To critique the available homeopathic research on musculoskeletal disorders a literature review was undertaken with searches conducted on databases Cinahl, AMED, Embase and Cochrane using terminology including, ‘homeopathy’, ‘musculo-skeletal’, ‘musculoskeletal’, ‘skeletal’, ‘arthritis’, ‘joints’ and ‘muscle’ or combinations of these words. This literature review found a range of research trials, systematic reviews and literature relating to homeopathy and musculoskeletal complaints. This report will give evidence of why research into homeopathy in musculoskeletal disorders is important; examine the findings of 3 relevant systematic reviews and critically examine 9 homeopathic musculoskeletal trials found in more detail (see appendix 1). An overview of common problems found in the research designs used and the impact these designs have on the current state of knowledge in this field will then be discussed.
Why musculoskeletal disorders are important
Over the last decade the average life expectancy has risen by almost 20 years, however, even though much public attention is focused upon fatal diseases, chronic musculoskeletal and rheumatic diseases are a major cause of morbidity in the world today (World Health Organisation [WHO], 2003). Persistent pain associated with musculoskeletal disorders often increases with age, and morbidity factors include pain, inflammation, depression, anxiety, physical disability, sleep disturbance and impaired cognitive function (Weiner & Ernst, 2004). These factors have a major impact on quality of life and create substantial economic burden on health systems (WHO, 2003). Approximately one tenth of all General Practitioner consultations are related to musculoskeletal complaints, impacting indirect costs, as one fifth of incapacity claims sought in Great Britain are for musculoskeletal disorders (Clarke & Symmons, 2006).
Common medications used to conventionally treat musculoskeletal disorders such as arthritic pain are nonsteroidal anti-inflammatory drugs (NSAIDs) or opiods prescribed as analgesics. However, whilst these drugs may mask the pain experienced by patients, the chronicity of the complaint usually requires medication for a long period of time, which often leads to adverse drug reactions (Weiner et al., 2004). Thus patients often seek other forms of alternative or complementary treatments to improve their quality of life. This move to alternative treatment is illustrated by the patient benefit survey, conducted at the Department of Homeopathic Medicine in Liverpool, where the largest diagnostic group treated were for musculoskeletal conditions, representing 261 of the 1100 completed questionnaires (23.7%) (Richardson, 2001). A similar survey conducted by the Tunbridge Wells Homeopathic Hospital also lists musculoskeletal conditions as the second highest diagnostic group treated, accounting for 13% of its patients surveyed (Clover, 2000). Therefore, given that musculoskeletal complaints are highly prevalent in modern society with many patients seeking alternative treatment to manage their condition, it is important that we understand the available evidence for how effective homeopathy is in treating the symptoms associated with musculoskeletal disorders.
Summary and critique of the current state of knowledge
Soeken (2004) who examines systematic reviews of CAM therapies for arthritis-related pain cites one review of homeopathy and rheumatic disease by Jonas, Linde & Ramirez (2000). This review included 6 trials using either random assignment or double-blinding. Three trials were on rheumatoid arthritis (RA) and 3 other trials on osteoarthritis (OA), fibromyalgia, and myalgia. Soeken (2004) states that, using the Jadad scale, 5 of the 6 trials rated as high quality, but the interventions varied as individualized (classical) homeopathy, where the physician selects a remedy based on each patient’s individual symptoms, and complex homeopathy (where one or more remedies are given for a clinical condition) were used. Meta-analysis of these trials shows homeopathy to be twice as effective as placebo although the number of studies is small to reach a definitive conclusion (Soeken, 2004).
However, Kleignen, Knipschild & Riet (1991), reviewed 107 clinical trials of homeopathy generally, stating that most trials conducted were of low quality but that there were many exceptions. Trials were scored out of 100 for methodological soundness with points awarded for criteria including:
· Patient characteristics described
· Number of patients analysed
· Intervention well described
· Double blinding
· Effect measurement relevant and described
· Clarity of results presentation enabling results to be checked by readers.
Of the 107 trials examined by Kleignen et al. (1991) only 6 investigated rheumatological disease, with 4 showing positive and 2 showing negative results. However, generally the methodological scores were low, ranging from 33 to 50 (see details in appendix 2).
Kleignen et al. (1991) states, only 14 trials used individualised classical homeopathy, 58 used one homeopathic remedy for a comparable diagnosis, 26 trials used complex homeopathy and 9 trials used isopathy. This is an important factor in homeopathic research generally, as many classical homeopaths would not count isopathy, complex homeopathy or using one homeopathic remedy for a diagnosed complaint as the practice of homeopathy. Classical (individualised) homeopathy, based on the principle that ‘like treats like’, requires the homeopath to select the most similar homeopathic remedy to the totality of symptoms presented by each individual patient (Hahnemann, 1996). Thus it is questionable if these non-classical interventions used in clinical trials have clinical significance, as it does not represent conventional practice.
Long & Ernst (2001) conducted a systematic review on the homeopathic treatment of OA. In this, 4 randomized controlled trials (RCTs) met the inclusion criteria and were described to be of high average methodological quality, with 2 of the studies giving a positive result in the efficacy of homeopathic treatment in OA. Overall this review concluded that the research data shows homeopathy to be more effective than placebo, but that the small number of studies again prevents firm conclusions being made (Weiner et al., 2004).
Three trials were found assessing if homeopathy can reduce delayed onset muscle soreness (DOMS). Vickers, Fisher, Smith, Wyllie & Lewith (1997) conducted a randomised, double blind placebo controlled trial (RDBPCT), in which 58 subjects undertook a bench stepping exercise test and then the treatment group took a combination of potentised (30c) arnica, rhus tox and sarcolactic acid. Participants were then asked to complete an outcome form for the next 5 days. Homeopathy was shown to be no more effective than placebo (Vickers et al., 1997). However, the trial was flawed as: the sample size was too small to reach statistical significance; there were big variations in participant muscle soreness scores which could effect comparisons of the groups, due to many variations in individual fitness and lifestyle post or prior to the exercise test; and the homeopathic remedy was not individualised for the subjects, thus not reflecting the practice of homeopathy.
A second RDBPCT conducted by Vickers, Fisher, Smith, Wyllie and Rees (1998) examined the effect of arnica 30x compared to placebo on DOMS following long distance running. Vickers et al. (1998) concluded that arnica 30x is ineffective for DOMS. This trial had a larger sample size, but subjects ran in one of 114 different races of varying distances, with the difficulty of the races not being specified. There were also no exclusion criteria. These variables, coupled with the potentially differing fitness levels / characteristics of participants could have a significant impact on muscle soreness scores and thus the outcome for participants. This point is highlighted as the trial showed that muscle soreness was higher the greater the distance ran (Vikers et al., 1998). Again the treatment was not individualized in remedy or potency, thus not reflecting homeopathic practice and potentially leading to negative results due to ineffective medication of the treatment group.
Tveiten, Bruset, Borchgrevink & Norseth (1998) conducted another RDBPCT examining the effect of potentised arnica (D30) on muscle soreness in Oslo marathon runners. This trial concluded that arnica did have a statistically significant effect on lowering muscle soreness immediately after the marathon, but there was no difference between the treatment and placebo group 3 days after the marathon, thus showing that arnica does not seem to shorten the recovery time of marathon runners (Tveiten et al., 1998). Tveiten et al.’s trial does not contain as many race variables (i.e. in distance and difficulty) as the Vikers et al. trial (1998), but there are still many uncontrolled variables in the population characteristics and their pre and post race activities, such as diet and exercise levels that could effect the trial’s results. Once again this research does not reflect classical homeopathic practice, as it is not based on individualised homeopathy.
Sajedi, Alizad, Alaeddini, Fatemi & Mazaherinezhad (2008) conducted a RDBPCT to examine the effect of individualized homeopathic treatment on the muscle tone of 24 children with spastic cerebral palsy over a 4 month period. Sajedi et al. state that whilst some children on homeopathic treatment showed some positive effect on motor development criteria, there was no evidence to show a reduction of muscle spasticity (2008). Whilst this trial did replicate the practice of classical homeopathy in terms of individualised remedy selection, there was a very small sample size and a short space of time to obtain results, to give conclusive evidence. In addition, the fact that some children did show some signs of motor development also indicates that with homeopathic treatment, a positive effect may be happening in participants, but not necessarily where the trial objectives and measures are set. With individualised classical homeopathy, the homeopath is basing a prescription on the symptom totality of the patient (Hahnemann, 1996), not necessarily a specific tissue or organ focused on in a trial. Therefore, this presents a problem with researching the effects of homeopathy using RDBPCTs when often only specific quantifiable outcomes are measured. Perhaps more qualitative outcomes should be included in homeopathy trials to examine the patient effect as a whole.
Fisher & Scott (2001) undertook a cross-over, RDBPCT to test if individualised homeopathy is effective in reducing joint inflammation in 112 patients with rheumatoid arthritis who were also using either NSAIDs or disease modifying anti-rheumatic drugs (DMARDs). 58 patients completed the trial. This study gives clear details regarding methodology, randomisation, drop-out rates, lists the 42 remedies available to the homeopath for selection (in potencies of 6c or 30c) and states the cross-over of treatment groups after a 3 month period. Fisher et al. (2001) concluded that there is no evidence that homeopathy reduces the symptoms of RA, which contradicts Gibson, Gibson, MacNeill & Buchanan’s (1980) trial findings that are discussed below. However, another problematic point in homeopathic research when a cross-over method is used, is that how do we know how long a homeopathic remedy is active for in a patient? This question makes it difficult to know when it would be appropriate to switch the participant group’s treatment over in a cross-over trial, as the homeopathic remedy could still be active when the treatment is switched to placebo. This point is also highlighted by Gibson et al. (1980) who explains that no clear-cut information can be gained from putting patients who had been on homeopathic treatment, on to placebo for further study in a trial.
Gibson et al. (1980) conducted a double-blind RCT that compared 23 people with RA on anti-inflammatory drugs and individualised homeopathic treatment to 23 people with RA on anti-inflammatory drugs and placebo for a 3 month period. This research found that there was, “…a significant improvement in subjective pain, articular index, stiffness and grip strength in those patients receiving homeopathic remedies” (Gibson et al., 1980, p.453) compared to placebo. In this study patients were also grouped depending upon if they were considered to have strong or weak prescribing symptoms. It was found that patients with strong prescribing symptoms improved more than those with weak symptoms. This result could be due to the homeopath finding it easier to make a more accurate prescription sooner for the patients presenting with a clearer symptom picture. This highlights another significant problem with conducting RCTs with individualised homeopathy, as the trial relies on the effective prescribing skills of the homeopaths deciding upon the remedies to be used for individual patients.
This trial also sought to correct faults in execution and design from a previous study conducted by Gibson, Gibson, MacNeill, Gray, Dick & Buchanan in 1978 (Turner, 1988). Although the Gibson et al. (1980) trial still only scores 40 out of 100 for methodology in Kleignen et al.’s, (1991) review of homeopathy clinical trials (see appendix 2). This relatively low score may be due to the fact that: remedy potencies were not displayed; patients were not randomly allocated to groups, although group allocation was made independently by a third physician; and there were only a relatively low number of patients analysed to reach a good score in Kleignen et al.’s review (1991). These factors also highlight the fact that systematic reviews of clinical trials can often be quite subjective depending on the reviewer’s scoring protocol and philosophical stance as to what constitutes good research.
Shealy, Thomlinson, Cox & Borgmeyer (1998) compared a potentised combination of rhus tox 12x, causticum 12x and lac vaccinum 30x to acetaminophen for the treatment of OA knee pain in a double blind RCT. Shealy et al. (1998) showed that 55% of homeopathy patients achieved pain relief of 40% or greater compared to 38% of acetaminophen patients achieving pain relief of 40% or greater, however this result did not reach statistical significance and it unfortunately did not use individualised homeopathy.
However, Bell, Lewis, Brooks, Schwartz, Lewis, Walsh & Baldwin (2004) did use individualised homeopathy in a RDBPCT to assess the efficacy of homeopathic treatment for fibromyalgia. Bell et al. state, “Fifty-three people completed the treatment protocol. Participants on active treatment showed significantly greater improvements in tender point count and tender point pain, quality of life, global health and a trend towards less depression compared with those on placebo” (2004, p.577). Bell et al. (2004) recognised some shortcomings of other previous homeopathic studies within its methodology by:
· Recruiting two homeopaths to jointly interview patients to agree on remedy selection. This reduces the risk of selecting an incorrect / ineffective remedy for the patients.
· Giving clear explanations on treatment, inclusion / exclusion criteria, drop-out rates, measures used, blinding and randomization protocols and statistical results. Helping to facilitate easy trial replication in the future.
· Using remedies in LM potency to minimize any anti-doting effect on homeopathic treatment.
· Using individualised prescription that is more representative of classical homeopathic practice.
· Conducting the trial over 3 months instead of 1. Although this may still be a relatively short period of time in assessment of a chronic condition.
· Measurement of global outcomes as well as disease specific outcomes, to show an overall effect of homeopathic treatment.
However, although Bell et al. (2004) negated many problems of previous studies, it was noted that the sample size was still too small to give statistical significance to outcome measures other than tender point pain and there was a lack of objective measures directly relating to patient’s fibromyalgia status.
Fisher (1986) conducted a double blind, placebo controlled trial using one of three remedies (arnica, bryonia or rhus tox) in a 6c potency to treat the symptoms of fibrositis, with measurements on pain, the number of tender spots and sleep. These remedies were selected as they were thought to cover the symptom picture of fibrositis in most cases. Each patient was also scored on how well they fitted the prescribed remedy. Fisher concluded that, “…homeopathy produced a statistically significant improvement, but only when the prescribed remedy was well indicated.” (1986, p.142). This result could indicate that trials using individualised homeopathy could elicit more realistic results. However, where the condition being investigated contains many symptoms and many possible remedy options available to the prescriber, prescribing errors could still influence results negatively.
Common problems in research designs
From the reviews and trials examined above, it can be seen that the quality of reporting, methodologies used and the results obtained differ immensely. Many trials have used one remedy to treat a specific condition for all participants, which does not mirror homeopathic practice in reality and could lead to negative results, as the sample population will invariably not all have the symptom totality for the one remedy included within the trial, or require the same potency as provided within the trial. The same problem is shown in trials using complex homeopathy, as a ‘one size fits all’ approach to prescribing is not considered homeopathic prescribing based upon the law of similars (Hahnemann, 1996). The sample sizes required to reach statistical significance also seem to be a recurring problem in musculoskeletal trials, illustrating that recruitment is problematic and that the piloting of trials is essential to avoid some of these pitfalls.
However, trials using individualised prescribing methods also have pitfalls, as incorrect prescriptions can be made, which could lead to ineffective treatment and a negative outcome. Although reliability could be improved using several prescribing homeopaths. Prescribing is made even more difficult as double blinded, placebo controls complicate normal practice conditions, as homeopaths will not know if a patient is failing to respond because they have selected the wrong remedy or if they are on the placebo treatment (Linde & Melchart, 1998).
In addition, placebo patients, who still undergo the consultation experience, could experience some form of healing response from the consultation alone, which could decrease the difference between the control and treatment groups, skewing results. (Linde et al, 1998). Milgrom, who has investigated the complexities of the patient, practitioner and remedy relationship in homeopathy and the difficulties that this presents to researchers, states that it is better to, “…avoid the ideological trench warfare of double-blind placebo-controlled trials and discover new ways of experimentally getting to grips with homeopathy.” (2003, p.13). Thus it becomes apparent that if we do not fully understand the mechanism of action in homeopathy, it becomes very difficult to know how to investigate it reliably, as there are many possible variables.
Most homeopathic trials on musculoskeletal conditions use specific outcome measures focused on a diagnosed condition. Therefore, it may be beneficial to set research objectives that look at both disease specific outcomes and the overall / global aspects of the patient, as demonstrated in the Bell et al. (2004) trial. These additional measures are perhaps more appropriate, as homeopathy treats individuals as a whole and not the diagnosed disease. Perhaps this could be why patient benefit surveys, based on the patient’s own assessment of treatment outcome, (shown in appendices 3 & 4) show such a high level of improvement. This would also have more relevance for people deciding if they want to try homeopathic treatment as an alternative to conventional medicine.
In addition, focus on quantitative outcome measures for diagnosed conditions is possibly due to homeopathic researchers wishing to stem the flow of criticism aimed at homeopathy, as this creates pressure to justify homeopathy within a hostile scientific environment by providing more evidence in the form of RDBPCTs, rather than to research to find new knowledge or for self-critical evaluation (Linde et al., 1998).
Impact of research designs and on the current state of knowledge
“Recommendations for homeopathic care itself can hardly be extrapolated from the clinical trials, as they seem rarely to reflect what happens in everyday practice.” (Linde et al., 1998, p.385)
Unfortunately if research designs are of poor quality and do not reflect everyday practice as highlighted in many trials above, the results present an unrealistic picture of homeopathy to the public, homeopathic, medical and scientific communities. Additionally, if trials achieve negative results, it becomes easy for critics to present articles, which may not consider any bias, quality or design problems in the research conducted, thus presenting an unbalanced picture of evidence to the reader.
In summary, the 9 trials reviewed assessing the efficacy of homeopathy in treating musculoskeletal conditions show 5 positive and 4 negative results in favour of homeopathy. However, the quality of reporting clarity, design, methodology, sample size, types of measures and intervention used has been variable, with many trials not having much clinical relevance, as they do not represent classical homeopathic practice in reality. Nevertheless, there have been some trials of reasonable methodological quality, using individualised homeopathy and a range of disease specific and global outcome measures to show that there is evidence to suggest that homeopathy is effective in some musculoskeletal complaints. There are still hurdles to overcome in resolving how to reliably research individualised homeopathy using RDBPCT, when we are still unsure of homeopathy’s mechanism of action. Thus the current state of knowledge in homeopathy’s success in treating musculoskeletal conditions is generally positive, but not completely certain in all cases, as improvements to the quality and design of research is required.
Article written by:
Sharon Townhill, BSc. RSHom.