Acupuncture for osteoarthritis
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
To review all randomized or quasi‐randomized (not strictly randomized methods of allocation e.g., by date of birth, hospital record number or alternation) controlled trials pertaining to the efficacy of acupuncture in the treatment of OA of the knee, hip or hand. The three specific objectives of this systematic review are:
1) To assess the efficacy of acupuncture compared to standard medical treatment for osteoarthritis.
2) To assess the efficacy of true acupuncture compared to sham acupuncture for osteoarthritis of the knee.
3) To assess the efficacy of acupuncture plus standard medical treatment compared to standard medical care alone for osteoarthritis of the knee.
Background
Osteoarthritis (OA) is the most common form of arthritis, and is a major cause of morbidity, physical limitation, and increased health care utilization, including total joint arthroplasty, especially in the elderly (Scott and Hochberg 1993; Hochberg 1997; Lawrence 1998).
Recent years have seen advances in our understanding of this condition. For example, it is now believed that OA may not be a single disorder but rather a group of overlapping distinct diseases. Presently there is no cure for OA (Creamer and Hochberg 1997; Creamer 1998). Therefore, the treatment of OA is primarily focused managing the condition by minimizing morbidity. Current recommendations for the management of OA, including guidelines published by the American College of Rheumatology, focus on the relief of pain and stiffness and maintenance or improvement in functional status as important goals of therapy (Hochberg 1995).
Many pharmacological therapies for OA can be associated with significant adverse effects (Blackshear 1985; Herman 1986; Brandt 1993, Schlegal 1986). Therefore, the need for effective, nontoxic therapies for OA, has become evident. Non pharmacologic physical modalities include patient education, physical and occupational therapy, aerobic and resistive exercises and weight loss (Puett and Griffin 1994).
Acupuncture, one non pharmacologic modality, is gaining popularity among OA patients, especially those who have failed to respond to standard therapy. Approximately one million consumers utilize acupuncture annually in the United States (Paramore 1997) and that a large percentage of these users are patients with musculoskeletal disorders such as OA (Diehl 1996). Although the safety of acupuncture has been well demonstrated (Lytle 1993) and the utilization of acupuncture is increasing, the efficacy of acupuncture for OA, either as a stand alone treatment or an adjunct to standard medical care, remains in question.
Several models explaining the possible mechanisms of action for acupuncture analgesia have been proposed. The most widely researched and documented hypothesis, known as the neurohumoral model of acupuncture analgesia (Han 1997; Huang 1987; Sims 1997) will be briefly described here. This model proposes that the acupuncture needle stimulates the A delta afferents (types II and III) of a peripheral neuron. This neuron, terminating in the dorsal horn of the spinal cord, transfers the impulse to a second neuron within the same spinal segment. This second neuron then activates three levels of the nervous system. In ascending order, these levels are: 1) the spinal cord itself, within the same segment, 2) the midbrain, where two of the three anatomical structures of the body's internal analgesia mechanism reside: the periaqueductal gray matter (PAGM) and raphe magnus nucleus (RMN), and 3) the pituitary‐hypothalamus complex.
As each of these three levels are stimulated, specific endorphins and monomines (serotonin and noradrenalin) become involved in a chemical cascade, which inhibits pain. Additionally, low frequency (2‐4 Hz) electro‐ acupuncture is believed to stimulate the release of ACTH, stimulating the adrenal cortex and the release of cortisol, producing anti‐inflammatory effects.
Objectives
To review all randomized or quasi‐randomized (not strictly randomized methods of allocation e.g., by date of birth, hospital record number or alternation) controlled trials pertaining to the efficacy of acupuncture in the treatment of OA of the knee, hip or hand. The three specific objectives of this systematic review are:
1) To assess the efficacy of acupuncture compared to standard medical treatment for osteoarthritis.
2) To assess the efficacy of true acupuncture compared to sham acupuncture for osteoarthritis of the knee.
3) To assess the efficacy of acupuncture plus standard medical treatment compared to standard medical care alone for osteoarthritis of the knee.
Methods
Criteria for considering studies for this review
Types of studies
Randomized and quasi‐randomized controlled trials will be included. Because quasi‐randomized studies will be included in the review, a sensitivity analysis will be performed to assess whether these studies differ qualitatively from the randomized trials.
Types of participants
Participants greater than 18 years of age, of either gender, and having a diagnosis of osteoarthritis of the extremities (hip, knee and small joints) will be included. Osteoarthritis of the spine will not be included. Studies stating the participants had "knee pain" or "hip pain" but not explicitly stating that the pain was due to osteoarthritis will be excluded from the review.
Types of interventions
a. Acupuncture will be limited to the stimulation of acupuncture points by needle insertion. Both forms of needle manipulation (manual or electrical) will be accepted. However, methods of stimulating acupuncture points other than needle insertion, e.g., digital pressure or surface electrodes (TENS or infrared) will be excluded.
b. Sham (fake) acupuncture will be defined as the superficial needling of non acupuncture points, usually without needle manipulation, done either proximally and/or distally to the true acupuncture points.
c. Standard medical care will be defined as the use of non steroidal anti‐inflammatory medications (NSAIDs) or localized steroid injections in the affected joint(s). If physiotherapy or physical therapy is the comparison group intervention, this will be analyzed separately from NSAIDs.
Types of outcome measures
The primary outcome measures will include those proposed by Outcome Measures in Rheumatology Clinical Trials (OMERACT) (Bellamy 1997). These are: 1) a validated measure for pain; 2) functional change; and 3) global assessment made by the patient. Secondary measures, which will be included providing there is sufficient data, are: 1) side effects of treatment; 2) quality of life measures; and 3) medical utilization.
Search methods for identification of studies
Identification of published studies:
a. We will search the Cochrane Complementary Medicine Field Registry of randomized clinical trials and controlled clinical trials. This registry, contains approximately 3500 studies, which have been retrieved from MEDLINE, EMBASE, Acubase (a French‐based acupuncture database), Dissertation Abstracts, other electronic complementary medicine databases and hand searches.
b. MEDLINE will be searched (1966‐ September 1998) using the optimally‐sensitive search strategy (Dickersin, 1997) for retrieving all randomized trials. We will couple this search with the following keywords:'arthritis', 'osteoarthritis', 'rheumatology', 'knee pain', 'hip pain', 'joint pain,' 'oriental medicine,' 'Traditional Chinese Medicine,' 'meridians,' 'acupuncture,' and 'osteoarthrosis'
c. The CAMPAIN (Complementary and Alternative Medicine and Pain) database which includes about 9000 citations (not limited to RCTs or CCTs) from assorted electronic and hand searched sources, will be searched using the keywords 'arthritis', 'osteoarthritis', and 'acupuncture'.
d. Aculars (an acronym for Acupuncture Literature Analysis and Retrieval System) which contains approximately 40,000 citations from more than 500 biomedical journals and related conferences, published in China and other countries since 1984, will also be searched, using the same keywords described in B.
e. The Science Citation Index will be searched for additional studies, once we have identified studies to be included in the review, from the already described sources. We will search for articles which have cited one or more of the included studies in order to identify other relevant studies.
f. Bibliographies from retrieved articles will be searched for additional studies.
2. Identification of unpublished studies: In order to minimize the impact of publication bias, we will actively seek conference abstracts and unpublished trials. Potential publication bias will be evaluated with inverted funnel plot techniques.
a. Acupuncture journals which are neither indexed electronically in one of the above databases, nor presently being hand searched by the Complementary Medicine Field, will be identified and hand searched. Additionally, serial conference proceedings with the word(s) arthritis, rheumatology, rheumatism, musculoskeletal, pain or acupuncture in the conference title, will be hand searched. Also, an inquiry will be sent to the investigators of included studies and the research directors of acupuncture schools requesting information of additional trials.
b. Conferences from the following Rheumatology Associations will be hand searched:
American College of Rheumatology
British Society of Rheumatology
Canadian Rheumatology Association
c. Available conference proceedings from known Acupuncture Associations will be hand searched.
d. Additional Conference Abstracts and Posters for hand searching:
Proceedings from the World Congresses on Pain
Pain in Osteoarthritis International Medical Conference in: Seminars in arthritis and rheumatism v 18(4) Suppl 2.
International Pain Symposium (Atlanta: 1992)
Advances in pain research and therapy
e. Rheumatology journals to be hand searched
Journal of Rheumatology
Arthritis and Rheumatism
British Journal of Rheumatology
Clinical and Experimental Rheumatology.
Data collection and analysis
Data Extraction
Data will be extracted independently by two trained reviewers (MH, JE). Inter‐rater agreement between reviewers will be assessed using the kappa statistic. Disagreements on methodological quality ratings will be discussed by the two reviewers to reach a consensus.
Further patient data to be extracted and presented in an evidence table will include: details on the patients' basic disease, duration of disease, age and gender of patients, reason and number of withdrawals and drop‐outs, year(s) of the study, country of study, manner of recruitment (e.g., media advertisements, outpatient clinic, inpatient pain clinic) and adverse effects. Although these are not internal validity issues, per se, they may be used to investigate heterogeneity among studies, or serve as predictors of outcomes. Acupuncture treatment data to be extracted will include: the type of needle manipulation used (e.g., manual, low frequency electrical, high frequency electrical), total number of treatments, frequency of treatments, duration of treatment, whether or not de chi was elicited (and in the case of sham acupuncture, whether de chi was confined to the treatment group), type of acupuncture (e.g., traditional Chinese acupuncture, five‐element acupuncture, auriculotherapy), number of points used, location of points used, and the training and experience of the acupuncturist in treating osteoarthritis of the knee.
Assessment of the quality of acupuncture procedure
The adequacy of the acupuncture treatment will be assessed in two ways. First, acupuncture textbooks will be reviewed using a method previously cited for other types of chronic pain (Birch 1997). This method ascertains whether there are common denominators across textbooks for the acceptable acupuncture treatment for osteoarthritis of the knee. By this standard, the acupuncture treatments in the studies will be rated as "adequate" or "inadequate." Secondly, two acupuncturists with prior experience in treating OA will be given the description of the study population and the acupuncture procedure, and will be blinded to the results of the study. The clinicians will be asked, "Given your clinical experience would you rate the acupuncture treatment "adequate" or "not adequate" for this population?" Kappa statistics will be calculated between panelists.
Statistical Analysis
A fixed effects model will be used, approximate chi square tests for heterogeneity will be used to assess outcome data for compatibility with the assumption of a uniform risk ratio (p>0.10). In the event that there is significant (p>0.10) heterogeneity among studies, we will use a random effects model (der Simonian and Laird, 1986) and perform appropriate subgroup analysis. Separate assessment according to the type of arthritic joint i.e., hip, knee, hand (joints of the spine are excluded) will be performed as part of the primary analysis, and results from the separate joint assessments will be combined if results are similar across joints. Appropriate subgroup analyses may include length of follow up time, type of needle manipulation of acupuncture procedure (various manipulations are known to stimulate different endorphin/monoamine cascades), the total number of treatments received, and/or the overall adequacy of the acupuncture treatment (discussed in the next section).
All trials included in the systematic review will be entered into Review Manager 4.0.4 (RevMan 4.0.4). For pain measures, standardized effects sizes will be estimated (change divided by the baseline standard deviation of the sample). The results for each intervention group will be weighted by the sample size of the group. Dichotomous results will include dropouts and withdrawals as non responders (intention‐to‐treat analysis) and will be summarized as odds ratios (ORs). The summary OR will be obtained by weighting each individual OR by the inverse of the variance of the estimate for each trial. Because overall effects will only be estimated for groups of trials using the same type of reference group (e.g., sham acupuncture, NSAIDS, steroid injection), it is likely that more than one meta‐analysis will be performed.
A sensitivity analysis will be conducted to evaluate the robustness of the meta‐analysis. This analysis will examine the effects of methodological quality on study outcomes.
