetarian diet for nine months with a control group who continued their usual diet. Food items that led to return of pain or rheumatic symptoms were eliminated on an individual basis. Fasting followed by this vegetarian diet regimen led to significant improvement in all clinical variables including pain (effect size d=0.67; P<0.01) and tender and swollen joints up to the final follow-up at 13 months.187 Adverse events were not reported.

Little is known about the long term effects of fasting and its effect on disease progression and long term outcome. Taken together, preliminary evidence exists for symptomatic benefits of fasting in rheumatoid arthritis. However, newer trials with higher methodological quality and patients treated with current drugs are needed.

Elimination diet

Among a broad variety of food and nutrients that are associated with increases in disease activity, meat, milk and dairy products, wheat gluten, citrus fruit, alcohol, and coffee are ranking high.172173 Doctors may encourage patients to find individual associations and try individual elimination after confirmation by re-challenging. Clinical research in elimination diets is difficult and complex to perform. So far, only one RCT with 53 patients has examined the concept of an elimination diet with a complex study design, including an initial washout period and specific responder analyses.188 Foods least likely to cause intolerance were reintroduced first in a stepwise manner, and any foods inducing symptoms of rheumatoid arthritis were removed from the diet. There was a significant reduction in pain with the diet in both groups during the dietary elimination phase, but differences between groups were not reported.


Patients with rheumatoid arthritis may benefit and experience pain relief from fasting, a plant based diet, a Mediterranean diet, and possibly an elimination diet. However, much of the evidence was produced before the current era of intensified DMARD therapy including biologic drugs. The effect of diet in the setting of newer potent drug therapies remains unclear and needs to be tested in future randomized trials.

Clinical evidence in back and neck pain

The association between obesity and low back pain is established, and increased adiposity is associated with higher levels of back pain intensity and disability.189190 In an uncontrolled pilot study of 46 obese patients (mean body mass index 45), a multidisciplinary caloric restriction and weight loss program over 26 weeks led to significant weight loss and reduced pain (P<0.001) and improved function after 52 weeks.191 There are no further data from specific nutritional and dietary interventions, so whether specific diets affect chronic back and neck pain independently of weight loss is unclear.

Adverse effects of diet and fasting

The Mediterranean diet and vegetarian diets are not associated with adverse effects. Patients on vegan diets should be monitored for vitamin B12 deficiency. Periods of fasting up to 10 days are not associated with severe adverse effects unless contraindications such as underweight, previous weight loss, ulcer disease, or eating disorder exist. However, fasting should be supervised by trained therapists.

Herbal medicine and nutrients

Research data on mechanisms of herbal medicine and nutrients

Preclinical research on herbs is conducted according to standard pharmacological research and focuses on the specific pharmacological mechanisms of the leading substances in plant extracts. Herbal medicine is popular in Europe and a cornerstone in most of the traditional medical systems worldwide. Mechanisms of nutrients are less intensively examined, and more research is needed on their pharmacological actions.

Clinical evidence for herbal medicine and nutrients in rheumatoid arthritis

Despite their longstanding use, only limited data from high quality trials on the use of herbal medicine and specific plant based nutrients in rheumatic pain conditions are available.

Borago officinalis

Borago officinalis (borage seed oil) containing γ-linolenic acid (GLA) has been tested in two small placebo controlled RCTs over six months. One included 37 patients with rheumatoid arthritis using 1.4 g GLA, and the other included 65 patients using 2.8.g GLA daily from borage seed oil.192193 Both studies found significant improvements in clinical parameters of rheumatoid arthritis including pain intensity (P<0.001 each) after six months.

Nigella sativa

Nigella sativa (blackcurrant seed oil) and Oenothera biennis (evening primrose oil) are further rich sources of GLA. Trials on these GLA sources used smaller daily doses (<1.4 g/day) of GLA or did not report outcomes for pain.194 Overall effects on rheumatoid arthritis were inconsistent, and rates of withdrawal were high owing to the number of large capsules needed for intake. A systematic review on herbal medicine in people with RA, comparing higher dose GLA (≥1.4 g daily) with placebo (including the two studies on borage seed oil), found significant improvements in pain (VAS) compared with placebo (mean difference −32.8, −56.2 to −9.4) and a non-significant increase in adverse events.194

Tripterygium Wilfordii Hook F

Tripterygium Wilfordii Hook F (TWH; thunder god vine) is traditionally used in China for the treatment of a broad spectrum of autoimmune and inflammatory diseases such as rheumatoid arthritis and ankylosing spondylitis. In preclinical studies, triptolides were the major components accounting for the anti-inflammatory effect. An RCT including 121 patients with rheumatoid arthritis compared 2 g of sulfasalazine with 180 mg of a specific TWH extract. Outcomes were available for only 62 patients after 24 weeks. TWH was superior to sulfasalazine in terms of American College of Rheumatology (ACR) 20% clinical response (65% v 32.8%; P=0.001) and pain relief (mean differences not given).195 Adverse events were more frequent with sulfasalazine.

A further three armed RCT compared methotrexate (12.5 mg/week orally) with TWH (60 mg daily) and the combination of TWH and methotrexate in a multicenter trial of more than 200 rheumatoid arthritis patients. A 50% clinical response after 24 weeks was achieved in 46% with methotrexate, 55% with TWH, and 77% with the combination therapy (P=0.0014 for non-inferiority).196 Adverse events were balanced between groups, with the exception of more irregular menstruation with TWH.

A recent meta-analysis that included all available studies on TWH concluded that TWH could be as effective as synthetic DMARDs in the treatment of rheumatoid arthritis.197 However, the efficacy of TWH has to be confirmed with better designed RCTs. Furthermore, there is major concern about the safety of TWH.198 The herb has a profound toxic potential if not extracted properly. Studies outside the field of rheumatic disease have noted an adverse effect on male fertility in both animals and humans and dysmenorrhea and amenorrhea in women.


Curcuma (turmeric) has shown profound anti-inflammatory effects in preclinical studies, with curcumin being the leading substance. The effectiveness of curcumin (500 mg twice daily) was compared with diclofenac (50 mg twice daily) or combination therapy in a small pilot clinical trial in 45 patients with rheumatoid arthritis. No differences were reported between the groups after eight weeks.199 A recent meta-analysis on turmeric extracts including three studies reporting pain in arthritis concluded that eight to 12 weeks of treatment with standardized turmeric extracts can reduce pain due to arthritis compared with placebo (VAS SMD −2.04, −2.85 to −1.24). Adverse events were not increased with turmeric.200

Omega-3 fatty acids

Omega-3 fatty acids, such as fish oil, are often used by patients as an additive therapy for rheumatoid arthritis. A meta-analysis of 17 studies with a total of 823 patients investigating the role of omega-3 fatty acid supplements in patients with rheumatoid arthritis showed modest benefits of fish oil in reducing pain (SMD −0.26, −0.49 to −0.10), duration of morning stiffness, and the number of tender or swollen joints.201 Furthermore, intake of fish oil significantly reduced the consumption of NSAIDs. Thus, current data support a moderate beneficial effect of omega-3 fatty acids on disease activity and pain in rheumatoid arthritis.

A plant derived source of omega-3 fatty acids is flax seed (Linum usitatissimum). The effect of flax seed oil (30 g/day, containing 30% α-linolenic acid) was tested for a period of three months in a small randomized trial in 22 patients with rheumatoid arthritis, but no specific effect could be found at the end of the treatment period.202

Salix alba

Salix alba (willow bark) extract has been used for thousands of years as an anti-inflammatory and analgesic remedy. Salix alba is mostly used in standardized preparations containing 120 or 240 mg salicin. Only one small randomized study has compared 240 mg salicin daily with placebo in 26 patients with rheumatoid arthritis. There was no difference in pain after six weeks.203

Rosa canina

Rosa canina (rose hip) contains a high amount of vitamin C and galactolipids, which are claimed to have anti-inflammatory properties. In a placebo controlled RCT in 89 patients with rheumatoid arthritis, intake of 5 g standardized rose hip extract significantly improved quality of life (P=0.032) and disease activity (DSA28; P=0.056) but not pain intensity. Between group differences were not reported.204

Boswellic acid

Boswellic acid (frankincense, olibanum) is the resin from trees of the genus Boswellia, native to the Arabian Peninsula and India. Traditionally, it is used against inflammatory and rheumatic diseases. In the only randomized placebo controlled study to date in 78 patients intake of 3600 mg of a standardized extract had no significant effect on rheumatic disease activity and pain. Between group differences were not reported.205


Overall, no strong evidence shows that herbal medicine reduces pain in rheumatoid arthritis. However, borage seed oil and curcumin/turmeric have slightly more evidence and may be tried in treatment resistant rheumatoid arthritis.

Clinical evidence in chronic back and neck pain

A systematic review identified five trials on Capsicum frutescens. In three trials of moderate quality, capsicum cream or plaster was compared with placebo in a total of 755 patients with chronic back pain. All three trials found significant differences favoring capsicum. Only minor adverse events were noted.206207Harpagophytum procumbens (devil’s claw) is popular as herbal remedy for non-specific pain. The systematic review identified three RCTs that found that harpagophytum reduced pain more than placebo in the short term, but poor methodological quality precluded firm recommendations being made.

Evidence that herbal medicine reduces pain in chronic back and neck pain is weak. Herbal medicine and nutritional supplements have only minimal adverse effects when adequately dosed and are well tolerated. However, as in the case of TWH, some herbal preparations that have more profound pharmacologic effects also show increased risk of serious adverse effects and can have important pharmacologic interactions with standard drugs.208

Other traditional treatments and whole medical systems

Besides Chinese medicine, further whole medical systems exist worldwide. Ayurveda or Ayurvedic medicine is the traditional medicine of India and south Asia. It comprises different therapeutic modalities such as nutritional therapy, herbal medicine, specific massages, external oil application, and lifestyle advice. Ayurveda has regained its popularity in its native countries and has become popular and more accessible recently in Europe and the US.209

Other medical systems gaining popularity include Japanese Kampo medicine, traditional Korean medicine, traditional Arabian medicine, and anthroposophical medicine. Cupping of the skin and subcutaneous tissue is a traditional treatment method mainly applied in pain and has great importance in various whole medical systems, such as traditional Chinese, Ayurvedic, and traditional Arabian and European medicine. Each of the various cupping techniques uses a glass cup to create suction over a painful area. With dry cupping, the cups are applied to the intact skin; with so called wet or bloody cupping, the skin is incised before the cups are applied. Cupping is claimed to increase the local circulation of blood and lymph and to relieve painful muscle tension.210

Clinical evidence for traditional medicine and whole medical systems in rheumatoid arthritis

A three armed pilot RCT compared a standardized Ayurvedic polyherb and a monoherb formulation with hydroxychloroquine sulfate in 121 patients with mild to moderate rheumatoid arthritis. No significant difference in clinical outcomes was found between hydroxychloroquine sulfate and the polyherb formulation after 24 weeks (ACR 20% response 44% v 51%); however, because of the preliminary nature of the study, the results are inconclusive.211 Only a few mild adverse effects occurred in all three groups. In a further pilot RCT, 43 patients with rheumatoid arthritis were randomized to individualized classic Ayurveda herbal formulations, methotrexate, and their combination and followed over 36 weeks.212 ACR 20% response was 86% for methotrexate, 100% for Ayurveda, and 82% for the combination. DAS28 responses were comparable: −2.4 with methotrexate, −1.7 with Ayurveda, and −2.4 with combination. No significant differences in adverse events were found. Larger and higher quality studies comparing Ayurveda with conventional treatments in rheumatoid arthritis are warranted.

Clinical evidence in chronic back and neck pain

Ayurvedic treatment

External Ayurvedic treatments with classic massage techniques that apply medicated oils were compared with standard local thermal therapy over two weeks in a randomized study in 64 patients with chronic low back pain. After two and four weeks, Ayurvedic treatment significantly reduced pain intensity and discomfort compared with control treatment (mean difference of pain VAS after two weeks −18.7, −28.7 to −8.7).213

Anthroposophical medicine

This treatment was evaluated in a cohort study and a further prospective non-randomized comparative study. The comparative study included 86 patients with back pain and compared a complex anthroposophical therapy with the standard approach, finding comparable improvements in both groups after six and 12 months for pain, function, and quality of life.214 A longer term cohort study that assessed effects of anthroposophical medicine in 75 patients with chronic back pain found sustained benefits after two years.215 Future randomized trials are warranted.


A recent systematic review identified three randomized trials of wet cupping for the treatment of chronic low back pain.216 One study with 80 patients applied six cupping sessions according to traditional Arabian techniques over two weeks. Compared with rescue medication, only cupping reduced pain significantly (mean difference on numeric rating scale −29.2, −24.6 to −33.8). The second RCT in 96 patients applied three wet cupping sessions of traditional Arabian style in six days and found significant pain reduction after three months compared with usual care (mean difference −2.17, −1.72 to −2.60). The third RCT in 32 patients applied cupping according to traditional Korean medicine and found no significant difference in pain compared with usual care after two weeks.216

In a randomized study on 50 patients with chronic neck pain, a series of five wet cupping treatments according to traditional European medicine over a period of two weeks was compared with no intervention. After two weeks of treatment, cupping decreased pain (mean difference on numeric rating scale –17.9, −29.2 to −6.6) and reduced pressure pain thresholds.217 Pulsating cupping (pneumatic pulsation therapy) was examined in a randomized trial of 50 patients with chronic neck pain. Five pneumatic pulsation treatments over a period of two weeks decreased pain compared with a waiting list control group (VAS mean difference −11.2, −16.2 to −6.2).218 A further study of this group examined the effectiveness of a partner delivered, home based cupping massage compared with progressive muscle relaxation in patients with chronic neck pain. After 12 weeks of treatment, both groups showed comparable pain reduction.219

Needle pads

Needle stimulation pad (NSP) has often been used for the treatment of neck and back conditions in some cultures. An NSP has up to several hundred sharp but non-penetrating plastic needles. It is typically placed on soft ground and the patient lies for a limited time period on top of the mat with the painful part of the body in contact with the needles. The putative mechanism of NSP is a peripheral antinociceptive effect, and NSP is mostly self administered.

An RCT assessed the effect of NSP on pain and nociceptive thresholds in patients with chronic neck and back pain. Forty patients with chronic neck pain and 42 patients with chronic back pain were randomized to either NSP treatment or a waiting list control group over a period of 14 days. Pain was significantly reduced in the NSP group for neck pain (mean difference −1.6, −2.8 to −0.3) and back pain (−2.3, −3.2 to −1.3), accompanied by a significant increase in pressure pain threshold.220


Taken together, some traditional therapies and whole medical systems may effectively contribute to pain management in rheumatoid arthritis and back and neck pain, although high quality studies are needed. With regard to safety, much variation exists among the various traditional and whole medical systems. For example, Ayurvedic external therapies, cupping, and NSP have minimal adverse effects. Patients need to be informed that after cupping hematoma can occur for several days. Ayurvedic internal herbal preparations have a risk profile similar to that of European herbal medicine. In addition, concern has been expressed about impurities and contamination of Ayurvedic products with heavy metals.219 Patients should be advised not to use products sold via the internet and that are not certified and controlled by pharmacies for any impurities.


The American (ACR) and the European (European League Against Rheumatism) guidelines on rheumatoid arthritis do not mention the use of non-pharmacological treatments or components of CIM.221222 For back pain, the National Institute for Health and Care Excellence’s guidelines recommend self management without specifying modalities.223 Acupuncture is not recommended. The American College of Physicians and the American Pain Society published guidelines on back pain in 2007.224 Self care options were strongly recommended; for patients who do not improve, the addition of acupuncture, yoga, and some relaxation methods received a weak recommendation.

Emerging treatments

The field of clinical research in CIM is growing rapidly, including in the area of mind-body medicine and the whole medical systems, acupuncture, and worldwide traditional medicine systems. The broad spectrum of interventions covered under CIM makes it impossible to list or prioritize individual studies that are in the pipeline. Broadly speaking, many studies on herbal medicine and acupuncture from China and India are expected. In the non-drug/herbal medicine area, several pragmatic effectiveness studies and comparative effectiveness trials with head-to-head comparison of different methods of CIM and conventional medicine are expected in the near future.

Challenges for future research

The mechanisms and clinical effectiveness of these treatment modalities remain to be better determined. Challenges of assessing the clinical effectiveness of acupuncture may include limited knowledge about its mechanisms; effective blinding of study participants; problems with proper study design such as acupuncture “dosing,” sham or placebo treatment, and standardized treatment methods; and informative clinical assessment tools. For mind-body interventions, aspects of adherence and the implementation in medical practice remain challenging, and comparative effectiveness studies should clarify their role in pain management. The clinical effectiveness of nutritional therapies and fasting has to be further investigated in well designed randomized trials. Worldwide traditional medicine and whole medical systems have promising potential, as indicated by the World Health Oorganization,225 and warrant further scientific evaluation. All nutritional and mind-body interventions face the challenge that blinding is almost impossible, making non-specific effects difficult to evaluate.


Demand for complementary and intergrative medicine approaches in the treatment of chronic pain is high, and their use is increasing. Despite the increased use of acupuncture in clinical practice, research data remain inconclusive about its effectiveness in the management of chronic pain. Some of the challenges related to the use and study of acupuncture are related to study methods and assessment tools. For example, the scientific merits of these studies are often limited owing to the poor study design, with non-randomized controls. Also, maintaining true blinding in clinical studies remains difficult, which makes differentiation of a placebo effect from a true treatment effect complicated.

Evidence for the effectiveness of various mind-body interventions in chronic back and neck pain is increasing rapidly. In rheumatoid arthritis, mind-body interventions may improve quality of life and psychological wellbeing. As most of these interventions support active self care, they seem to be useful for comprehensive chronic pain management. Basic research points to a relevant role of fasting and distinct dietary interventions in the treatment of rheumatic disease. However, most clinical evidence stems from studies before the era of modern DMARD therapy, so the effect of diet remains to be assessed in newer clinical trials. Many patients are interested in herbal medicine for chronic back and neck pain and rheumatic disease, but the clinical effects of herbs from Western complementary medicine seem to be modest. Some herbs and herbal mixtures from traditional Chinese and Ayurvedic medicine might be beneficial in rheumatoid arthritis, but their safety is not yet proven. Further traditional therapies and whole medical systems may effectively contribute to pain management in rheumatoid arthritis and back and neck pain, although the evidence is still preliminary. Taken together, CIM has an increasing role in the management of chronic pain, but high quality research is needed.

Questions for future research

  • Will it be possible to rank complementary and conventional therapies with proven effectiveness by means of comparative effectiveness trials?

  • Are traditional concepts for individualization of treatment approaches valid for modern concepts of personalized medicine in order to increase the precision of therapy?

  • What is the long term cost effectiveness of relevant complementary and integrative medicine (CIM) therapies in the management of chronic pain?

  • Will programs with multimodal interventions be superior to single interventions of CIM?

  • Will the combination of CIM modalities with Western medicine lead to better clinical outcomes?

Glossary of abbreviations

ACR—American College of Rheumatology

CBT—cognitive behavioral therapy

CIM—complementary and integrative medicine

CLBP—chronic low back pain

CNP—chronic neck pain

DMARD—disease modifying anti-rheumatic drug

GABA—γ-aminobutyric acid

GLA—γ-linolenic acid

MBSR—mindfulness based stress reduction

MRI—magnetic resonance imaging

NHIS–National Health Interview Survey

NSAID—non-steroidal anti-inflammatory drug

NSP—needle stimulation pad

RCT—randomized controlled trial

SMD—standardized mean difference

TNF-α—tumor necrosis factor-α

TWH—Tripterygium Wilfordii Hook F


  • Series explanation: State of the Art Reviews are commissioned on the basis of their relevance to academics and specialists in the US and internationally. For this reason they are written predominantly by US authors

  • Contributors: Both authors contributed equally to the manuscript.

  • Competing interests: We have read and understood the BMJ policy on declaration of interests and declare the following interests: none.

  • Provenance and peer review: Commissioned; externally peer reviewed.


View Abstract

Log in

Log in through your institution


* For online subscription