Search J Rheum

Advanced Search


Current Issue


Guidelines for Authors

Classified Ads


Search PubMed


Subscriber Registration

Guidelines for Website Users

JRheum Update Service

Contact Info


Still a Blind Spot for Osteoporosis Prevention and Treatment for Rheumatoid Arthritis

To the Editor:

In the September issue of The Journal, Drs. Zochling and March commented1 on our editorial2 in which we noted the absence of recommendations for the prevention and treatment of osteoporosis in patients with rheumatoid arthritis (RA).

They referred to the American College of Rheumatology Subcommittee on Rheumatoid Arthritis Guidelines update 20023, which, they state, addressed the issue of osteoporosis prevention for patients with RA. However, these guidelines contain only the following sentence within the text of the article: "RA is associated with an increased risk of osteoporosis independently of glucocorticoid therapy." There is a recommendation for bone mineral densitometry as well as vitamin D and calcium supplement only for RA patients who are taking glucocorticosteroid. But these recently published guidelines, like the 2 we reviewed4,5, as well as the recent consensus statement on pharmacological management of early RA6, failed to deal with the need for osteoporosis prevention in RA patients in general. Only the recent article of Haugeberg, et al7 also raised the question of identifying RA patients at high risk for osteoporosis, identifying 5 criteria: age, weight, inflammation, immobility, and ever-use of corticosteroids.

The role of corticosteroids in RA was the principal focus of Drs. Zochling and March, and this was well described in this Guidelines update3. However, every patient undergoing glucocorticoid treatment needs bone mineral densitometry evaluation and consideration of treatment for osteoporosis, as recently reiterated by Saag, et al8, so this is not specific to patients with RA.

This unfortunately reinforces our contention that the prevention and treatment of osteoporosis in patients with RA continues to be insufficiently addressed in published guidelines for management of RA. Our colleagues who work on the osteoporosis side are doing better — as evidenced by the Guidelines on Osteoporosis Management recently published by the Osteoporosis Society of Canada9, which list RA as a risk factor for the development of osteoporosis.

BRIGITTE M. JOLLES, MD, MSc, Hôpital Orthopédique de la Suisse Romande, Centre Hospitalier Universitaire Vaudois, University of Lausanne, 4 Avenue Pierre Decker, 1005 Lausanne, Switzerland. (E-mail:; EARL R. BOGOCH, MD, MSc, FRCSC, Department of Surgery, University of Toronto, Martin Family Centre for Arthritis Care, Mobility Program, St. Michael's Hospital, Suite 800, 55 Queen Street East, Toronto, Ontario M5C 1R6, Canada (E-mail:


1. Zochling J, March L. First do no harm — a bone of contention in rheumatoid arthritis [editorial]. J Rheumatol 2002;29:1818-20.

2. Jolles BM, Bogoch ER. Current consensus recommendations for rheumatoid arthritis therapy: a blind spot for osteoporosis prevention and treatment [editorial]. J Rheumatol 2002;29:1814-7.

3. Anonymous. Guidelines for the management of rheumatoid arthritis: 2002 Update. Arthritis Rheum 2002;46:328-46.

4. Emery P. Overview of current therapies for rheumatoid arthritis. J Rheumatol 2001;28 Suppl 62:1-3.

5. Wolfe F, Cush JJ, O'Dell JR, et al. Consensus recommendations for the assessment and treatment of rheumatoid arthritis. J Rheumatol 2001;28:1423-30.

6. Emery P. Is it time for a European consensus on the pharmacological management of early RA? J Rheumatol 2002;29 Suppl 66:1-2.

7. Haugeberg G, Orstavik RE, Uhlig T, Falch JA, Halse JI, Kvien TK. Clinical decision rules in rheumatoid arthritis: do they identify patients at high risk for osteoporosis? Testing clinical criteria in a population based cohort of patients with rheumatoid arthritis recruited from the Oslo Rheumatoid Arthritis Register. Ann Rheum Dis 2002;61:1085-9.

8. Saag KG, Pisu M. Balancing bones and bucks among new glucocorticoid users [editorial]. J Rheumatol 2003;30:1-3.

9. Brown JP, Josse RG. 2002 clinical practice guidelines for the diagnosis and management of osteoporosis in Canada. CMAJ 2002;167:S1-34.

Dr. Zochling and Dr. March reply

To the Editor:

Drs. Jolles and Bogoch1 have highlighted an important shortcoming of current rheumatoid arthritis (RA) guidelines, in the lack of direction for prevention and management of associated osteoporosis. We agree wholeheartedly, and suggest it is perhaps a result more of the lack of good scientific evidence on which to base such guidelines than a lack of recognition. Little is known about the natural history of bone loss and associated fracture risk in steroid-naive RA. Corticosteroid therapy is renowned for causing bone loss in other disease states, but in RA the association of osteoporosis and disease activity complicates the picture.

It may well be that with more aggressive use of disease modifying agents and biologic therapies, improved control of disease activity will bring with it an opportunity to more adequately assess the progression of osteoporosis in RA independent of corticosteroids in prospective studies, and to define the most appropriate management strategies.

Until that time, any guidelines must make it clear they are based on evidence that is incomplete. We agree that the management of any patients with RA should include assessment of the risk of osteoporosis, including underlying disease activity, demographics, and potential corticosteroid use.

JANE ZOCHLING, MBBS, FRACP, MMed (ClinEpi), Research Fellow in Rheumatology, Institute of Bone and Joint Research, Department of Rheumatology, Royal North Shore Hospital, Pacific Highway, St. Leonards, NSW, Australia 2065 (E-mail:; LYN MARCH, MBBS, MSc, PhD, FRACP, FAFPHM, Associate Professor of Medicine, Senior Staff Specialist in Rheumatology and Clinical Epidemiology, Institute of Bone and Joint Research, Department of Rheumatology, Royal North Shore Hospital (E-mail:


1. Jolles BM, Bogoch ER. Current consensus recommendations for rheumatoid arthritis therapy: a blind spot for osteoporosis prevention and treatment. J Rheumatol 2002; 29:1814-7.

Fatigue and Psychological States and Traits in Systemic Lupus Erythematosus: Association or Causation?

To the Editor:

We read with interest the article by Omdal and colleagues, with results from a study of psychological characteristics of fatigued patients with systemic lupus erythematosus (SLE). Applying a cross-sectional design, they found that fatigue, as measured by the Fatigue Severity Score, was correlated with measures of depression, hysteria, anxiety, and social dysfunctioning1. The authors have previously reported a lack of association between fatigue and biological markers of disease activity in the same population2. They conclude that psychological factors like response and adaptation to a chronic disease appear to be the most important determinant of fatigue, and the high prevalence of fatigue is probably caused by personality traits common in patients with SLE.

The study seems nicely performed, and the applied statistical methods seem appropriate. The authors, however, seem to mistake correlation with causation.

The correlation between fatigue and symptoms of depression is based on the Beck Depression Inventory (BDI). Fatigue is one of the key symptoms of depression and fatigue/energy questions are therefore embedded in most diagnostic instruments. Consequently, a fatigued population will necessarily score higher on a depression scale, clinically depressed or not. It has been suggested that the only way to overcome this methodological problem is to discharge fatigue/energy items from the depression inventory3. The authors themselves deliver an argument for this point of view, since no correlation was found with the "depression" factor in the General Health Questionnaire (GHQ-30) and the Fatigue Severity Score. The GHQ-30, and consequently the depression factor derived from GHQ-30, does not include any fatigue/energy questions4.

The reported correlation between fatigue and hysteria is based on the Hysteria-axis in the Multiphasic Personality Inventory (MMPI). The MMPI claims to detect stable personality traits. However, as Creed and Ash point out, the MMPI includes several items that rheumatologists commonly attribute to disease rather than change in psychological status; for example, "I have few or no pain"5. With respect to fatigue, the instrument includes a significant number of fatigue/energy items (Table 1). Again, any fatigued population would score higher on this scale, disregarding present psychopathology or not. It has been suggested that the MMPI for this reason should not be used in patients with rheumatological diseases because "their elevated score on some dimensions, e.g., hypochondriasis, depression and hysteria, may reflect the disease rather than psychopathology"6.

Table 1. Examples of fatigue-related items in the "Hysteria-axis" from the MMPI. Numbers refer to item numbers in MMPI-2.

Many aspects of the etiology and pathogenesis of SLE still remain to be solved, and no single valid biomarker for disease activity or organ involvement has been found. Gladman, et al compared disease activity in SLE patients with 5 health status instruments7: the Health Assessment Questionnaire, Functional Ability Index, the Fatigue Severity Scale, the Disability Days Measure, the Centre for Epidemiological Studies-Depression Scale, and the Medical Outcomes Study Short Form Health Survey (SF-20) during their clinic visit. Disease activity was measured using the SLE Disease Activity Index (SLEDAI). Inter-instrument correlation analysis revealed that components of the SF-20 correlated significantly with each of the other instruments used, while there was no correlation between any of the health status instruments used and the SLEDAI. They suggested that health status assessment as measured by the SF-20 is a valid independent outcome measure in patients with SLE. It could also be argued that the lack of correlation between self-reported health and objective measures of disease activity disclose the inability of existing clinical markers to detect disease burden as it is perceived by the patient.

We agree with the authors that fatigue in chronic disease is a complex phenomenon, but with the lack of valid biomarkers it is important to ask if cross-sectional studies add to our understanding of causal and modifying factors for the development and course of fatigue in SLE and other chronic diseases.

NIELS HENRIK HJOLLUND, MD, PhD; OLE NØRBY HANSEN, MSc, PhD, Department of Occupational Medicine, Herning Central Hospital, Gl. Landevj 61, DK-7400 Herning, Denmark. E-mail:


1. Omdal R, Waterloo K, Koldingsnes W, Husby G, Mellgren SI. Fatigue in patients with systemic lupus erythematosus: the psychosocial aspects. J Rheumatol 2003;30:283-7.

2. Omdal R, Mellgren SI, Koldingsnes W, Jacobsen EA, Husby G. Fatigue in patients with systemic lupus erythematosus: lack of associations to serum cytokines, antiphospholipid antibodies, or other disease characteristics. J Rheumatol 2002;29:482-6.

3. Visser MR, Smets EM. Fatigue, depression and quality of life in cancer patients: how are they related? Support Care Cancer 1998;6:101-8.

4. Huppert FA, Walters DE, Day NE, Elliott BJ. The factor structure of the General Health Questionnaire (GHQ-30). A reliability study on 6317 community residents. Br J Psychiatry 1989;155:178-85.

5. Creed FH, Ash G. Depression in rheumatoid arthritis: aetiology and treatment. Int Rev Psychiatry 1992;4:23-34.

6. Bowling A. Measuring disease: a review of disease-specific quality of life measurement scales. 2nd ed. Buckingham, UK: Open University Press; 2001.

7. Gladman DD, Urowitz MB, Ong A, Gough J, MacKinnon A. A comparison of five health status instruments in patients with systemic lupus erythematosus. Lupus 1996;5:190-5.

Dr. Omdal, et al reply

To the Editor:

We thank Drs. Hjollund and Hansen for their interest in our report. Their letter contains important views on the topic of fatigue in autoimmune diseases, and is written with insight concerning fatigue and psychology. With a few exceptions, we have no problem agreeing with most of their arguments, but would like to comment on them.

Considering mistaking correlation with causation, we cannot quite understand that the message from our study should be interpreted as that. It is clear from the text that this is a cross-sectional study, and it should also be emphasized that we use the word association throughout1. This word does not imply any specific causation, only that some variables covariate: correlation — as Hjollund and Hansen call it. Having said that, the procedure for unveiling the mechanisms of fatigue in these diseases would include prospective studies, intervention studies, and search for biomarkers of the fatigue phenomenon, as emphasized by Hjollund and Hansen.

Their comments regarding fatigue and depression are important. Mutually related factors may cause "circular argumentations" leading to conclusions of associations that do not exist. This is a well known phenomenon, but sometimes hard to exclude since the associations may be unknown, or the researcher unaware of relationships among the variables.

Hjollund and Hansen argue that the Beck Depression Inventory (BDI) contains several items related to fatigue, while the General Health Questionnaire (GHQ-30) does not. They also argue that we find no association between fatigue and the depression factor (Factor C) in the GHQ-30, while such a relationship is evident in the BDI.

Factor C in the GHQ-30 is significantly associated to fatigue (R2 = 0.24, p = 0.0001) by simple regression analysis. It is only in a stepwise regression model that Factor C does not significantly contribute to fatigue. This does not imply lack of association between fatigue and the items of Factor C, but indicates that Factor C does not independently contribute enough to the association with fatigue to be maintained in a multivariate model1. To some extent, we can therefore agree with Hjollund and Hansen's argument.

Further, the BDI is a 21 item questionnaire pertaining to cognitive/emotional and somatic manifestations of depression. The first 13 items assess cognitive/affective symptoms of depression; the final 8 items evaluate somatic phenomena of depression. Our study also showed an association between the cognitive/affective items of the BDI and fatigue, indicating that it is not the fatigue/energy items of the BDI that are the most important aspects of depression for patients with SLE.

It has been suggested2 that the Multiphasic Personality Inventory (MMPI) should not be used in patients with rheumatological diseases. This may be true for arthritis and other joint-related diseases, but our patients with SLE are characterized by few or no joint manifestations and have features more related to systemic disease. In this context, we see no reason for not applying the MMPI instrument in SLE.

Also, the Hysteria-axis in the MMPI-2 includes 60 items. Almost all items on Scale 3 are scores on other clinical scales; only 10 items are unique to Scale 3. Some of the items deal with a general denial of physical health and some specific somatic complaints. Another group of items involves a rather general denial of psychological or emotional problems and of discomfort in social situations. Although these 2 clusters of items are reasonably independent in normal subjects, persons displaying hysterical defenses seem to score high on both clusters3,4. Indeed, it is not possible to obtain a T score above 65 on Scale 3 without endorsing both kinds of items. Patients with bona fide medical problems lacking any evident psychological component tend to obtain T scores of about 55­60 on this scale3,4.

However, Hjollund and Hansen's concern about the BDI and MMPI-2 versus fatigue reflects to a considerable extent the problem of mutually related variables — a matter in which we fully agree.

Finally, Hjollund and Hansen refer to the study by Gladman, et al5 where 5 health status instruments in SLE were compared by correlation analysis, and suggest that health status assessment as measured by the Medical Outcomes Study Short Form (SF-20) is a valid independent outcome measure in SLE. Applying correlation statistics for comparing different instruments (measuring agreement between 2 methods) is in our opinion questionable. A more adequate approach is estimation of limits of agreement6. The results of such a comparison between the health instruments are unknown, as far as we know. Whether the SF-20 is superior in evaluating fatigue in patients with SLE therefore remains to be seen.

ROALD OMDAL, MD, PhD, Clinical Immunology Unit, Rogaland Central Hospital, N­4068 Stavanger, Norway; KNUT WATERLOO, PsyD, Department of Neurology; WENCHE KOLDINGSNES, MD, Department of Rheumatology, University Hospital of North Norway; GUNNAR HUSBY, MD, PhD, Department of Rheumatology, National Hospital, Oslo; SVEIN MELLGREN, MD, PhD, Department of Neurology, University Hospital of North Norway; BJØRN STRAUME, MD, Institute of Community Medicine, University of Tromsø. E-mail:


1. Omdal R, Waterloo K, Koldingsnes W, Husby G, Mellgren SI. Fatigue in patients with systemic lupus erythematosus: The psychosocial aspects. J Rheumatol 2003;30:283-7.

2. Bowling A. Measuring disease: a review of disease-specific quality of life measurement scales. 2nd ed. Buckingham, UK: Open University Press; 2001.

3. Graham JR. MMPI-2. Assessing personality and psychopathology. New York: Oxford University Press; 1990.

4. Greene RL. The MMPI-2/MMPI: An interpretive manual. Boston: Allyn and Bacon; 1991.

5. Gladman DD, Urowitz MB, Ong A, Gough J, MacKinnon A. A comparison of five health status instruments in patients with systemic lupus erythematosus. Lupus 1996;5:190-5.

6. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307-10.

Hands-on Treatment in Rheumatology

To the Editor:

The editorial by Fitzcharles1 regarding "hands-on" treatments in rheumatology is timely and balanced. She noted the seemingly low frequency of side effects of spinal manipulation and wonders whether this is due to underreporting or the rarity of complications.

We recently surveyed all UK neurologists asking them to note all neurological adverse effects of spinal manipulation seen within the last year2. Our response rate was 74% and 35 cases of often serious complications were reported. None of these had been published in the medical literature. It follows that, in our case, underreporting was exactly 100%. I therefore agree with Fitzcharles that (chiropractic) manipulation (particularly of the upper spine) is associated with serious complications of unknown frequency. The incidence rates reported by chiropractors are pure speculation and, in view of huge underreporting, even nonsensical. The inescapable conclusion is that we need conclusive incidence figures. Until they are available, caution seems well advised.

EDZARD ERNST, MD, PhD, FRCP, FRCPEd, Director, Complementary Medicine, Peninsula Medical School, Universities of Exeter and Plymouth, 25 Victoria Park Road, Exeter, EX2 4NT, UK. E-mail:


1. Fitzcharles M-A. Is it time for rheumatologists to rethink the use of manual therapies? [editorial]. J Rheumatol 2002;29:1117-20.

2. Stevinson C, Honan W, Cooke B, Ernst E. Neurological complications of cervical spine manipulation. J Roy Soc Med 2001;94:107-10.

Dr. Fitzcharles replies

To the Editor:

The comments of Prof. Ernst are appreciated. Once again caution in the use of manipulation therapies is advocated. Good medical practice requires that a physician should be fully aware of risks associated with a treatment before prescription. We do not currently have accurate information regarding risks related to "hands-on" or manipulation therapies. Prof. Ernst and colleagues have recently demonstrated the high rate of underreporting of neurological events following spinal manipulation1. This survey of physician habits, which is likely reflective of usual practice, raises further concerns about the true frequency of side effects due to any treatment without a formal monitoring procedure. The literature abounds with anecdote, case reports, and polls from physicians, none sufficient to give reliable estimates of risk.

Attempts at systematic review or metaanalysis of manipulation therapies concede that the quality of clinical trials, mostly examining efficacy, but also reporting on risks, is generally poor. Risk of harm is clearly of importance to prescribers. Although a recent metaanalysis reported that spinal manipulation was no better than other commonly used treatments for low back pain, advantages or risks associated with various treatments were not addressed2. In a population-based study examining the relationship of stroke and chiropractic manipulation, the rates for chiropractic visits were similar in the preceding year for both patients and the population controls. However, differences only emerged when the groups were divided according to age of 45 years, with younger stroke patients reporting more chiropractic visits3.

Once again there is a strong call for rigorous scientific evidence before these treatments can be universally endorsed.

MARY-ANN FITZCHARLES, MB, ChB, MRCPUK, FRCPC, Associate Professor of Medicine, Division of Rheumatology, Montreal General Hospital, McGill University Health Centre, 1650 Cedar Avenue, Montreal, Quebec H3G 1A4, Canada


1. Stevinson C, Honan W, Cooke B, Ernst E. Neurological complications of cervical spine manipulation. J Roy Soc Med 2001;94:107-10.

2. Assendelft WJJ, Morton SC, Yu EI, Suttorp MJ, Shekelle PG. Spinal manipulative therapy for low back pain. A meta-analysis of effectiveness relative to other therapies. Ann Intern Med 2003;138:871-81.

3. Rothwell DM, Bondy SJ, Williams JI. Chiropractic manipulation and stroke. A population-based case-control study. Stroke 2001;32:1054-9.

Another Look At Wegener's Granulomatosis-Associated Pachymeningitis

To the Editor:

Fam, et al recently described an unusual manifestation of Wegner's granulomatosis (WG) with cranial pachymeningitis and reviewed the literature1. We read their report with great interest, as we had reported 3 cases of severe central nervous system (CNS) manifestations in generalized WG and reviewed the literature ourselves2. In contrast to Fam, et al, we found 18 patients with meningeal involvement, all proven by biopsy. Yet Fam, et al did not mention that 9 out of 20 patients (45%) with meningeal involvement and known antinuclear cytoplasmic antibodies (ANCA) results were found to be ANCA negative. Although the sensitivity for cANCA/PR3-ANCA approaches almost 100% in acute generalized WG, there are published reports of about 12 patients with generalized WG and negative ANCA in the English literature. On top of that, out of these 12 patients with ANCA-negative WG, 83% (10 patients) had cerebral and/or meningeal involvement.

Fam, et al did not point out that according to the literature, 83% of patients with WG and cerebral involvement are persistently ANCA-negative versus just 10% with "classic" WG, as shown in larger cohorts. In addition to broader awareness of this data, further studies and other initiatives are needed to distinguish this subset of ANCA-negative WG with predominant CNS involvement from others.

PEER MALTE ARIES, MD; EVA REINHOLD-KELLER, MD; WILHELM LUDWIG GROSS, MD, Universitätsklinikum Schleswig Holstein, Poliklinik für Rheumatologie, Rheumaklinik Bad Bramstedt, Oskar Alexander Strasse 26, 24576 Bad Bramstedt, Germany. E-mail:


1. Fam AG, Lavine E, Lee L, Perez-Ordonez B, Goyal M, Cranial pachymeningitis: An unusual manifestation of Wegner´s granulomatosis. J Rheumatol 2003;30:2070-4.

2. Reinhold-Keller E, de Groot K, Holl-Ulrich K, et al. Severe CNS manifestations as the clinical hallmark in generalized Wegner's granulomatosis consistently negative for antineutrophil cytoplasmic antibodies. A report of 3 cases and a review of the literature. Clin Exp Rheumatol 2001;19:541-9.

Dr. Fam replies

To the Editor:

We appreciate the comments by Aries, et al and their interest in our study1. In its classic generalized form, WG chiefly affects the upper and lower respiratory tracts and kidneys. In the limited or partial form, the pathological findings of necrotizing granulomatous vasculitis are similar but the upper respiratory tract, orbit, or lung are primarily affected, in the absence of renal disease1.

Measurement of serum cytoplasmic antineutrophil cytoplasmic antibody (cANCA/PR3-ANCA) is highly specific for WG, with a sensitivity greater than 90% for active generalized WG, but only 67% for those with active limited disease. Thus, ANCA is absent in one-third of patients with limited WG1-3.

Reinhold-Keller, et al4 described 3 patients with active, ANCA-negative WG and severe neurologic manifestations: leptomeningitis in 2 and cerebrospinal lesions in one. However, the absence of renal disease in all 3 patients raises some question whether these subjects had limited rather than generalized WG4. Review of reported patients with WG-associated meningeal disease revealed that 45% of cases were ANCA-negative4.

Our study, which focused on meningeal inflammation in WG, showed that most reported cases occurred early (within 6 months of onset) in the course of active, limited WG, and that about one-third of patients were ANCA-negative1. In accord with other studies, our patient developed pachymeningitis in the setting of active, limited WG. ANCA was repeatedly negative initially, but she subsequently developed pANCA/myeloperoxidase (MPO) antibodies, which disappeared following successful immunosuppressive therapy for WG1. A possible association between limited WG, pANCA/MPO antibodies, and pachymeningitis has recently been described by Japanese investigators5.

Thus an association between active limited WG and pachymeningitis with variable ANCA results, but a trend toward cANCA negativity and pANCA positivity, is proposed but remains unproven. To confirm these clinical observations, further case-control studies of this rare manifestation of WG, and new insights into the pathogenetic roles of cANCA and pANCA antibodies, are required. Greater awareness, early recognition, and timely therapy of WG-associated pachymeningitis are important to minimize permanent neurologic damage.

ADEL G. FAM, MD, FRCPC, Division of Rheumatology, Sunnybrook and Women's Health Sciences Centre, University of Toronto, 2075 Bayview Avenue, Room M1-402, Toronto, Ontario M4N 3M5, Canada. E-mail:


1. Fam AG, Lavine E, Lee L, Perez-Ordonez B, Goyal M. Cranial pachymeningitis: an unusual manifestation of Wegener's granulomatosis. J Rheumatol 2003;30:2070-4.

2. Rao JK, Weinberger M, Oddone EZ, et al. The role of antineutrophil cytoplasmic antibody (C-ANCA) testing in the diagnosis of Wegener granulomatosis. A literature review and meta-analysis. Ann Intern Med 1995;123:925-32.

3. Hoffman GS, Specks U. Antineutrophil cytoplasmic antibodies. Arthritis Rheum 1998;41:1521-37.

4. Reinhold-Keller E, de Groot K, Holl-Ulrich K, et al. Severe CNS manifestations as the clinical hallmark in generalized Wegener's granulomatosis consistently negative for antineutrophil cytoplasmic antibodies. A report of 3 cases and a review of the literature. Clin Exp Rheumatol 2001;19:541-9.

5. Nagashima T, Maguchi S, Terayama Y, et al. P-ANCA-positive Wegener's granulomatosis presenting with hypertrophic pachymeningitis and multiple cranial neuropathies: case report and review of literature. Neuropathology 2000;20:23-30.

A Randomized, Double Blind, Placebo Controlled Trial of a Topical Cream Containing Glucosamine Sulfate, Chondroitin Sulfate, and Camphor for Osteoarthritis of the Knee

To the Editor:

I'm confused how the fatal methodological flaws in this study1 were not uncovered in editorial review, and that this study was allowed to be published.

The most glaring problem, the "treatment" group used a topical product that contained a known active ingredient, camphor, while this was absent from the placebo group. This does not support the author's conclusion that "Topical application of glucosamine and chondroitin sulfate is effective in relieving the pain from OA of the knee." The proper conclusion should be that camphor was effective in short term pain relief, or better yet, no conclusion could be drawn on glucosamine and chondroitin because the "active" group contained 3 variables compared to the placebo-treated group.

This is not the only flaw. Subjects in the study were apparently allowed to apply treatment or placebo "ad lib" and were not required to "dab" the product, as is often the standard in study of topical agents. The act of rubbing the joint alone may help reduce symptoms. The camphor-containing "active" product may have been used more often by the subjects, further biasing the results. Finally, camphor has a burning sensation not found with peppermint oil. The absence of camphor in the placebo group could have unblinded the trial.

There is no evidence that glucosamine and chondroitin are absorbed in the skin, and certainly none that shows any significant levels are achieved by topical application. Further, the active product contained only a small quantity of glucosamine and chondroitin. We know from oral studies that a minimum threshold concentration is required for effect.

JASON THEODOSAKIS, MD, MS, MPH, FACPM, Steering Oversight Committee Member, NIH Trial on Glucosamine and Chondroitin, Assistant Professor, University of Arizona, Tucson, Arizona, USA.

Dr. Theodosakis's activities and relationships with regards to companies and products related to the treatment of osteoarthritis are the following: Celebrex, Pfizer Pharmaceuticals (speaker); Hyalgan, Sanofi-Synthelabo (consultant); glucosamine/chondroitin and Avocado-Soy Unsaponifiables, NBTY (consultant); SAMe, glucosamine/chondroitin, Pharmavite (consultant); anti-inflammatory Philodendron extract, Next Pharmaceutical (research advisor).


1. Cohen M, Wolfe R, Mai T, Lewis D. A randomized, double blind, placebo controlled trial of a topical cream containing glucosamine sulfate, chondroitin sulfate, and camphor for osteoarthritis of the knee. J Rheumatol 2003;30:523-8.

To the Editor:

I have some major concerns regarding Cohen, et al's article on the benefits of topical glucosamine sulfate1. How did this article slip through your peer review process with the words "double blind" in the title?

The "active" cream contains not only glucosamine and chondroitin sulfate, but also peppermint oil and camphor. The placebo was a "simple cosmetic cream that used conventional skin emollients, petrolatum and mineral oil." Unless the experimental subjects were anosmic, they would be aware which cream they were using; indeed, if they applied it shortly before being assessed, the experimenters would also be unblinded. At best this study was single blind; at worst it is completely unblinded by the failure to match the odor of the placebo to that of the "active" cream.

MATTHEW L. GROVE, BSc, MRCP, Consultant Rheumatologist, North Tyneside Hospital, North Shields, Tyne and Wear, NE29 8NH, United Kingdom.


1. Cohen M, Wolfe R, Mai T, Lewis D. A randomized, double blind, placebo controlled trial of a topical cream containing glucosamine sulfate, chondroitin sulfate, and camphor for osteoarthritis of the knee. J Rheumatol 2003;30:523-8.

Dr. Cohen replies

To the Editor:

The letters from Drs. Theodosakis and Grove raise a number of issues concerning our study1, including issues to do with the study blinding as well as the administration of the intervention and the conclusions drawn regarding the contribution of glucosamine and chondroitin to the therapeutic results.

Regarding the study being de-blinded, this is certainly not the case. The active and placebo creams both contained peppermint oil, which effectively matched them with regard to odor. Further, while there may have been a slight difference between the creams due to the presence of camphor in the active preparation, both the subjects and investigators were naive to any difference, and the subjects did not bring their creams to clinic visits and did not meet in the waiting room, thereby ensuring that no contamination occurred. Further, blinding was formally checked with subjects being asked at each visit which group they thought they had been allocated to. The data from these assessments indicate that blinding was maintained (see Table 1).

Table 1. Treatment allocation versus subjects' assessment of allocation.

It is unlikely that the method of application influenced the results of this study. Having the subjects rub rather than dab the cream may have contributed to the observed pain reduction seen in the placebo group; however, this effect would not have led to the observed difference between the groups. Over the course of the trial the active group used a total of 5.5 tubes, while the placebo group used 5.7 tubes.

Finally, the contribution of glucosamine and chondroitin to the observed pain reduction in the active group is open to speculation and cannot be determined by the current study. Unfortunately, the doses of these ingredients were overstated in the original article by a factor of 10. Thus the active intervention actually contained glucosamine sulfate (3.0 mg/g), chondroitin sulfate (7.2 mg/g), and shark cartilage (14 mg/g), of which 10-30% is chondroitin sulfate, along with camphor (32 mg/g) and peppermint oil (9 mg/g).

The effective dose of glucosamine and chondroitin delivered to a painful joint is uncertain. Based on animal studies it is evident that even though there is active uptake of glucosamine by articular cartilage, only around 0.4% of an administered oral dose is delivered to cartilage tissue2. Extrapolating this to humans would equate to approximately 6 mg of a 1500 mg oral dose. This is in the range delivered by topical administration in our study, and the observed gradual and progressive improvement is consistent with studies using oral administration of these agents. Thus while the study title refers to glucosamine, chondroitin, and camphor, the exact mechanisms of action of the cream used in our study are yet to be determined, and it is premature to discount the therapeutic effect of topical glucosamine and chondroitin. The study certainly does provide evidence that the active cream containing camphor, glucosamine, and chondroitin is safe and effective in reducing the pain from osteoarthritis of the knee. The relative contribution of these ingredients is now the subject of ongoing research.

MARC COHEN, MD, Head, Department of Complementary Medicine, RMIT University, PO Box 71, Bundoora, Victoria 3083, Australia. E-mail:


1. Cohen M, Wolfe R, Mai T, Lewis D. A randomized, double blind, placebo controlled trial of a topical cream containing glucosamine sulfate, chondroitin sulfate, and camphor for osteoarthritis of the knee. J Rheumatol 2003;30:523-8.

2. Setnikar I, Giacchetti C, Zanolo G. Pharmacokinetics of glucosamine in the dog and in man. Arzneimittelforschung 1986;36:729-35.

Return to April 2004 Table of Contents

© 2004. The Journal of Rheumatology Publishing Company Limited.
All rights reserved.