Clinical trials evaluating acupuncture for osteoarthritis of the knee constitute some of the highest quality evidence available regarding acupuncture for any condition. As a consequence, researchers can be more certain in their conclusions.
Acupuncture for Chronic Pain: Update of an Individual Patient Data Meta-Analysis (2018) :
Osteoarthritis of the knee was one of the conditions included in this large systematic review, along with back pain, neck pain migraine, tension headaches, and shoulder pain. This review received data from a total of 20,827 patients from 39 trials. As far as we are aware, this is the largest high-quality systematic review that evaluates acupuncture for any condition. In addition to size, the review’s strengths are that it included only high-quality clinical trials and had access to the individual patient data. In many systematic reviews the meta-analysis combines the summary data from clinical trials: for example, the mean (average) pain scores. The meta-analysis in this systematic review used the pain scores from each participant, therefore, the analysis has greater statistical ‘precision’. In summary, the Acupuncture for Chronic Pain: Update of an Individual Patient Data Meta-Analysis is the most reliable assessment of acupuncture to date.
For the above chronic pain conditions the review found:
- acupuncture is superior to ‘no acupuncture controls’
- acupuncture is superior to placebo
- the clinically relevant effects of acupuncture on chronic pain persist overtime
‘No acupuncture controls’ refers to medication or physiotherapy or exercise and advice. In other words, no acupuncture controls refers to the therapies that many people with chronic pain are currently offered.
Some people worry acupuncture is purely a placebo, and that responding to treatment indicates that the pain was ‘all in their heads’. This systematic review demonstrates the benefits of acupuncture cannot be explained only by placebo effects.
Naturally, many people want to know whether the benefits of acupuncture last over time or does it simply make them feel better for a few days. This review demonstrates clinically relevant benefits last for year. Very few clinical trials have followed participants for more than a year, so whether there are benefits beyond a year has yet to be fully investigated.
The review included 10 clinical trials that compared acupuncture to no acupuncture controls for osteoarthritis of the knee. In nine out of 10 of these clinical trials acupuncture was superior to no acupuncture controls. In eight of these the difference was statistically significant. In other words, the difference was probably not due to ‘chance’. The one trial that did not show acupuncture to be superior compared advice/exercise plus acupuncture to advice/exercise. There was no difference between the groups. The combined results showed those receiving acupuncture had less pain with an effect size of 0.63 (95% CI: 0.56-0.69).
The effect size is a standardised way of comparing the size of the effect between groups. For example, the difference between the mean (average) change in pain scores in the groups. It quantifies how much more effective the treatment, acupuncture, is compared to a control group usually sham acupuncture or no acupuncture control.
By convention, 0.2 is considered a small effect, 0.5 medium and 0.8 large. In the Acupuncture for Chronic Pain: Update of an Individual Patient Data Meta-Analysis, for all the chronic pain conditions combined, the effect sizes were:
- acupuncture compared to no acupuncture controls 0.5
- acupuncture compared to sham acupuncture 0.2
To illustrate effect sizes in more clinically applicable terms the authors give the following example. If baseline pain score [before treatment] in a typical clinical trial was 60 on a scale of 0–100, with a standard deviation of 25, follow-up scores might be:
- 30 among acupuncture patients
- 35 in a sham acupuncture group
- 43 in a no acupuncture control group
Comparison of acupuncture with other physical treatments for pain caused by osteoarthritis of the knee: a network meta-analysis (2017) 
This is another important study that investigated acupuncture for osteoarthritis of the knee. A network meta-analysis is a way of comparing different treatment options for a particular condition. The different treatment options can then be ranked in order of relative effectiveness. The rankings when compared to usual care, using only better-quality clinical trials, were as follows:
- Sham acupuncture
- Muscle-strengthening exercise
- Tai ji quan (T’ai chi)
- Weight Loss
- Aerobic exercise
- No intervention
Cochrane Review: Acupuncture for peripheral joint osteoarthritis (2010) 
This review found beneficial effects for acupuncture. For acupuncture versus usual care the effect size was 0.96 for pain and 0.89 for function. When compared to sham acupuncture effect sizes were 0.28 for pain and 0.28 for function. The study authors concluded that sham-controlled trials show statistically significant benefits; however, these benefits were small, and probably due at least partially to placebo effects from incomplete blinding. Waiting list-controlled trials of acupuncture for peripheral joint OA suggest statistically significant and clinically relevant benefits.
Three out of five of these clinical guidelines find in favour of using acupuncture for osteoarthritis of the knee. Please see the Commentary for further discussion.
National Institute for Health and Clinical Excellence: CG177 (2014) 
- The NICE guidelines do not recommend acupuncture for osteoarthritis of the knee.
EULAR recommendations for the non-pharmacological core management of hip and knee osteoarthritis (2013) 
- The European League Against Rheumatism (EULAR) conducted a delphi process prior to undertaking the main review. At this stage the panel of experts decided not to include acupuncture. The reason for this exclusion was not given although two of the recent clinical guidelines, which the panel appears to have considered at this stage, recommended acupuncture (see below ACR & OARSI).
American College of Rheumatology (2012) 
- ACR conditionally recommends acupuncture for severe to chronic osteoarthritis of the knee when the patient is unwilling or unable to undergo total knee arthroplasty.
Osteoarthritis Research Society International (OARSI) (2008/2010) [7, 8]
- OARSI recommend acupuncture for the symptomatic treatment of patients with osteoarthritis of the knee.
The Joint Federal Committee of Physicians and Health Insurance Plans in Germany (Gemeinsamer Bundesausschuss, G-BA)
- Recommend acupuncture for osteoarthritis of the knee, since 2006.
In its idealised form, science is meant to be an objective process. This begs the question as to why the expert panels, which write clinical guidelines, come to different conclusions. The clinical guidelines included in this factsheet are American/ European and are concerned with similar populations. They are not from China or other Asian countries, with different cultural backgrounds that may influence the acceptance of acupuncture. Nor is it the case at some clinical guidelines rely on clinical evidence from China where others do not. These reviews consider basically the same clinical evidence.
Placebo or pragmatic clinical trials
The different conclusions and recommendations can be largely attributed to differences of opinion regarding placebo/sham acupuncture. There are two things to consider: statistical significance and clinical significance.
Statistical significance evaluates whether the differences between groups might be due to chance. This is a bit like playing a betting game with someone by tossing a coin. If the person says heads and wins, we put this down to luck. If we play again and it is heads again there’s a one in four chance of that happening, still lucky: then 1:8, 1:16. The question gradually becomes is the person cheating or just being extremely lucky? At what point should we accuse them of cheating? In clinical research the question is at what point do we conclude that one therapy is better than another? If a clinical trial shows acupuncture to be better than placebo but there is a one in four chance that this was due to luck, we cannot be confident enough in the result to start spending precious resources on providing acupuncture to patients. The solution is to do more research to become more confident about the result. However, if there is only a one in one thousand chance the difference was due to luck, then it is reasonable to recommend acupuncture. Acupuncture for Chronic Pain: Update of an Individual Patient Data Meta-Analysis indicates that acupuncture is superior to placebo and the chance of this result being due to luck is greater the one in one thousand.
Clinical significance is about the size of the difference: the effect size. Clinical significance can also be described as minimal important differences (MID). If the benefits of a therapy are small, it might not be worth allocating precious resource to provide that therapy. Naturally, the value of any improvement is a subjective issue. For someone with the condition any improvement may be significant. Clinical guidelines must, however, consider the cost of delivering a therapy compared to its therapeutic gains.
The 2014 NICE guidance (CG177) reports acupuncture was superior sham/placebo for osteoarthritis knee pain with an effect size of 0.34, measured using the WOMAC scale. When compared to waiting list control the effects size was 0.89. These were both statistically significant results. However, based on these figures the panel concluded:
Ten studies with 2290 people with osteoarthritis of the knee suggested that acupuncture and sham acupuncture may be similarly effective in decreasing pain measured on the WOMAC scale.
Seven studies with 893 people with osteoarthritis of the knee showed that acupuncture was clinically more effective than waiting list control in decreasing pain measured on the WOMAC pain scale.
The expert panel that wrote the guidelines set the MID at 0.5. As a consequence, acupuncture was assessed as being similarly effective to sham/ placebo: because 0.34 is less than the required 0.5 MID. The panel chose to prioritise the comparison with sham/placebo over pragmatic trials of acupuncture versus non acupuncture controls. Therefore, acupuncture was deemed not to be effective and subsequently not included within the guidelines. In other words, acupuncture fell at the sham/placebo hurdle.
An alternative approach is: First, assess whether there is a difference between acupuncture and sham/placebo. There is: in this case an effect size of 0.34. This shows that the effects of acupuncture are not purely psychological. Second, assess the clinical significance in terms of acupuncture versus other therapies and waiting lists controls. This is the practical real-world choice that confronts someone with osteoarthritis of the knee. The authors of Acupuncture for chronic pain and depression in primary care: a programme of research took this approach and concluded acupuncture should be consider chronic pain conditions including osteoarthritis of the knee. This was announced in a National Institute of Health Research Signal.
Active sham/placebo interventions
There is a second problem in emphasising the acupuncture versus sham comparison. There is an underlying assumption that sham procedures are inert placebos. If this were the case, the effect size would be purely the clinical benefits attributable to the physical process of inserting needles at acupuncture points. However, sham acupuncture procedures are mostly likely not inert: they are active. The more active the placebo acupuncture, the higher the hurdle acupuncture must jump before it is included in guidelines.
Not all sham acupuncture procedures are the same and it seems likely some are more active than others. For example, in one clinical trial the sham acupuncture points were so close to real points that there was probably some overlap in practice9. For a detailed explanation of the potential problems with this trial see Appleyard et al 2014. The effect size of acupuncture versus waiting list control in this trial was calculated as 1.0. There was no difference between acupuncture and sham acupuncture. This means the effect size of the sham acupuncture compared to waiting list control was also in the region of 1.0. This is more effective than treatments currently recommended.
The NICE guidelines (CG177) identify the following core treatments:
Activity and exercise
Interventions to achieve weight loss
A network meta-analysis2, discussed in the Summary section, showed that the effect size of acupuncture compared to: sham was 0.34 (same as the NICE guidelines); muscle strengthening exercises 0.49; weight loss 0.93; aerobic exercise 1.09. Therefore, it seems that if the criteria set for acupuncture was MID of 0.5 compared to the current core treatments it would have been included in the guidelines. However, because the primary criterion was the assessment compared to sham acupuncture it was not included.
An integrative approach and normal practice
One of the potential weaknesses with clinical trials investigating acupuncture is the tendency to provide only a short course of treatment. For example, some trials investigating osteoarthritis knee provided six treatments and then measured the effect a year later10. This does not necessarily reflect actual practice. Commonly, people may be treated for initial course of six weeks but then come back for top-up treatments at widening intervals.
The value of weight loss interventions and exercise programmes will naturally be dependent on individuals. Therefore, in some respects is not necessarily helpful to see acupuncture compared to these interventions. In practice, there is no reason why someone cannot have acupuncture and do exercise. So, the best approach for some individuals maybe to integrate these therapies.
1. Vickers AJ, Vertosick EA, Lewith G, et al. Acupuncture for Chronic Pain: Update of an Individual Patient Data Meta-Analysis. J Pain 2018;19(5):455-74. doi: 10.1016/j.jpain.2017.11.005
2. MacPherson H, Vickers A, Bland M, et al. Acupuncture for chronic pain and depression in primary care: a programme of research. 2017 doi: 10.3310/pgfar05030
3. Manheimer E, Cheng K, Linde K, et al. Acupuncture for peripheral joint osteoarthritis: John Wiley & Sons, Ltd, 2010.
4. NICE. Osteoarthritis: care and management: NICE Guidelines [CG177]: National Institute for Health and Clinical Excellence, 2014.
5. Fernandes L, Hagen KB, Bijlsma JWJ, et al. EULAR recommendations for the non-pharmacological core management of hip and knee osteoarthritis. Annals of the Rheumatic Diseases 2013;72(7):1125. doi: 10.1136/annrheumdis-2012-202745
6. Hochberg MC, Altman RD, April KT, et al. American College of Rheumatology 2012 recommendations for the use of nonpharmacologic and pharmacologic therapies in osteoarthritis of the hand, hip, and knee. Arthritis Care & Research 2012;64(4):465-74. doi: https://doi.org/10.1002/acr.21596
7. Zhang W, Nuki G, Moskowitz RW, et al. OARSI recommendations for the management of hip and knee osteoarthritis: part III: Changes in evidence following systematic cumulative update of research published through January 2009. Osteoarthritis And Cartilage / OARS, Osteoarthritis Research Society 2010;18(4):476-99. doi: 10.1016/j.joca.2010.01.013
8. Zhang W, Moskowitz RW, Nuki G, et al. OARSI recommendations for the management of hip and knee osteoarthritis, Part II: OARSI evidence-based, expert consensus guidelines. Osteoarthritis And Cartilage / OARS, Osteoarthritis Research Society 2008;16(2):137-62. doi: 10.1016/j.joca.2007.12.013
9. Suarez-Almazor ME, Looney C, Liu Y, et al. A randomized controlled trial of acupuncture for osteoarthritis of the knee: effects of patient-provider communication. Arthritis Care & Research 2010;62(9):1229-36. doi: 10.1002/acr.20225
10. Foster NE, Thomas E, Barlas P, et al. Acupuncture as an adjunct to exercise based physiotherapy for osteoarthritis of the knee: randomised controlled trial. BMJ Clinical Research 2007;335(7617):436-36. doi: 10.1136/bmj.39280.509803.BE