Dextrose Prolotherapy for Knee Osteoarthritis: A Randomized Controlled Trial

After a long gestation, including five years of recruitment, it is good to read the results of the second randomised trial of dextrose/lignocaine prolotherapy injections for osteoarthritis (OA) of the knee in the May/June edition of this journal. The investigators had the ambition and foresight to include both a saline injection control group to address the question of attributable effect of dextrose/lignocaine and a non-injection exercise group to test compare with usual care. This ambitious agenda made recruitment challenging, hence the protracted recruitment period to reach their predetermined sample size. There was no 'wait and see' to give a comparison with the natural history of the osteoarthritis over the year-long follow-up period, but would have made recruitment even more difficult.

The trial seems to have been rigorously conducted using validated outcome measures and reporting good retention rates. A strong majority of participants in all groups were happy to recommend their respective treatments to others. Blinding to the injection solution in the injection groups was successful. The investigators used a comprehensive injection protocol involving three to five treatments with the injection of up to 22.5 mls of 15% dextrose/0.4% lignocaine or saline solution into tender extra-articular structures using a peppering technique and a 6 ml intra-articular injection of 25% dextrose/0.4% lignocaine or saline solution. Notably, the exercise group were given web-based exercise program with telephone contact, but not in-person contact at the same intervals as injection treatments. The primary outcome was the well validated WOMAC knee OA scale. This provided a composite score of the pain, stiffness, and function subscales. Secondary outcomes included pain severity and frequency from the knee pain scale. They showed statistically significant and clinically important improvements in WOMAC composite scores and pain and function subscales for the dextrose/lignocaine group from baseline to 24 and 52 weeks of around 25%. Unlike an analogous study on prolotherapy for chronic low back pain, the differences between dextrose/lignocaine and saline at these points were statistically significant, although they did not reach the threshold of clinical importance.1 The corresponding reductions in pain scores for dextrose/lignocaine of 0.92 from baseline to 24 and 52 weeks achieved statistical significance on a 0-5 pain scale. This represents a 37% reduction in pain levels, well above the commonly quoted figure of 30% for clinically significant reductions in chronic pain and similar to those reported for low back pain.1 Interestingly saline injections had a very similar effect to exercise, but within the exercise group, there was no correlation between exercise compliance and WOMAC composite improvements at 52 weeks. The authors went some way to testing the generalizability and reproducibility of the results for the dextrose/lignocaine injections by demonstrating almost identical results over 12 months in a parallel cohort of 36 participants who declined participation in the randomised trial.3 Comparison with other treatments across studies is difficult due to differing selection criteria and outcome measures. This applies to the much more conservative injection protocol used in the previous, less rigorous knee OA study with intra-articular injections of dextrose/lignocaine versus lignocaine alone. In this trial, at 6 months, both treatments showed significant improvements in pain, subjective swelling and goniometric knee measures, however only the dextrose- lignocaine group showed a superior response with analysis of all measures combined and a reduction in the number of knee-buckling episodes.2 So the dextrose/lignocaine component of the injections seems to have an attributable effect but it is not a magic cure for patients who have had symptomatic knee OA pain for 5-10 years. Prolotherapy also seems to be more effective than an exercise regime. When applying the results in clinical practice, it should be noted that the cohort had an average duration of pain of 5-10 years, was younger than in some other studies and was lacking in patients with the most severe forms of osteoarthritis and those requiring chronic opioids for their OA. Patients should be told to expect mild to moderate post- injection pain for 3-5 days.

Through persistence and attention to detail, the study investigators have made an important contribution to the growing literature on prolotherapy for chronic musculoskeletal pain. They have provided us with a good estimate of both the efficacy and the effectiveness of prolotherapy for knee OA that should aid both practitioners and patients in difficult choices of the myriad of treatments promoted for this condition.

1. Yelland MJ, Glasziou PP, Bogduk N, Schluter P, McKernon M. Prolotherapy injections, saline injections, and exercises for chronic low-back pain: a randomized trial. Spine. 2004;29(1):9-16.
2. Reeves KD, Hassanein K. Randomized prospective double-blind placebo-controlled study of dextrose prolotherapy for knee osteoarthritis with or without ACL laxity. Alternative Therapies in Health and Medicine. 2000;6(2):68-70,2-4,7-80.
3. Rabago D, Zgierska A, Fortney L, Kijowski R, Mundt M, Ryan M, et al. Hypertonic dextrose injections (prolotherapy) for knee osteoarthritis: results of a single-arm uncontrolled study with 1-year follow-up. Journal of alternative and complementary medicine. 2012;18(4):408-14. Epub 2012/04/21.

Competing interests: The author of this commentary has published a paper with the lead author on prolotherapy for lateral epicondylaglia and is currently a coinvestigator with the lead author on a randomised trial of prolotherapy and exercises for lateral epicondylalgia and on three small grants that support this trial.

Response to Beth A Fox

We appreciate Dr. Fox's positive impression of our study. We acknowledge that there are relatively few RCTs assessing prolotherapy. However, the number is growing and the data are promising, as noted in Distel and Best (2011) and cited in Dr. Fox's comments above.

It is appropriate in the early assessment of a therapy to argue for "more rigorous scientific trials demonstrating significant improvement over extended periods" before "extensive use can be advocated;" we do the same in the penultimate paragraph of the paper. However, we also consider prolotherapy to be appropriate care for selected patients. In clinical life, optimal care exists in an evolving environment. We must often make choices based on best available, not best imaginable, evidence.

One way to judge the evidence is the Strength of Recommendation Taxonomy (SORT) criteria, developed by several prominent evidence based primary care journals.1 SORT identifies the current study's quality as level 1 ("High quality individual RCT") and the overall level of recommendation as B ("inconsistent [read 'few'] studies of good quality using patient-oriented evidence"). Evidence-based therapies for symptomatic knee OA are few and identifying optimal treatment can be challenging. For example, the American Academy of Orthopedic Surgeons has recently published knee OA treatment recommendations (http://www.aaos.org/research/guidelines/OAKSummaryofRecommendations.pdf). Among therapies that the AAOS is not able to recommend based on inconclusive evidence, or evidence suggesting ineffectiveness, are acupuncture, manual therapy, valgus force bracing, lateral wedge insoles, glucosamine and chondroitin, acetaminophen, opioids, pain patches, intra- articular corticosteroid injections, hyaluronic acid injections, growth factor injections and/or platelet rich plasma, needle lavage, arthroscopy and other surgical procedures. Among therapies recommended by the AAOS are physical therapy and exercise, and NSAIDS, two therapies that most of our participants failed prior to enrollment.

Some patients will be refractory to conservative care; they still merit our best treatment efforts. The data for prolotherapy in recent small but rigorous RCTs for tendinopathies2-4 and now OA suggest both safety and efficacy for a variety of conditions. Given these findings, and the limited conservative treatment options for knee OA, we stand by our conclusions that prolotherapy may be an appropriate therapy for selected patients. Further study will inform its application to more general populations.

1. Ebell MH, Siwek J, Weiss BD, Woolf SH, Susman J, Ewigman B, et al. Strength of recommendation taxonomy (SORT): a patient-centered approach to grading evidence in the medical literature. Am Fam Physician 2004;69:548- 56.

2. Scarpone M, Rabago D, Zgierska A, Arbogest J, Snell ED. The efficacy of prolotherapy for lateral epicondylosis: a pilot study. Clin J Sport Med 2008;18:248-54.

3. Yelland MJ, Sweeting KR, Lyftogt JA, Ng SK, Scuffham PA. Prolotherapy injections and eccentric loading exercises for painful Achilles tendinosis: a randomised trial. Br. J. Sports Med 2011;45:421-8.

4. Rabago D, Lee KS, Ryan M, Chourasia AO, Sesto ME, Zgierska AE, et al. Hypertonic dextrose and morrhuate sodium injections (prolotherapy) for lateral epicondylosis (tennis elbow): Results of a single-blind, pilot- level randomized controlled trial. Am J Phys Med Rehab 2013;E-pub ahead of print PMID: 23291605.

Competing interests: None declared

Prolotherapy for injured or painful musculoskeletal problems has been around for many years with the use of various injectants but responses described in the literature were varied and studies were of poor quality. It is hypothesized that prolotherapy encourages the stimulation and proliferation of growth factors, such as platelet-derived growth factor, basic fibroblast growth factor, and transforming growth factor-beta.1 The therapy has been utilized in tendon injuries or diseases, muscular sports injuries, and for the treatment of osteoarthritis particularly of the back, and less so of the hand and the knee.1 2 There has been an increasing interest in finding alternate therapies especially when traditional treatments fail to relieve pain or are unacceptable to patients. Injectants of recent interest include autologous conditioned serum, dextrose, and platelet-rich plasma.3 4 Most studies involving dextrose have lacked scientific rigor and several have not demonstrated significant improvement. Most of the convincing evidence has been demonstrated in the treatment of tendonopathies but that therapy still suffers from lack of standardization of injection technique, lack of large randomized controlled trials, and lack of extended follow- up.1 2 The current study is one of only a few randomized controlled trials evaluating dextrose prolotherapy for knee osteoarthritis symptoms. These results in knee osteoarthritis are promising but more rigorous scientific trials demonstrating significant improvement over extended periods are needed before extensive use can be advocated.

References:
1.Distel LM and Best TM. Prolotherapy: a clinical review of its role in treating chronic musculoskeletal pain. PM R. 2011;3:S78-S81.
2.Rabago D, Slattengren A, and Zgierska A. Prolotherapy in primary care practice. Prim Care. 2010;37(1):65-80.
3.Sampson S, Gerhardt M, and Mandelbaum. Platelet rich plasma injection grafts for musculoskeletal injuries: a review. Curr Rev musculoskelet Med. 2008;1:165-174.
4.Fox BA and Stephens MM. Treatment of knee osteoarthritis with orthokine? -derived autologous conditioned serum. Expert Rev Clin Immunol. 2010;6(3):335-345.

Competing interests:   None declared

We appreciate Dr. Finnegan's response to our RCT and respect his having opinions regarding prolotherapy. However, 1) we disagree with his view regarding the outcome measures in our study and 2) our experience regarding prolotherapy and the community of physicians who practice prolotherapy differs markedly from his.

Clinically, our university-based clinical and teaching experience over the past 30 years suggests that the practice of prolotherapy by a trained physician is safe, clinically useful and of potentially high impact for appropriately selected patients.

But individual experiences (both ours and his) do not carry substantive weight in the advancement of the scientific understanding of prolotherapy. Over the years, we have sought to understand prolotherapy scientifically through the conduct of well-designed studies; in addition to the referenced RCT, we have conducted systematic reviews, animal model studies and pilot-level clinical trials as noted in the reference section.

Because our clinical experience and our prior studies have been promising, we undertook a rigorous, 3-arm, NIH-funded, blinded RCT with strong scientific measures that include the most relevant outcomes - those that matter to patients, anchored by the well-validated Western Ontario McMaster University Osteoarthritis Index (WOMAC). The WOMAC is recommended as an essential outcome measure by the Osteoarthritis Research Society International among many other relevant medical societies, and has been used hundreds of times in validation studies, comparative studies against other health status measures, and in clinical trials and clinical practice settings. The WOMAC, not functional, biomarker or other surrogate measures, is the gold standard in this context. Our strong study design is necessary to overcome the biases that can occur in our individual clinical experience. The findings, that prolotherapy showed biological efficacy and pragmatic effectiveness compared with blinded injection and exercise control therapies respectively, are important and useful because they address many of the methodological weaknesses of prior studies, and certainly the limitations of individual experience. Future planned trials will indeed include functional measures; however their absence here does not diminish the usefulness of these findings. While no single study is perfect, we feel the study adds to the body of scientific knowledge on prolotherapy and stand by our conclusions.

Respectfully

David Rabago MD

Competing interests: None declared

Dear Editor:

Your recent publication of the article by Dr. Rabago et al in the May/June 2013 Annals of Family Medicine represents a good try, but a frightening potential for misrepresenting the issues exists.

The article was entitled Dextrose Prolotherapy for Knee Osteoarthritis: A Randomized Controlled Trial.

On the surface, it is a reasonable effort to be both randomized and controlled; however, all the parameters for measuring outcomes are purely subjective or quasi subjective at most. Simply put, no fully objective criteria were utilized!

As an orthopedic surgeon with four decades experience, I have had the privilege (perhaps "privilege" is overly-complimentary) of following the course of prolotherapy since the mid-1970s, and it continues to scare me, since there has never been one (or more!) convincingly-objective study to document its efficacy.

In a recent letter (yet unpublished) which I wrote to an editor of an orthopedic journal, I stated that the simple fact is that prolotherapy in some form has been around for over 60 years, and still has no recognition as being an effective intervention. I was referencing lower back pain, and my opinion was well-supported by an outstanding article in the first issue of the Spine Journal this year.

As an aside, the editor replied to me directly and stated he had never used prolotherapy in any of his patients and still did not recommend it, but we unfortunately have a situation where certain physicians and their PA's are using it on a frequent basis, even with the lack of appropriate studies.

My own experience with this modality began when I was an orthopedic resident at the University of Pennsylvania nearly four decades ago, at which time I became aware of some deceptive tomfoolery [i.e., prolotherapy] being conducted at a few suburban hospitals. Suffice it to say that this purportedly therapeutic intervention was a great money generator for the so-called physicians who used it, but it was a scam then, and it continues to be so today, until much further evidence is generated on a randomly- controlled, blinded, prospective basis, a situation which is most unlikely to occur in the near term, if ever.

The above sentiments have been further reinforced by my attendance at the recent annual meeting of the American Academy of Pain Medicine, where I saw/heard nothing to contradict these conclusions.

To propagate or perpetuate a myth such as this, notwithstanding the efforts of Dr. Rabago et al, represents a total disservice to all our physician colleagues at this time, but also to the lay public who are particularly well-informed these days and may also be misled as a consequence. If and when objective scientific evidence arises, I would be most happy to reassess and perhaps change my viewpoint, but right now I see it as a major misrepresentation to disseminate unsupported information to our colleagues and patients.

Very truly yours,

Walter J. Finnegan, MD, JD

Competing interests:   None declared