Article Text

Is acupuncture effective for knee osteoarthritis? A protocol for a systematic review and meta-analysis
  1. Chuan-Yang Liu1,2,
  2. Jian-Feng Tu3,
  3. Myeong Soo Lee4,
  4. Ling-Yu Qi3,
  5. Fang-Ting Yu3,
  6. Shi-Yan Yan3,
  7. Jin-Ling Li3,
  8. Lu-Lu Lin3,
  9. Xiao-Wan Hao3,
  10. Xin-Tong Su3,
  11. Jing-Wen Yang3,
  12. Li-Qiong Wang3
  1. 1School of Traditional Chinese Medicine, Beijing University of Chinese Medicine, Beijing, China
  2. 2School of Traditional Chinese Medicine, Capital Medical University, Beijing, China
  3. 3School of Acupuncture-Moxibustion and Tuina, Beijing University of Chinese Medicine, Beijing, China
  4. 4KM Science Research division, Korea Institute of Oriental Medicine, Daejon, South Korea
  1. Correspondence to MD Li-Qiong Wang; wangliqiongwork{at}; MD Jing-Wen Yang; yangjw0626{at}


Introduction Knee osteoarthritis (KOA) is one of the leading causes of disability. The effectiveness of acupuncture for treating KOA remains controversial. This protocol describes the method of a systematic review and meta-analysis evaluating the effectiveness and safety of acupuncture for treating KOA.

Methods and analysis Four English databases (PubMed, Embase, Cochrane Library databases and Web of Science) and four Chinese databases (China National Knowledge Infrastructure, Chinese Biomedical Literature Database, VIP Database for Chinese Technical Periodicals, and Wanfang) will be searched from the database inception to 1 September 2021. All randomised controlled trials related to acupuncture for KOA will be included. Extracted data will include publication details, basic information, demographic data, intervention details and patient outcomes. The primary outcome will be pain intensity. Risk of bias will be assessed using the Cochrane Collaboration’s tool for assessing risk of bias. Article selection, data extraction and risk of bias assessment will be performed in duplicate by two independent reviewers. If the meta-analysis is precluded, we will conduct a descriptive synthesis using a best-evidence synthesis approach. The strength of recommendations and quality of evidence will be assessed using the Grading of Recommendations Assessment Development and Evaluation working group methodology.

Ethics and dissemination Ethics approval is not required because individual patient data are not included. This protocol was registered in the international Prospective Register of Systematic Reviews on 25 February 2021. The systematic review and meta-analysis will be submitted for publication in a peer-reviewed journal. The findings will also be disseminated through conference presentations.

Trial registration number CRD42021232177.

  • knee
  • protocols & guidelines
  • complementary medicine

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See:

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Strengths and limitations of this study

  • This systematic review and meta-analysis will evaluate the effectiveness and safety of acupuncture in treating knee osteoarthritis by collecting comprehensive evidence.

  • We want to focus on many different factors in subgroup analysis and to explore the applied law of different doses of acupuncture.

  • This study will explore the difference in the effectiveness between manual acupuncture and electroacupuncture for knee osteoarthritis by synthesising the evidence from direct comparison and indirect comparison.

  • We plan to search multiple Chinese and English language databases to ensure a comprehensive search of the literature.

  • Transformation of pain scores will result in loss of some accuracy; however, we believe that it is clinically irrelevant.


Description of the condition

Osteoarthritis (OA) is a common clinical degenerative disease and is one of the leading causes of disability.1 The excess costs of adults with OA are considerable, estimated at $45 billion annually in USA.2 Knee osteoarthritis (KOA) accounts for approximately 85% of global OA burden.3 With the trends of an ageing population and increasing obesity, the incidence of KOA is increasing for both sexes.4 5 In addition, pain symptoms associated with KOA result in physical and walking disability, which in turn give rise to an excess risk of all-cause mortality.6 7

Exercise and weight loss, two effective nonpharmacological treatments, are strongly recommended in all people with clinical OA.8 However, for patients with KOA, it is difficult to continue exercising and losing weight. Representatives of pharmacological interventions include analgesics and non-steroidal anti-inflammatory drugs (NSAIDs). However, acetaminophen (paracetamol) is not associated with long-term pain improvement.9 Furthermore, many NSAIDs are associated with serious side effects such as cardiovascular, renal adverse effects and gastrointestinal bleeding.10 11 In addition, the healthcare systems of Western countries are overstretched because of the increasing joint replacement requirements.12 In this context, identification of the efficacy of existing treatments or development of novel therapies remains an important priority.

Description of the intervention

Acupuncture has long been recognised as a non-pharmacological therapy in treating various disorders by inserting fine needles into specific anatomic points (acupoints) on the skin of the patient’s body. As an important component of traditional Chinese medicine (TCM), acupuncture has been used in clinical practice for more than 3000 years.

The WHO has recommended acupuncture therapies for 107 diseases. The effectiveness of acupuncture for different kinds of pain diseases has been verified by a great deal of high-quality clinical trials.13–15 Recently, two individual patient data meta-analyses also reported that acupuncture was effective for the treatment of chronic pain, with treatment effects persisting over time.16 17 In addition, acupuncture appears to be a safe intervention that has rare adverse effects in the hands of competent practitioners.18 19

How the intervention might work

KOA is a prevalent, chronic joint disorder, characterised by synovitis, overgrowth of subchondral bone, development of osteophytes, erosions and loss of the articular cartilage. Previous study found that cartilage damage is the origin and result of KOA. With the further study of KOA, synovitis has been verified to play a crucial part in the pathological development and the maintenance of pain in KOA.20

In recent decades, preclinical investigations of acupuncture mechanisms in KOA pain have increased. These studies show that acupuncture relieves symptoms of KOA by activating a variety of bioactive chemicals through peripheral, spinal and supraspinal mechanisms.21 For example, acupuncture can desensitise peripheral nociceptors and reduce proinflammatory cytokines peripherally and in the spinal cord.21–23 In addition, acupuncture dampens the transmission of noxious inputs at the spinal level with the involvement of spinal opioids, serotonin (ie, 5-hydroxytryptamine), norepinephrine, glial cell/cytokines and signal molecules.21 24–26 In addition, CBR1-GABA-5-HT may be a novel pathway contributed to the effect of electroacupuncture (EA) on KOA pain.27 EA downregulated IL-1β expression via activating the peripheral CBR2 to inhibit the KOA pain.28

Why it is important to perform this review

Research on acupuncture for KOA has been growing, but the findings have been inconsistent. Different guidelines do not reach an agreement on whether acupuncture should be recommended as an effective non-pharmacological treatment for KOA.8 29–31 In 2014, a clinical trial showed that acupuncture did not confer a benefit over sham treatment for pain or function.32 In 2019, however, a review suggested that acupuncture provided relief of pain associated with KOA.33

Most meta-analyses mainly focused on chronic pain and peripheral joint OA and were not specific to KOA.16 17 34–37 Although there were some systematic reviews conducted to establish the association of acupuncture with KOA, few drew a definitive conclusion.38 39 One systematic review has looked at the comparative effectiveness of manual acupuncture (MA) and EA, but considered only direct evidence.40 Furthermore, some rigorous randomised clinical trials (RCTs) in this field published within recent years were not included in previous systematic reviews. For example, a multicentre RCT published in 2020 by our team suggested that acupuncture had potential benefits for KOA.41 Thus, it is important to perform a systematic review and meta-analysis to inform clinical practice.


We aim to evaluate the effectiveness and safety of acupuncture for treating patients with OA of the knee by conducting a systematic review and meta-analysis. For this purpose, we put forward the following questions about this review:

  1. Is acupuncture effective for treating OA of the knee compared with sham control or no-acupuncture control?

  2. Is there a difference in the effectiveness between MA and EA?

Methods and analysis

Patients and public involvement

There will be no patients or public directly involved in this review. Only data already existent in the literature and the aforementioned sources will be used for this study.

Protocol registration

This protocol was registered in PROSPERO (CRD42021232177). It will be followed the standard methods of systematic review and meta-analysis. It will adher to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses reporting guidelines (see online supplemental appendix 1).42 43

Criteria for including studies in this review

Types of studies

RCTs (with or without blinding, including crossover design) of acupuncture therapy for KOA will be included. We will consider including older RCTs that were cited in previous reviews of acupuncture for OA.

Types of participants

Studies enrolling participants diagnosed as KOA will be included. The diagnostic criteria should be based on the American College of Rheumatology clinical criteria, National Institute for Health and Clinical Excellence guidelines or any other accepted guidelines.8 44 There will be no restrictions on their age, sex, race, education, economic status, Kellgren-Lawrence score or Outbridge score.45 46

Types of interventions

The eligible intervention is acupuncture including MA and EA. There will be no restriction on the sessions of acupuncture, needling techniques or stimulation methods.

Types of control groups

In this review, we plan to compare needle acupuncture with sham acupuncture, analgesic, usual care or waiting list control groups. Acupuncture plus one or more therapies with the same therapies also will be included.


Primary outcome

Pain intensity: The WOMAC Pain Subscale, Visual Analog Scale (VAS), Brief Pain Inventory, Numerical Rating Scale (NRS), Verbal Rating Scale or other validated outcome measures.

Secondary outcomes

  1. Function: The WOMAC Function Subscale, Lysholm Scale or other validated scales.

  2. Quality of life: The 12-Item Short Form Health Survey, 36-Item Short Form Health Survey, Assessment of Quality of Life Instrument or other validated scales.

  3. Adverse events: incidence and severity of adverse events.

  4. Drug use: number of people using emergency analgesics, frequency or dosage of medication for KOA.

  5. Cost: incremental cost-effectiveness ratio of acupuncture treatment.

Criteria for excluding studies in this review

  1. Participants with knee pain but no other criteria of KOA.

  2. The intervention group received transcutaneous electrical nerve stimulation.

  3. Studies reported only improvement rates.

  4. Studies comparing one type of acupuncture with other type of acupuncture (except EA vs MA) and studies comparing acupuncture with complementary therapies or TCM.

Search methods for identification of studies

Electronic searches

We developed search strategies for four English databases (PubMed, Embase, Cochrane Library databases and Web of Science) and four Chinese databases (China National Knowledge Infrastructure, Chinese Biomedical Literature Database, VIP Database for Chinese Technical Periodicals, and Wanfang) from database inception to 1 September 2021. Additional trials will be identified by searching previous systematic reviews. No language or publication status restrictions are applied. The search strategy components are clinical condition (OA, chondromalacia patellae, knee, knee pain and gonarthrosis), intervention (acupuncture, EA and acupuncture points) and study type (RCT). We will adapt the search strategies to medical subject headings terms and keywords as necessary for each database (see online supplemental appendix 2 for the search strategy used in the PubMed database). A pilot of the systematic search was conducted on 28 February 2021 (see online supplemental appendix 3). We (F-TY and CL) will rerun the searches before submission of the manuscript to identify any eligible articles published since our first search.

Searching other sources

We will search the following websites as a supplement: the WHO International Clinical Trials Registry Platform and the National Institutes of Health clinical registry and the Chinese Clinical Registry. The search will also include a manual search for grey literature (eg, unpublished conference articles).

Data collection and analysis

Selection of studies

All search results will be exported to EndNote, where we will check for and exclude duplicates. Two of us will screen all titles and abstracts independently to identify potentially relevant studies. Full texts will be downloaded and printed for further assessment. Two reviewers will screen the whole-length articles to confirm whether the studies meet the inclusion criteria. Any disagreement will be settled by discussion. If an agreement cannot be reached, a third reviewer will be consulted. The reasons for excluding studies will be recorded. The study selection process is shown in figure 1. Besides, we will add a table of excluded studies with reasons for exclusion to the appendix of our meta-analysis.

Figure 1

Flow diagram of the study selection process. KOA, knee osteoarthritis.

Data extraction and management

All data will be extracted independently and in duplicate by two reviewers with a predesigned data extraction template. Disagreements will be settled by discussion. A third reviewer will be consulted if discrepancies cannot be resolved. All data will be cross-checked by two reviewers and transferred into Microsoft Office Excel. If required, we will contact the corresponding authors for more information by email.

The predefined variables for extraction are the following:

  1. Publication details (study year, first author, funding source).

  2. Basic information (location, study type, number of centres, sample size, study duration and length of follow-up).

  3. Participants (type and/or stage of KOA, mean age, sex and pain intensity before treatment).

  4. Interventions (type of acupuncture, choice of acupuncture points, number of sessions, treatment frequency, duration of each session and needling techniques).

  5. Control (if there is any control, details of the treatment, including the name, dosage, frequency and course).

  6. Outcomes (data and time points for each measurement, type and number of adverse events in each group).

Risk-of-bias assessment in included studies

Two reviewers will assess the risk of bias in the included studies by using the Cochrane Collaboration’s tool for assessing risk of bias. We will assess each RCT a low, high or unclear risk of bias for six domains: selection bias (random sequence generation and allocation concealment), performance bias (blinding of researchers and participants), attrition bias (incomplete outcome data), ascertainment bias (blinding of outcome assessment), reporting bias (selective outcome reporting) and other sources of potential bias. Disagreements will be resolved by discussion, according to the published articles and supplementary materials. We will consult the third reviewer and contact the study authors when needed.

Acupuncture adequacy assessment

We will use the adequacy assessment instrument to assess treatment adequacy in acupuncture RCTs from the following four aspects of acupuncture treatment: choice of acupuncture points, number of sessions, needling technique and experience of the acupuncturists.47 Two assessors who are experienced acupuncturists will assess adequacy independently and reach an agreement by discussion. They will be blinded to the results of the study and the publication and conduct the assessments only based on the description of the study population and the acupuncture procedure. To test the success of the blinding, we will ask the assessors to guess the identity of each study.

Heterogeneity assessment

If there are sufficient data, we will conduct a meta-analysis to determine the effectiveness of acupuncture and the related factors. I2 testing will be used to quantify heterogeneity among the included studies.48 We will present summary estimates in forest plots. If the I2 is more than 50%, we will explore the possible sources of heterogeneity via metaregression and subgroup analyses. If a meta-analysis is not appropriate, we will conduct a descriptive synthesis using a best-evidence synthesis approach.

Reporting bias assessment

We will also consider assessing the reporting bias and small-study effects by using funnel plots when there are 10 or more trials. We will assess funnel plot asymmetry by using Begg’s and Egger’s tests and will define significant publication bias as a p value<0.1. We will also use a trim-and-fill computation to estimate the effect of publication bias on the interpretation of the results.49

Data synthesis

When the meta-analysis is performed, Stata V.16.0 and RevMan V.5.3 will be used for all statistical calculations. All the analyses will be based on the random-effect model because the RCTs included by us may come from different populations. For dichotomous variables, the Mantel-Haenszel method will be used for analyses and effect size will be reported as relative risk with 95% CIs. For continuous variables, the inverse variance method will be used for analyses and treatment effect will be reported as mean difference with 95% CIs. The standardised MD with 95% CIs will be used if different scales are used to evaluate a predesigned outcome.

For pain variance, we plan to pool data from previous studies reporting VAS 100 mm, VAS 10 cm, and NRS by transforming it to a ‘0–100-pain measure’ using an appropriate multiplier. We also intend to analyse pain intensity by independently reporting the aforementioned scales.

Subgroup analysis

Subgroup analyses will be performed to explain the heterogeneity. Predefined subgroups include the location of studies, the type of intervention, the dosage of acupuncture, the stage of KOA and the TCM types of KOA.

Sensitivity analysis

We will conduct a sensitivity analysis to verify the robustness of the review conclusions. We will consider removing one study at a time to observe its effect on heterogeneity and effect size. In addition, the meta-analysis will be repeated after studies with lack of allocation concealment are excluded.

Other analysis

If MA and eletroacupuncture are effective for KOA compared with sham acupuncture, we will conduct the exploratory research to compare the difference in the effectiveness between MA and EA by synthesising the evidence from direct comparison and indirect comparison. For direct comparison results, we will use Revman to analyse. For the indirect comparison, we will choose sham acupuncture as a common comparator and use R software to analyse. Finally, we will conduct a mixed treatment comparison meta-analysis to synthesise the evidence from direct comparison and indirect comparison.

Strength of recommendations and the quality of evidence

We will assess the strength of recommendations based on the Grading of Recommendations Assessment Development and Evaluation working group methodology. The two categories of weak/conditional evidence and strong evidence will be used.

We will also assess the quality of evidence. The quality of evidence will be assessed according to the domains of risk of bias, consistency, directness, precision and publication bias. The assessments will be adjudicated into four levels: high, moderate, low or very low.50 51


This systematic review will be performed based on previous studies of acupuncture for KOA. Conclusions drawn from this review may be beneficial to patients with KOA, clinicians and policy-makers. We will summarise and explain the characteristics and findings of the included studies by conducting a systematic narrative synthesis.

Based on the above, we want to conduct some exploratory studies. (1) Is there a difference in the effectiveness between MA and EA? (2) Is the effectiveness (if any) related to the stage of KOA according to the Kellgren-Lawrence score or Outbridge score, some characteristics of acupuncture (eg, treatment frequency), type of control group, measurement time points of outcomes or other variables?

MA and EA are the most commonly used acupuncture therapies. MA maintains a moderate dose of stimulation by lifting, inserting and twisting needles to acupoints. However, it is laborious and difficult to reach an agreement on standards because of the different needle techniques. EA, which is widely used in clinical practice, refers to the pulse current input to acupoints on the basis of needle acupuncture. This approach can accurately control the dose of stimulation and save labour. In clinical trials for pain conditions, better analgesia appears to be obtained when electrical stimulation is added to manual stimulation than with MA needle stimulation alone.52 However, the findings may not be generalisable because of the different pain types.

There are many factors affecting the effectiveness of acupuncture. One review presented ‘the challenge of adequacy of dose’ recently.33 Our group built a scoring instrument to calculate the dose of acupuncture from four parameters.53 Based on the sum of the scores, we defined three doses of acupuncture treatment: high dosage, medium dosage and low dosage. And we designed three subgroups according the three kinds of dosage to explore the relationship between doses of acupuncture and effectiveness. Deqi response is a comprehensive sensation of soreness, numbness, heaviness, aching at and around acupoints produced by manipulation of the needles. It plays a role in acupuncture dosage so it is only one dimension of our scoring instrument. On one hand, not all types of acupuncture need a Deqi response during sessions. For example, MA and EA are required to cause a Deqi response while wrist-ankle acupuncture is not. Furthermore, electrical stimulation can enhance Deqi response elicited by manipulation of needles. Based on the different duration of active stimulation, it is necessary to compare the effectiveness of MA and EA.52 On the other hand, Deqi response is more emphasised in China than Western countries.54

Acupuncture has both specific effects caused by intervention itself and non-specific effects including patient–acupuncturist relationship, patient expectations, and so on. Sham acupuncture group has usually been set in order to eliminate non-specific effects. The sham acupuncture can be divided into superficial insertion and non-penetrating insertion at traditional acupuncture points or not.55 Superficial insertion is not a physiologically inert procedure and thus decreases the difference between groups.56 Therefore, more and more trials choose non-penetrating sham acupuncture at non-acupoints as control to minimise the physiological effects of sham acupuncture.

The proposed review has several strengths. We plan to search multiple Chinese and English language databases to ensure a comprehensive search of the literature. Any meta-analyses will be performed according to the Cochrane Handbook for Systematic Reviews of Interventions. A further strength is that stringent eligibility criteria will be applied to ensure the quality of the included RCTs. In addition, pain intensity was selected as the targeted outcome because it plays an important role in the pain management of KOA. Transformation of pain scores measured by different pain scales to a 0–100 pain measure will result in loss of some accuracy; however, we believe that it is clinically irrelevant.

Ethics and dissemination

Ethics approval is not required because individual patient data are not included. This protocol was registered in the international Prospective Register of Systematic Reviews on 25 February 2021. The systematic review and meta-analysis will be submitted for publication in a peer-reviewed journal. The findings will also be disseminated through conference presentations.

Ethics statements

Patient consent for publication


The authors are grateful to Dr. Hong-ping Li and Dr. Mei Han for their helpful assistance.


Supplementary materials


  • Contributors C-YL is the guarantor and first author of the protocol. C-YL and L-QW designed the systematic review. C-YL and L-LL drafted the manuscript. J-WY, J-FT, L-QW and M-SL provided help to design and edited the manuscript. C-YL and F-TY will independently screen the eligible studies. C-YL and X-WH will extract data from included articles. J-LL and J-FT will assess the risk of bias. C-YL and X-TS will assess acupuncture adequacy, strength of recommendations and the quality of evidence. C-YL, L-YQ and S-YY will finish data synthesis. L-QW will arbitrate any disagreements during the review. All authors have read the manuscript and approved the final publication of the protocol.

  • Funding This study was supported in part by the National Natural Science Foundation of China (grant number 82004223).

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.