A network for students interested in evidence-based health care

Acupuncture vs. physiotherapy for severe knee osteoarthritis

Posted on 15th February 2017 by

Evidence Reviews


This is a summary of a randomized control trial (RCT) comparing the effects of standardized western acupuncture and physiotherapy on pain and functional ability in patients with severe osteoarthritic knee pain awaiting knee arthroplasty (i.e. knee replacement).

Osteoarthritis (OA) is a degenerative disease of joint cartilage, and is a major cause of pain, disability and healthcare utilization. One of the treatments of late stage OA is a total joint arthroplasy, or joint replacement surgery. The authors looked at the impact of OA on knee pain and function as well as post-operative recovery.


The authors compared acupuncture and physiotherapy against each other (treatment groups), and against current standard management (control group). Both treatment groups attended a weekly 1 hour group session with 6-10 people for 6 weeks led by a senior physiotherapist.

The acupuncture group had needles inserted for 20 minutes at points most commonly used in previous reported research studies in the field, and up to three additional needles in trigger or traditional points at the physiotherapist’s discretion.

The physiotherapy group attended an exercise circuit devised and supervised by the same physiotherapist who provided the acupuncture. There were 10 exercises in the circuit which included balance training and lower extremity strengthening.

The control group received an exercise and advice leaflet, which had been designed by consensus between the physiotherapy, rheumatology and orthopaedic departments to reflect best current practice. At enrolment, patients were told that they were in the ‘home exercise group’.

The primary outcome measure was the Oxford Knee Score (OKS), a patient-completed questionnaire consisting of twelve questions on knee function and pain. Other outcomes were the Western Ontario MacMaster (WOMAC) Score, 10-cm pain visual analogue scale (VAS), Hospital Anxiety and Depression score (HAD). Fifty metre timed walk and weight were also measured at each visit.


Compared to patients in the control group, patients with severe knee OA achieved a short-term improvement in their Oxford Knee Score when treated with acupuncture. The difference between the control group and the physiotherapy group was not significant. Additionally, no other clinically or statistically significant effects were found between the groups.

Although the improvement in OKS in the acupuncture group was the only positive result out of a series of non-significant results, the study demonstrated that there may be some potential for improvements to knee pain and function, perhaps with longer or more intensive treatment courses. The effects observed in this study were smaller than those found in other knee OA studies, which all involved longer treatment courses in less disabled patients. Additionally, while not significant, the physiotherapy group had a shorter post-operative hospital stay (mean 6.5 days, S.D. 2.0) than the acupuncture group (mean 7.77, days S.D. 3.96). More research is needed to determine the impact of longer, more intensive, and combined treatments.

Study Quality


  • This study had high dropout rates: A hundred and sixty-one (89%) patients were assessed at 7 weeks, 120 (66%) patients where assessed at 12 weeks and 69 (38%) patients were assessed at 3 months post-operatively.
  • A 6-week course of treatment was used due to funding constraints, which was the minimum anticipated treatment time needed to document any meaningful acupuncture improvement, based on clinical experience and current practice. This may not have been long enough to build up the full effects of the physiotherapy.
  • The treatment groups were limited to standardized programs, rather than individual treatment plans.
  • The participants were not blinded as to which intervention they received, nor was sham acupuncture used. 


  • The research fellow carrying out all the assessments was blind to patient treatment allocation until the data analysis was completed.
  • Analysis was performed using intention-to-treat analysis, with the baseline values used in place of any missing follow-up values..
  • The study population was similar between groups at baseline, and the treatment groups were treated similarly (by the same practitioner in a group of 6-10, for 1 hour each week for 6 weeks) to minimize sources of variation.
  • The study used reproducible therapies and primary outcome measures, which are directly relevant to current clinical practice.


Williamson, L., Wyatt, M.R., Yein, K., and Melton, J.T.K. (2007). Severe knee osteoarthritis: a randomized controlled trial of acupuncture, physiotherapy (supervised exercise) and standard management for patients awaiting knee replacement. Rheumatology 2007;46:1445–1449. https://academic.oup.com/rheumatology/article/46/9/1445/1788894/Severe-knee-osteoarthritis-a-randomized-controlled 

See also: Manual therapy vs Exercise for Knee OA 


Kristen Dufresne

Physical Therapy student at the University of Toronto View more posts from Kristen

Leave a Reply

Your email address will not be published. Required fields are marked *

No Comments on Acupuncture vs. physiotherapy for severe knee osteoarthritis

  • Kristen Dufresne

    Hello Mike, thanks for your comment. It has allowed me to delve into this subject more and I now have a better understanding of this topic and where this literature falls in the body of evidence for knee OA pain. I became interested in the topic of pain management for severe knee OA recently when I was able to observe an initial assessment, but the surgeon told them the soonest opening would be in the Fall. The patient was devastated – they were already struggling to participate in their activities of daily living, let alone their job, and would now have to continue like that for almost another year.

    When I got home I quickly searched for evidence on osteoarthritis and pain management. Unsurprisingly, NSAIDs were a commonly found solution, but at the most severe stages of the disease it can be difficult to control pain even with additional medications. When I found this study by Williamson, Wyatt, Yein & Melton comparing physiotherapy and acupuncture, and surprisingly the physiotherapy was less effective, I knew I wanted to critically appraise it. Less surprisingly, this article had several limitations and was not of high quality.

    Here is a similar study of higher quality that included a placebo/sham acupuncture treatment, a major flaw in the Williamson, Wyatt, Yein & Melton article, as you said: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1962890/?tool=pmcentrez. A key point from this article was that “participants receiving either acupuncture intervention were significantly more confident that treatment could help their knee problem than those receiving advice and exercise alone,” which implies that acupuncture may indeed have the impact expected by patients (i.e., a placebo effect). Similarly to the article by Williamson, Wyatt, Yein & Melton, this study concluded “we have shown that acupuncture offers limited additional benefits when delivered alongside an effective package of physiotherapy led advice and exercise.”

    Of note, I do think it’s worth reviewing studies including treatments such as acupuncture as in the future I will be treating patients who will ask my opinion on a variety of available treatments, and simply telling them that acupuncture is no better than a placebo without discussing the potential benefits and risks with them would be poor patient education and care. To conclude, from the article you cited in your comment: “All that remains to be discussed is whether or not the placebo effect is big enough to be useful, and whether it is ethical to prescribe placebos.”

    24th February 2017 at 12:23 am
    Reply to Kristen
  • Mike

    This study is fatally flawed. There was no blinding nor any control for placebo effects. Other studies have shown that when placebo effects are controlled for (e.g. via blinding), any supposed effect of acupuncture goes away. Even in this study, the effect did not last.

    Here is Cochrane contributor Edzard Ernst discussing a study that had better blinding:

    Acupuncture is a theatrical placebo, not effective for any condition.

    15th February 2017 at 11:08 pm
    Reply to Mike

Subscribe to our newsletter

You will receive our monthly newsletter and free access to Trip Premium.