ARTHRITIS & RHEUMATISM Vol. 46, No. 11, November 2002, pp 3102–3114 © 2002, American College of Rheumatology LETTERS 1. American College of Rheumatology Subcommittee on Rheumatoid Arthritis Guidelines. Guidelines for the management of rheumatoid arthritis: 2002 update. Arthritis Rheum 2002;46:328–46. 2. Patrono C, Coller B, Dalen JE, FitzGerald GA, Fuster V, Gent M, et al. Platelet-active drugs: the relationships among dose, effectiveness, and side effects. Chest 2001;119 Suppl:39S–63S. 3. Bombardier C, Laine L, Reicin A, Shapiro D, Burgos-Vargas R, Davis B, et al, VIGOR Study Group. Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis. N Engl J Med 2000;343:1520–8. 4. Konstam MA, Weir MR, Reicin A, Shapiro D, Sperling RS, Barr E, et al. Cardiovascular thrombotic events in controlled, clinical trials of rofecoxib. Circulation 2001;104:2280–8. 5. Silverstein FE, Faich G, Goldstein JL, Simon LS, Pincus T, Whelton A, et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid arthritis: the CLASS study: a randomized controlled trial. JAMA 2000;284:1247–55. 6. Garcia Rodriguez LA, Varas C, Patrono C. Differential effects of aspirin and non-aspirin nonsteroidal antiinflammatory drugs in the primary prevention of myocardial infarction in postmenopausal women. Epidemiology 2000;11:382–7. 7. Catella-Lawson F, Reilly MP, Kapoor SC, Cucchiara AJ, DeMarco S, Tournier B, et al. Cyclooxygenase inhibitors and the antiplatelet effects of aspirin. N Engl J Med 2001;345:1809–17. 8. Garcia Rodriguez LA, Hernandez-Diaz S. The risk of upper gastrointestinal complications associated with nonsteroidal antiinflammatory drugs, glucocorticoids, acetaminophen, and combinations of these agents. Arthritis Res 2001;3:98–101. DOI 10.1002/art.10505 Aspirin antiplatelet therapy and nonsteroidal antiinflammatory drugs: comment on the 2002 update of the American College of Rheumatology Guidelines for the Management of Rheumatoid Arthritis To the Editor: In the 2002 update of the American College of Rheumatology Guidelines for the Management of Rheumatoid Arthritis (1), the authors stress the need for antiplatelet therapy with low-dose aspirin for patients at risk for cardiovascular disease who are taking selective cyclooxygenase-2 (COX-2) agents, reasoning that, unlike nonselective nonsteroidal antiinflammatory drugs (NSAIDs), the selective COX-2 inhibitors have no effect on platelet adhesion or aggregation. This seems to imply that antiplatelet therapy with aspirin is required only when selective COX-2 inhibitors are used. I believe nonselective NSAIDs neither ensure appropriate antiplatelet prophylaxis, nor do they appear better than selective COX-2 inhibitors for patients at cardiovascular risk who are receiving concomitant aspirin antiplatelet therapy. Unlike the potent and irreversible inhibition of platelet COX-1 induced by aspirin, the antiplatelet effect of NSAIDs is highly variable and depends on two factors: their potency for inhibiting COX-1 at pharmacologic doses and their half-life (2). Thus, neither NSAIDs with a short half-life, such as ibuprofen, nor those with a relatively low COX-1 inhibitory effect, such as diclofenac, provide efficient and sustained antiplatelet therapy. Only drugs with a long half-life and a potent COX-1 inhibitory effect, such as indobufen, flurbiprofen, and naproxen, have been suggested to be potentially useful as antiplatelet agents (2). This may explain the observation of a higher rate of thrombotic events with rofecoxib than with naproxen in a large trial (3), something not observed when rofecoxib or celecoxib is compared with other nonselective NSAIDs (4,5). However, due to the reversibility of the antiplatelet effect of NSAIDs, the cardioprotective effect of these drugs as shown in clinical trials, where compliance with the dose and treatment schedule is usually better, does not necessarily apply to clinical practice, where intermittent use by patients is common. Accordingly, epidemiologic studies have failed to demonstrate cardioprotective effects of NSAIDs, in contrast with aspirin (6). Furthermore, recently published data demonstrate that in patients taking aspirin, ibuprofen antagonizes the irreversible platelet inhibition induced by aspirin, whereas other drugs with lower COX-1 inhibitory effects, such as diclofenac or rofecoxib, do not interfere with aspirin’s effects (7). An additional matter of concern when using aspirin with NSAIDs is their synergistic action in terms of gastrointestinal toxicity, something that has not been sufficiently evaluated for COX-2 selective drugs (8). In conclusion, neither relying on the antiplatelet effect of NSAIDs in patients at cardiovascular risk, nor using NSAIDs instead of selective COX-2 inhibitors when low-dose aspirin is needed, can be recommended based on available evidence. DOI 10.1002/art.10707 Hidden hazards and practical problems: comment on the 2002 update of the American College of Rheumatology Guidelines for the Management of Rheumatoid Arthritis To the Editor: The remarkable proliferation of improved diagnostic tests, disease activity measurements, and effective yet potentially toxic therapies more than justifies the 2002 update of the American College of Rheumatology Guidelines for the Management of Rheumatoid Arthritis (American College of Rheumatology Subcommittee on Rheumatoid Arthritis Guidelines. Guidelines for the management of rheumatoid arthritis: 2002 update. Arthritis Rheum 2002;46:328–46). Although individuals may argue about some of the details of the guidelines, in general I believe members of our College will welcome and appreciate the comprehensive and organized document prepared by the Subcommittee on Rheumatoid Arthritis Guidelines. As a practicing physician in a small office, however, I am concerned about two things. The first concern arises from some of the language used in the sections devoted to documentation, for by now, not only have most rheumatologists read these guidelines, but so have most third-party payers and malpractice attorneys. When “guidelines” such as these are couched in terms like “should document” and “must assess” and are published under the imprimatur of the American College of Rheumatology (ACR), they may become much more than just guidelines; they may become requirements. One need only consider how the original intent of ACR classification criteria for various diseases has been misinter- José L. Pablos, MD Hospital 12 de Octubre Madrid, Spain 3102 LETTERS 3103 preted and misused, to understand the point I am trying to make. Second, I wonder if, as they wrote the documentation recommendations, the members of the Subcommittee considered their potential impact on the actual clinical practice of rheumatology. Had they done so, then along with the guidelines perhaps would have come practical suggestions to enable us to fulfill the Subcommittee’s documentation goals, for example, a template acceptable to Medicare which would sufficiently document the history, physical examination, joint count, initial and interval laboratory monitoring results, imaging reports, Health Assessment Questionnaire findings, global assessments, pain, fatigue, and stiffness scores, and impressions and plans, all within the constraints of the usual 1-hour consultation and 15–20-minute return visit times we have to devote to our patients. By the time such recommended documentation, as currently outlined in the guidelines, is completed to the satisfaction of the Subcommittee (not to mention the other interested parties), there will be no time left to actually talk to our patients about what is going on in their lives, the changing evolution and course of their disease, and the therapeutic changes we might suggest, and their ramifications. Documentation, while important, has become so cumbersome and burdensome that it impairs a physician’s ability to provide good medical care. Computerization of records is not necessarily the answer: most of the computerized records I receive from primary care physicians (which supposedly comply with current Medicare rules) are so burdened with repetitive negative and unchanging data as to make them unreadable. Perhaps teaching institutions and large groups have the time, money, and manpower to meet the standards implied by these guidelines. As for the rest of us (who now better understand how frustrated states feel when the federal government imposes unfunded mandates), we would appreciate the Subcommittee now turning its attention to formulating a practical way for us to comply with its goals, serve our patients, and avoid (or at least satisfactorily reply to) third-party audits while at the same time limiting our legal liabilities. Sidney R. Block, MD Bangor, ME DOI 10.1002/art.10506 Reply To the Editor: One of the benefits of the exercise of the creation, dissemination, and critical review of treatment guidelines is to highlight gaps in evidence and focus attention on areas where well-designed studies are needed to clarify areas of uncertainty. The points made by Dr. Pablos are well-taken and raise important issues surrounding the risks and benefits of treatment with selective and nonselective NSAIDs in rheumatoid arthritis (RA) patients and other patient populations (1). We would suggest that additional evidence, preferably from randomized controlled trials, is needed to address the following questions: 1) Does low-dose aspirin negate the gastrointestinal (GI)–protective effect of the coxibs over the nonselective NSAIDs, as suggested by data from the Celecoxib Long-Term Arthritis Safety Study (2)? Is this a class effect of the coxibs? 2) Do coxibs and/or nonselective NSAIDs negate the cardioprotective benefits of low-dose aspirin (3)? If so, which ones have this effect? 3) For older patients who need an NSAID for osteoarthritis (OA) or RA, and need low-dose aspirin for cardioprotection but are at higher risk for NSAIDinduced adverse GI events, is the safest regimen a combination of low-dose aspirin and/or a proton pump inhibitor with either a coxib or a nonselective NSAID? Data from longitudinal cohort studies may be needed to determine the following: 1) In clinical practice, what is the effectiveness (versus efficacy in clinical trials) of various nonselective NSAIDs in terms of cardioprotective benefits as compared with low-dose aspirin in patients receiving long-term NSAID treatment for OA or RA? 2) In clinical practice, what is the effectiveness (versus efficacy) of the coxibs versus nonselective NSAIDs in terms of greater GI safety. Is this a class effect for all coxibs? In response to the comments of Dr. Block, the Subcommittee on Rheumatoid Arthritis Guidelines agrees that increasing documentation requirements for each clinical encounter have placed new and onerous burdens on the rheumatology practitioner. However, these challenges are generated not by practice guidelines, but rather by larger economic forces. It is Medicare and private payers who have linked provider reimbursement to level of service coding and, more recently, to documentation of the clinical encounter. In an effort to assist members in meeting these requirements, the ACR has created written documentation templates, provided multiple training sessions in Evaluation and Management coding, and is in the process of evaluating new practice management technologies such as electronic medical records and handheld devices. Many of these newer technologies use software templates for structured data entry that are designed to calculate and suggest a level of service based on the documentation of the visit. Information about these activities is accessible through the ACR Web site and the Socioeconomic Affairs Department. While we are sympathetic with the points raised by Dr. Block, we believe practice guidelines cannot be compromised by concerns about their being misused by legal professionals. These guidelines are indeed just guidelines and should never be “requirements” that substitute for good clinical judgment. The primary purpose, as Dr. Block affirms, was to update the RA guidelines to address clinical changes in the treatment of the disease. C. Kent Kwoh, MD Larry G. Anderson, MD Jerry M. Greene, MD Dorothy A. Johnson, DNSc, FNP James R. O’Dell, MD Mark L. Robbins, MD, MPH W. Neal Roberts, Jr., MD Robert W. Simms, MD Robert A. Yood, MD American College of Rheumatology Subcommittee on Rheumatoid Arthritis Guidelines 1. Mukherjee D, Nissen SE, Topol EJ. Risk of cardiovascular events associated with selective COX-2 inhibitors. JAMA 2001;286:954–9. 2. Silverstein FE, Faich G, Goldstein JL, Simon LS, Pincus T, Whelton A, et al. Gastrointestinal toxicity with celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthritis and rheumatoid 3104 LETTERS arthritis: the CLASS study: a randomized controlled trial. JAMA 2000;284:1247–55. 3. Catella-Lawson F, Reilly MP, Kapoor SC, Cucchiara AJ, DeMarco S, Tournier B, et al. Cyclooxygenase inhibitors and the antiplatelet effects of aspirin. N Engl J Med 2001;345:1809–17. DOI 10.1002/art.10490 Intolerance to caffeine may increase susceptibility to rheumatoid arthritis: comment on the article by Mikuls et al To the Editor: In the interesting report by Mikuls et al (1) on the independent and positive association between decaffeinated coffee and rheumatoid arthritis (RA), the authors discuss various reasons for the association, all implicating decaffeinated coffee. However, they did not consider the alternative explanation that those who are intolerant to caffeine may be more susceptible to developing RA. Rather than decaffeinated coffee, it may be the need to use decaffeinated coffee that defines the group at risk. Caffeine increases cortisol levels (2). RA patients have deficient cortisol responses due to a poorly functioning hypothalamic–pituitary axis (3). They can be expected to be hypersensitive to cortisol, just as those with hypothyroidism are hypersensitive to thyroxine. Insomnia, hyperactivity, and other central nervous system manifestations of hypersensitivity to cortisol, a central nervous system stimulant, would encourage the use of decaffeinated beverages. This hypersensitivity is in accord with the observed hypersensitivity of RA patients to the therapeutic and toxic effects of corticosteroids (4). The authors’ findings are no less interesting with this alternative explanation, because it supports decreased resistance to stress, i.e., a deficient hypothalamic response to severe inflammation and other stressors, as the critical feature predisposing one to RA. This has been described as well in the Lewis rat, an animal model of RA. Diminished pituitary– hypothalamic axis function is observed in this model, in which a severe, chronic, proliferative, and erosive inflammatory polyarthritis develops after injection of streptococcal cell wall antigen (5). The condition cannot be induced in the histocompatible F344 rat, which has a normal hypopituitary– hypothalamic axis. Replacement doses of corticosteroids significantly suppress the severity of the arthritis. Thus, the authors of the present study may be identifying a subset that is not only prone to developing RA but one in which the disease may be averted or attenuated by replacement doses of steroids. So why does regular coffee, which increases cortisol levels, not decrease the severity of RA? Unfortunately, regular coffee also inhibits adenosine receptors (6). Adenosine, a sleep promoter (7), is also a powerful suppressor of inflammation (8,9); the inhibition of adenosine by coffee, although adding to the wakefulness induced by cortisol, would tend to counter its antiinflammatory action. The authors are to be commended for their long-term comprehensive study of coffee, which has opened the door to links between coffee, an everyday beverage, and RA. These links are fascinating and warrant further study. Hugh McGrath, Jr., MD Louisiana State University Medical Center New Orleans, LA 1. Mikuls TR, Cerhan JR, Criswell LA, Merlino L, Mudano AS, Burma M, et al. Coffee, tea, and caffeine consumption and risk of rheumatoid arthritis: results from the Iowa Women’s Health Study. Arthritis Rheum 2002;46:83–91. 2. Lovallo WR, Al’Absi M, Blick K, Whitsett TL, Wilson MF. Stress-like adrenocorticotropin responses to caffeine in young healthy men. Pharmacol Biochem Behav 1996;55:365–9. 3. Chikanza IC, Petrou P, Kingsley G, Chrousos G, Panayi GS. Defective hypothalamic response to immune and inflammatory stimuli in patients with rheumatoid arthritis. Arthritis Rheum 1992;35:1281–8. 4. Firestein GS. Etiology and pathogenesis of rheumatoid arthritis. In: Kelley WN, Harris ED, Ruddy S, Sledge CB, editors. Textbook of rheumatology. 5th ed. Philadelphia: WB Saunders; 1997. p. 851–97. 5. Sternberg EM, Hill JM, Chrousos GP, Kamilaris T, Listwak SJ, Gold PW, et al. Inflammatory mediator-induced hypothalamicpituitary-adrenal axis activation is defective in streptococcal cell wall arthritis-susceptible Lewis rats. Proc Natl Acad Sci U S A 1989;86:2374–8. 6. Radulovacki M, Miletich RS, Green RD. N6 (L-phenylisopropyl) adenosine (L-PHA) increases slow-wave sleep (S2) and decreases wakefulness in rats. Brain Res 1982;246:178–80. 7. Van Dongen HP, Price NJ, Mullington JM, Szuba MP, Kapoor SC, Dinges DF. Caffeine eliminates psychomotor vigilance deficits from sleep inertia. Sleep 2001;24:813–9. 8. Ohta A, Sitkovsky M. Role of G-protein-coupled adenosine receptors in downregulation of inflammation and protection from tissue damage. Nature 2001;414:916–20. 9. Cronstein BN, Montesinos MC, Weissmann G. Salicylates and sulfasalazine, but not glucocorticoids, inhibit leukocyte accumulation by an adenosine-dependent mechanism that is independent of inhibition of prostaglandin synthesis and p105 of NFB. Proc Natl Acad Sci U S A 1999;96:6377–81. DOI 10.1002/art.10491 Reply To the Editor: We thank Dr. McGrath for his interest in our report and compliment him on his novel hypothesis. His theory, that patients with a relative cortisol deficiency, who are perhaps more prone to developing rheumatoid arthritis (RA), may have a heightened sensitivity to cortisol and thus avoid caffeine intake due to its adrenal-stimulating effects, is provocative. Although his hypothesis may ultimately prove correct, our findings are not entirely consistent with this explanation. Specifically, we observed no association (relative risk ⫽ 0.94; 95% confidence interval 0.58–1.52 for highest quartile) between daily caffeine intake and the development of RA. If the proposed hypothesis were true, one might anticipate women who are less sensitive to the effects of caffeine to have a significantly lower risk for RA. Moreover, we observed only a modest inverse correlation between daily caffeine intake and decaffeinated coffee consumption (r ⫽ ⫺0.3), suggesting that caffeine intake and decaffeinated coffee drinking were not necessarily mutually exclusive behaviors in this cohort. Notwithstanding, our data may be underpowered to detect a protective association between caffeine intake and RA risk, if such an association truly exists. LETTERS 3105 The association of decaffeinated coffee use with RA, also reported by investigators from the Black Women’s Health Study (Formica MK, Palmer JR, Rosenberg L, McAlindon TE. Lifestyle factors associated with development of rheumatoid arthritis: results from the Black Women’s Health Study [abstract]. Arthritis Rheum 2001;44 Suppl 9:S376), represents a novel finding and, as suggested by Dr. McGrath, warrants further investigation to define the mechanism underlying this putative association. Ted R. Mikuls, MD, MSPH James R. Cerhan, MD, PhD Lindsey A. Criswell, MD, MSPH Linda A. Merlino, MSc Kenneth G. Saag, MD, MSc University of Alabama at Birmingham DOI 10.1002/art.10507 Importance of T cells in rheumatoid synovitis: comment on the review by Firestein and Zvaifler To the Editor: In 1990, Firestein and Zvaifler wrote a cogent editorial on the limited role of T cells in rheumatoid (RA) synovitis, which remains a landmark in our understanding of the disease (1). The sequel (2) reaffirms the evidence, but the conclusions fail on the central question: “why does she have RA and not he?” Firestein and Zvaifler show that rheumatoid synovitis is not due to a local T cell effector mechanism. They could add that there is little evidence for T cell autoreactivity in RA. Neither point, however, is incompatible with the evidence from HLA–DR4 studies that RA is a T cell–dependent adaptive immune response (3). A systemic T cell–dependent response can initiate distant inflammation through antibody. It is now generally accepted that immune complexes generate inflammation not through complement, but by inducing cytokine release via IgG Fc␥ receptor type III (Fc␥RIII) (4). Unlike autoreactive T cells, immune complexes small enough to traverse endothelium are abundant in RA, and Fc␥RIII is selectively expressed in synovium (5). Many rodent models of RA have now been shown to be antibody and Fc␥RIII dependent (6,7). That adjuvant used to boost an adaptive immune response may also have an early activating effect on synovial cells through Tolllike receptors is probably a red herring. Another red herring, as Firestein and Zvaifler effectively point out, is that synovial macrophage cytokine release tends to recruit bystander “Th1” cells. It is very difficult to explain the extraarticular manifestations of RA without immune complexes. The distribution of Fc␥RIII indicates that pericarditis, nodules, and synovitis should all have the same origin (5). Tissue-specific amplification mechanisms almost certainly apply in synovium and bone marrow (8), but it is difficult to justify two means of initiation where one will do. The flaw in the common assumption that T cell– dependent autoimmune responses require autoreactive T cells is more subtle. The most consistent immune response in RA, rheumatoid factor production, is almost certainly T cell depen- dent, but there is no T cell autoreactivity to IgG. This paradox is explained by the ability of rheumatoid factor B cells to obtain help from T cells recognizing foreign antigen (9,10). Evidence from other disorders also suggests that T cell responses to autoantigens may not determine corresponding autoantibody production (11,12). “Loss of T cell tolerance” may not apply. Stastny not only confirmed the role of genes in RA (3), but also pointed out that the other factors may be not environmental, but stochastic (random). Genes, and probably smoking, set thresholds, but the kinetics of disease initiation are of a sequence of internal stochastic events. Studies of monozygotic twins suggest that a rate-limiting event occurs approximately once in 200 woman-years (4 adult lifetimes) (13). Scandinavian studies of antibody profiles evolving over several years before diagnosis indicate that other random events, occurring every few months or years, are probably involved (14). They also confirm that an adaptive response comes before synovitis, and it is an antibody response. Firestein and Zvaifler suggest that B cell activation is due to “undefined processes.” These processes may have remained undefined because they are engineered by the B cells themselves. “Anti-foreign” B cells control their own activation by interacting with antigen through surface and secreted antibody. Autoreactive B cells should not, but if a rare subset can do so through abnormal means, including the way rheumatoid factor B cells obtain T cell help, they will be selfperpetuating, and the kinetics of autoimmunity become comprehensible. Plausible mechanisms are not hard to identify (10). The generation of all antibody species is random. Antibodies able to subvert all control mechanisms and initiate RA might well arise only once in 200 woman-years, being what survival pressures will allow. Recruitment of other clones to create clinical disease might well take months or years. In advanced RA, B cell depletion often lowers the C-reactive protein level to normal as autoantibody levels fall (15,16). Relapse occurs following B cell return, but again, may evolve over many months. Firestein and Zvaifler are right to argue that T cells in rheumatoid synovitis are largely bystanders, but the answer to their question “what comes before T cells” is well documented (14): it is antibody. Jonathan C. W. Edwards, MD University College London, UK 1. Firestein GS, Zvaifler NJ. How important are T cells in chronic rheumatoid synovitis? [editorial]. Arthritis Rheum 1990;33: 768–73. 2. Firestein GS, Zvaifler NJ. How important are T cells in chronic rheumatoid synovitis? II. T cell–independent mechanisms from beginning to end. Arthritis Rheum 2002;46:298–308. 3. Stastny P. Association of the B cell alloantigen DRw4 with rheumatoid arthritis. N Engl J Med 1978;298:869–72. 4. Janeway CA, Travers P, Walport M, Capra JD. Allergy and hypersensitivity. In: Immunobiology, 4th ed. London: Elsevier Science; 1998. p. 478–80. 5. Bhatia A, Blades S, Cambridge G, Edwards JCW. Differential distribution of Fc␥RIIIa in normal human tissues and co-localization with DAF and fibrillin-1: implications for immunological microenviroments. Immunology 1998;94:56–63. 6. Matsumoto I, Staub A, Benoist C, Mathis D. Arthritis provoked by 3106 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. linked T and B cell recognition of a glycolytic enzyme. Science 1999;286:1732–5. Kleinau S, Martinsson P, Heyman B. Induction and suppression of collagen-induced arthritis is dependent on distinct fcgamma receptors. J Exp Med 2000;191:1611–6. Edwards JCW. Synovium. In: Klippel JH, editor. Primer on the rheumatic diseases. 12th ed. Atlanta: Arthritis Foundation; 2001. p. 22–6. Roosnek E, Lanzavecchia A. Efficient and selective presentation of antigen-antibody complexes by rheumatoid factor B cells. J Exp Med 1991;173:487–9. Edwards JCW, Cambridge G, Abrahams VM. Do self-perpetuating B lymphocytes drive human autoimmune disease? Immunology 1999;97:1868–96. Kuwana M, Feghali CA, Medsger TA Jr, Wright TM. Autoreactive T cells to topoisomerase I in monozygotic twins discordant for systemic sclerosis. Arthritis Rheum 2001;44:1654–9. Davies ML, Taylor EJ, Gordon C, Young SP, Welsh K, Bunce M, et al. Candidate T cell epitopes of the human La/SSB autoantigen. Arthritis Rheum 2002;46:209–14. Ollier WE, MacGregor A. Genetic epidemiology of rheumatoid disease. Br Med Bull 1995;51:267–85. Halldorsdottir HD, Jonsson T, Thorsteinsson J, Valdimarsson H. A prospective study on the incidence of rheumatoid arthritis among people with persistent increase of rheumatoid factor. Ann Rheum Dis 2000;59:149–51. Edwards JCW, Cambridge G. Sustained improvement in rheumatoid arthritis following a protocol designed to deplete B lymphocytes. Rheumatology 2001;40:205–11. Leandro MJ, Edwards JCW, Cambridge G. Clinical outcome in 22 patients with rheumatoid arthritis treated with B lymphocyte depletion. Ann Rheum Dis. In press. DOI 10.1002/art.10508 Reply To the Editor: We thank Dr. Edwards for his thoughts on RA and appreciate the fact that there are more similarities than differences between his view and our own. He acknowledges that elements of the innate immune system play a critical role in the initiation phase of inflammatory arthritis. His particular focus is the engagement of Fc␥ receptors on synovial intimal cells by immune complexes “small enough to traverse endothelium.” Presumably the autoantibodies in these complexes are produced in peripheral sites, gain access to the joints, and induce synovitis. This view is supported by rodent models of inflammatory arthritis caused by antibodies directed against type II collagen or glucose-6-phosphate isomerase, which require Fc␥RIII and components of the alternative (not the classical) complement pathway (1). However, experimental arthritis is not necessarily RA (2). The idea that immune complexes are important in the pathogenesis of RA joint inflammation hearkens back to concepts popular in the 1960s. Then, however, the emphasis was on the production of IgG and IgM antibodies by immune cells in the synovium; these formed large intraarticular immune complexes, engaged the classical complement system, and were removed by phagocytosis (for review, see ref. 3). The demonstration of complexes containing immunoglobulins, LETTERS rheumatoid factor, and early complement components in Fc␥R-bearing synovial fluid leukocytes and phagocytic intimal lining cells formed the basis for the notion that RA might be “an extravascular immune complex disease” (4). Indeed, it remains true that seropositive RA is the only form of chronic inflammatory arthritis associated with unequivocal consumption of early complement components in joint effusions. While we would agree that autoantibodies represent one potential mechanism for the initiation of synovitis, there is insufficient evidence to exclude an equally important role for engagement of Toll-like receptors, mannose receptors, or other primitive host defense mechanisms. Experimental evidence bearing on these facets of innate immunity will be clarified in future studies. The subsequent influx of immune and inflammatory cells into the joint could lead to either transient or persistent inflammation depending on the sex or genetic makeup of the individual, enviromental influences, and other still-to-be-defined factors. The production of autoantibodies in advance of clinical features of RA is well established (5), but does this mean that self-directed humoral responses are the only form of immunity (innate or acquired) that can prepare the synovium for subsequent invasion? The risk of developing RA in individuals who are persistently seropositive for rheumatoid factor is intriguing, and Dr. Edwards’ explanations are provocative. We anticipate, however, that the conversion from seropositivity to joint disease is likely to be more complex and will depend on a number of stochastic events (even including autoantibodies) as he suggests. Alternative mechanisms are also needed to explain the onset of RA in seronegative patients who later convert to seropositivity (6). In any case, we have once again stepped into the fray by fine-tuning our model of the initiation and perpetuation of RA. Our intent is to be provocative and raise questions, rather than answer all of them. Hopefully, this model will provide an impetus for dissecting the undefined mechanisms that lead to this uniquely human destructive arthritis. Gary S. Firestein, MD Nathan J. Zvaifler, MD University of California, San Diego School of Medicine La Jolla, CA 1. Matsumoto I, Maccioni M, Lee DM, Maurice M, Simmons B, Brenner M, et al. How antibodies to a ubiquitous cytoplasmic enzyme may provoke joint-specific autoimmune disease. Nat Immunol 2002;3:360–5. 2. Corr M, Firestein GS. Innate immunity as a hired gun: but is it rheumatoid arthritis? J Exp Med 2002;195:F33–5. 3. Zvaifler NJ. The immunopathology of joint inflammation in rheumatoid arthritis. Adv Immunol 1973;16:265–336. 4. Zvaifler NJ. Rheumatoid synovitis: an extravascular immune complex disease. Arthritis Rheum 1974;17:297–305. 5. Halldorsdottir HD, Jonsson T, Thorsteinsson J, Valdimarsson H. A prospective study on the incidence of rheumatoid arthritis among people with persistent increase of rheumatoid factor. Ann Rheum Dis 2000;59:149–51. 6. Dixon AStJ. Rheumatoid arthritis with negative serologic reaction. Ann Rheum Dis 1960;19:209–28. LETTERS 3107 DOI 10.1002/art.10570 Genetic drift as an explanation for the reduced incidence of rheumatoid arthritis To the Editor: With great interest we read the unique report by Doran et al on the decreasing incidence of rheumatoid arthritis (RA) (1). In the thoughtful accompanying editorial, Silman indicated that the most likely explanation for the worldwide trend of decline in incidence is a birth cohort effect (2). Since this effect is apparent in various populations such as American whites, Pima Indians, Finnish, and Japanese patients, the explanation should be a factor that had an identical effect in all populations throughout the world in the birth cohorts from 1880 to 1950. Such an effect could be either an environmental effect or an effect caused by a change in the population genome. We propose that this last explanation may be (partly) responsible for the decrease in incidence of RA. In previous times, human reproductive success was very unevenly distributed, with a minority of fertile women who gave birth to the majority of newborns. For example, in the 1912 Australian census, 50% of the children were the offspring of 1 in 7 of the women (3). However, in recent times this predominance steadily decreased since both fertile and less fertile women have equally contributed to the next generation. Genetic factors partly control fertility. Recently, we identified one of those factors, a high innate interleukin-10 (IL-10) production is associated with high fertility (4). Thus, before 1880, strong pressure within the genetic composition of the population was present that favors high IL-10 production. We previously demonstrated that IL-10 production is related to the composition of the IL-10 locus as defined by microsatellites; more specifically, the IL-10 R3 haplotype is associated with reduced IL-10 production (5). Moreover, we demonstrated that IL-10 R3 protects against RA in different ethnic populations (Scottish [Glasgow, UK]: odds ratio [OR] 0.6, 95% confidence interval [95% CI] 0.39–0.92; British [Oxford, UK]: OR 0.58, 95% CI 0.38–0.87; African American [Atlanta, GA]: OR 0.38, 95% CI 0.18–0.84) (6). The fact that the number of children per woman has decreased in the birth cohorts from 1880 to 1940 must have had the consequence that more women with a genetic makeup for impaired fertility contributed to the offspring. With respect to the IL-10 locus, this must have led to an increase of IL-10 R3 (which protects against RA) in the general population. This genetic drift may explain the lower incidence of RA in those born in the birth cohorts after 1880. Tom W. J. Huizinga, MD, PhD Suzanne Linn-Rasker, MD Leroy R. Lard, MD Rudi G. J. Westendorp, MD, PhD Leiden University Medical Center Leiden, The Netherlands 1. Doran MF, Pond GR, Crowson CS, O’Fallon WM, Gabriel SE. Trends in incidence and mortality in rheumatoid arthritis in Rochester, Minnesota, over a forty-year period. Arthritis Rheum 2002; 46:625–31. 2. Silman AJ. The changing face of rheumatoid arthritis: why the decline in incidence? Arthritis Rheum 2002;46:579–81. 3. Cummins J. Evolutionary forces behind human infertility [letter]. Nature 1999;397:557–8. 4. Westendorp RG, van Dunne FM, Kirkwood TB, Helmerhorst FM, Huizinga TW. Optimizing human fertility and survival [letter]. Nat Med 2001;7:873. 5. Eskdale J, Gallagher G, Verweij CL, Keijsers V, Westendorp RG, Huizinga TW. Interleukin 10 secretion in relation to human IL-10 locus haplotypes. Proc Natl Acad Sci U S A 1998;95:9465–70. 6. Eskdale J, McNicholl J, Wordsworth P, Jonas B, Huizinga T, Field M, et al. Interleukin-10 microsatellite polymorphisms and IL-10 locus alleles in rheumatoid arthritis susceptibility [letter]. Lancet 1998;352:1282–3. DOI 10.1002/art.10488 Successful rechallenge with anti–tumor necrosis factor ␣ for psoriatic arthritis after development of demyelinating nervous system disease during initial treatment: comment on the article by Mohan et al To the Editor: We read with interest the article by Mohan et al about the occurrence of neurologic events during anti–tumor necrosis factor ␣ (anti-TNF␣) therapy for inflammatory arthritides (1). Using the criteria proposed by Miller et al (2), Mohan and colleagues concluded that there were enough elements in their series to support the notion of an association between antiTNF␣ therapy and neurologic complications. Although they pointed out that reported cases were few, they recommended discontinuation of anti-TNF␣ agents when neurologic signs appear. We report herein the case of a 34-year-old man with severe psoriatic arthritis and very extensive skin involvement who developed a recurrence of Guillain-Barré syndrome during anti-TNF␣ treatment, followed later by a successful rechallenge. The patient’s disease began in 1986, when he was 19 years old, with polyarthritis, balanitis, and sacroiliac pain with radiologically evident sacroilitis, without skin disease. He was diagnosed at another institution as having reactive arthritis and treated elsewhere with corticosteroids. After 6 months, he developed a right facial palsy and instability while walking. On examination, ataxia, areflexia, and external ophthalmoplegia were observed. Electromyography demonstrated a generalized motor polyradiculoneuropathy. The cerebrospinal fluid (CSF) was normal, as were findings on computed tomography (CT) of the brain. Guillain-Barré syndrome (Miller-Fisher variety) was diagnosed, and he was treated with plasmapheresis and highdose corticosteroids. After 2 months, the Guillain-Barré syndrome had completely resolved. In the subsequent years, the patient had several flares of polyarthritis, treated with prednisone at various doses and nonsteroidal antiinflammatory drugs. In 1990, a typical and extensive psoriatic rash appeared, and methotrexate (MTX) therapy was started, with continuation of the steroids. During the next 9 years, he experienced a severe decline in functional capacity because of sustained polyarthritis and extensive cutaneous involvement, and had to leave his job. His erythrocyte sedimentation rate and C-reactive protein level were persistently elevated. He was referred to our institution in 1996. We 3108 LETTERS tried several therapeutic approaches to control the disease activity: high-dose MTX, cyclosporine, MTX plus azathioprine, MTX plus mycofenolate, and azathioprine plus mycofenolate, without improvement in either the articular disease or the skin involvement. In March 2000, the patient received his first dose of infliximab (200 mg), with an impressive response: 2 weeks after treatment, the joint and skin disease had resolved completely. A second dose of infliximab (100 mg) was given in May. In June 2000, he reported weakness in his legs, which quickly extended to the upper limbs and to respiratory muscles. After 2 days, he was placed on mechanical ventilatory support. Electromyography showed evidence of a demyelinating motor polyradiculoneuropathy. Magnetic resonance imaging of the cervical spine showed inestability at the C1–C2 level without evidence of myelopathy. A brain CT scan yielded normal results. CSF analysis revealed a glucose level of 49 mg/dl and a protein level of 54 mg/dl, with no leukocytes. Guillain-Barré syndrome was diagnosed, and all therapy was discontinued. Intravenous immunoglobulin was administered (2 gm/kg in 5 days), and the patient recovered fully after 3 weeks. One month later, the patient began to experience joint pain again, and, fearing a recurrence of his devastating disease, he requested that infliximab treatment be reinstituted. After discussing with him the uncertainty of this trial, we reinstituted the treatment, but in lower doses than used previously. In July 2000, he received 100 mg of infliximab. His joint pain disappeared and since then, he receives infliximab 100 mg once every 3 months, plus MTX 15 mg/week. He has no joint pain, and the only skin symptom is hyperpigmentation in some areas. He has resumed normal employment in an office. There are several reports in the literature about recurrences of Guillain-Barré syndrome 1–25 years after the first episode (3,4). Most of those recurrences were clinically similar to the initial episode. Some of them seemed to be related to a viral infection (5). Our patient had a variant of a GuillainBarré syndrome years before the start of anti-TNF␣ therapy. The recurrence might have been elicited by the blockade of TNF␣ or it might have been part of the natural history of the disease. The successful rechallenge with infliximab suggests that the latter could be the case. If so, then there might be differences among the various demyelinating diseases with regard to TNF blockade sensitivity. It is notable that the patient had such a good response with the low doses used. It might also be that these doses were sufficient to treat his condition but not high enough to elicit neurologic disease. However, even though our patient did not have any problems with the rechallenge, we do not believe we can make a definitive conclusion about how to continue therapy after the appearance of a demyelinating disease in the course of antiTNF␣ treatment. Marcela Cisternas, MD Miguel Gutiérrez, MD Sergio Jacobelli, MD Pontificia Universidad Católica de Chile Santiago, Chile 1. Mohan N, Edwards ET, Cupps TR, Oliverio PJ, Sandberg G, Crayton H, et al. Demyelination occurring during anti–tumor 2. 3. 4. 5. necrosis ␣ therapy for inflammatory arthritides. Arthritis Rheum 2001;44:2862–9. Miller FW, Hess EV, Clauw DJ, Hertzman PA, Pincus T, Silver RM, et al. Approaches for identifying and defining environmentally associated rheumatic disorders. Arthritis Rheum 2000;43:243–9. Taly AB, Gupta SK, Anisya U, Shankar SK, Rao S, Das KB, et al. Recurrent Guillain Barre’s syndrome: a clinical, electrophysiological and morphological study. J Assoc Physicians India 1995;43: 249–52. Combes A, Goulon M. Recurrence of Guillain-Barre syndrome. Ann Med Interne (Paris) 1992;143:515–8. Jacobs BC, Rothbarth PH, van der Meche FG, Herbrink P, Schmitz PI, de Klerk MA, et al. The spectrum of antecedent infections in Guillain-Barre syndrome: a case-control study. Neurology 1998;51: 1110–5. DOI 10.1002/art.10489 Reply To the Editor: In our case series, 4 patients had a flare of previously existing multiple sclerosis (MS) following administration of anti-TNF␣ therapy, as demonstrated in Table 1. However, in 3 of the 4 patients, the disease had remained completely quiescent for several years after initial diagnosis and flared only after institution of anti-TNF␣ therapy. Patient 4 had had several flares between diagnosis of MS and initiation of anti-TNF␣ treatment. However, the flare that occurred after anti-TNF␣ therapy was longer and more severe. Three patients in the reported series who continued to receive anti-TNF␣ therapy experienced worsening neurologic symptoms, while the rest, who discontinued these agents, had partial or complete resolution. Relapses in both Guillain-Barré syndrome (Romano JG, Rotta FT, Potter P, Rosenfeld V, Santibanez R, Rocha B, et al. Relapses in the Guillain-Barre syndrome after treatment with intravenous immune globulin or plasma exchange. Muscle Nerve 1998;21:1327–30) and MS (Amato MP, Ponziani G, Bartolozzi ML, Siracusa G. A prospective study on the natural history of multiple sclerosis: clues to the conduct and interpre- Table 1. Flares of multiple sclerosis (MS) after treatment with anti–tumor necrosis factor ␣ (anti-TNF␣) Patient with preexisting MS Duration between diagnosis of MS and flare, years Duration of anti-TNF␣ treatment prior to flare, months 1 10 10 2 26 12 3 4 5 4 6 1 Outcome Partial resolution at 4 months Complete resolution at 3 months Continued symptoms at 6 months Partial resolution at 3 months LETTERS 3109 tation of clinical trials. J Neurol Sci 1999;168:96–106), both of which are considered to be autoimmune disorders directed against central and peripheral myelin components, respectively, have been described. However, in Cisternas et al’s patient the fact that the relapse occurred within 3 months of initiation of anti-TNF␣ therapy, after being quiescent for 14 years (1986–2000), increases the likelihood of a true association. Hence, we reiterate that it would be prudent to discontinue anti-TNF␣ therapy in such a situation until the underlying etiology can be better delineated. Furthermore, the occurrence of a positive rechallenge phenomenon in patients such as patient 2 in our report, as well as more reports to the Food and Drug Administration of such events (personal communication), suggest that rechallenge is associated with a significantly increased risk of recurrence, even with a lower dose of the drug. Therefore, we believe rechallenge should be avoided unless no other treatment options are available. Niveditha Mohan, MBBS Thomas R. Cupps, MD Georgetown University Medical Center Washington, DC DOI 10.1002/art.10414 A nonsense mutation in exon 2 of the DNase I gene is not present in UK subjects with systemic lupus erythematosus and Graves’ disease: comment on the article by Rood et al To the Editor: The identification of genetic loci conferring susceptibility to the development of systemic lupus erythematosus (SLE) is important for our understanding of the disease process and the ultimate design of new therapies. Rood et al recently reported not only association between the HLA gene region on chromosome 6p21 and SLE but also an independent association with tumor necrosis factor ␣ (1). While the major histocompatibility complex (MHC) genes are inextricably linked with the autoimmune disease process, they have so far remained intractable to therapeutic modulation through peptide mimics. Identification of non-MHC loci, therefore, may hold the key to the development of novel therapies. The DNase I gene (DNASE1) on chromosome 16p13.3 (2) encodes an enzyme that hydrolyzes DNA forming deoxyribonucleotides. A murine model has shown that DNase I–deficient mice have an increased incidence of SLE, with classic findings including the presence of autoreactive antibodies and glomerulonephritis in a DNase I-dose– dependent manner. This suggests that DNase I could prevent the expansion of lymphocytes responsive to nucleosomal antigens, which contribute to SLE, by the removal of DNA from soluble or deposited autoantigenic nucleoprotein complexes (3). DNase I is, therefore, a potential target for therapeutic modulation. A heterozygous nonsense mutation in DNASE1 was reported in 2 of 20 female Japanese patients with SLE, resulting in an A3 G transversion in exon 2 at position 172 of the coding sequence, leading to the generation of a stop codon within the DNASE1 sequence (4). This resulted in decreased DNase I activity and an increased IgG titer against nucleoso- mal antigens, suggesting that DNASE1 could be involved not only in the genetic susceptibility to SLE but also the autoimmune disease process in general. The presence of a DNASE1 mutation could potentially identify a subgroup of patients with SLE with specific disease characteristics relating, for example, to progression of disease, prognosis, or response to therapy. The aim of this study, therefore, was to determine whether the A3 G mutation of DNASE1 plays a role in the genetic susceptibility to the development of SLE and the autoimmune disease process in a Caucasian UK population. One hundred eighty-two patients (12 of whom had a history of autoimmune thyroid disease) were recruited from the SLE clinic at the University Department of Rheumatology, Queen Elizabeth Hospital, Birmingham. Patients fulfilled the American College of Rheumatology criteria for the classification of SLE (5). Two hundred ninety-one patients with autoimmune hyperthyroidism (Graves’ disease), 5 of whom had a family history of SLE, were recruited from thyroid clinics, as described previously (6). Ethnically matched control subjects (285) with no family history of autoimmune disease, recruited as previously described, were also used (6). The study was approved by the local ethics committee and all subjects gave informed, written consent. Genomic DNA was extracted from whole blood using the Bacc II DNA extraction kit from Nucleon Biosciences (Lanarkshire, UK). Amplification of the target DNA in exon 2 of DNASE1 was carried out by polymerase chain reaction (PCR) and the resulting products were subjected to restriction fragment length polymorphism (RFLP) analysis with Nsp I using conditions previously described (4). To verify the Nsp I assay, RFLP was performed on a sample of DNA (HLA– DQB1*0602 cloned into the pCIneo expression vector) containing 6 Nsp I sites. Genotyping results obtained by PCRRFLP were verified by subjecting 15 randomly selected samples to sequence analysis using the ABI Prism 377 DNA sequencer (Applied Biosystems, Warrington, UK). Unambiguous genotyping data were obtained on all samples. The A3 G mutation in exon 2 of DNASE1 was not found in any of the 758 subjects studied. Although the A3 G mutation of DNASE1 has been shown to be associated with decreased DNase I activity and a high IgG titer against nucleosomal antigens in 2 Japanese patients with SLE (7), we examined 1,516 human chromosomes, including 364 in subjects with SLE, and found no evidence to suggest that this plays a role in the development of SLE in a Caucasian UK population. There are a number of factors that could explain the observed differences between the Japanese and UK populations. First, it is possible that the A3 G mutation is in linkage disequilibrium with another gene or haplotype that is associated with SLE in the Japanese population but not in the UK population. Second, certain genes may lead to the development of autoimmune disease in specific ethnic/geographic populations only. For example, chromosome 1p36 is linked to SLE in a Mexican American population but not in a Caucasian population (8). Third, the A3 G mutation may have produced a rare form of SLE in 2 Japanese subjects. Finally, however, it is worth noting that no further A3 G mutations were found in a second cohort of 180 Japanese patients, raising the possibility of a false positive result in the original data set (Yasutomo K: personal communication). 3110 LETTERS In conclusion, further studies are required to determine the mechanisms behind the role of the MHC in SLE and other autoimmune diseases. However, the study of non-MHC candidate genes is also important. Although the A3 G mutation in DNASE1 does not appear to be associated with SLE or the autoimmune disease process in the UK, this and other functional candidates require detailed investigation if novel disease-modifying therapies are to be identified. Matthew J. Simmonds, BSc Joanne M. Heward, PhD M. Ann Kelly, PhD Amit Allahabadia, MD Helen Foxall, BSc Caroline Gordon, MD, FRCP Jayne A. Franklyn, PhD University of Birmingham Queen Elizabeth Hospital Stephen C. L. Gough, MD, FRCP University of Birmingham Birmingham Heartlands Hospital Birmingham, UK 1. Rood MJ, van Krugten MV, Zanelli E, van der Linden MW, Keijsers V, Schreuder GM, et al. TNF-308A and HLA–DR3 alleles contribute independently to susceptibility to systemic lupus erythematosus. Arthritis Rheum 2000;43:129–34. 2. Yasuda T, Nadano D, Iida R, Takeshita H, Lane SA, Callen DF, et al. Chromosomal assignment of the human deoxyribonuclease I gene, DNASE 1 (DNL1), to band 16p13.3 using the polymerase chain reaction. Cytogenet Cell Genet 1995;70:221–3. 3. Napirei M, Karsunky H, Zevnik B, Stephan H, Mannherz HG, Moroy T. Features of systemic lupus erythematosus in Dnase1deficient mice. Nat Genet 2000;25:177–81. 4. Yasutomo K, Horiuchi T, Kagami S, Tsukamoto H, Hashimura C, Urushihara M, et al. Mutation of DNASE1 in people with systemic lupus erythematosus. Nat Genet 2001;28:313–4. 5. Tan EM, Cohen AS, Fries JF, Masi AT, McShane DJ, Rothfield NF, et al. The 1982 revised criteria for the classification of systemic lupus erythematosus. Arthritis Rheum 1982;25:1271–7. 6. Heward JM, Nithiyananthan R, Allahabadia A, Gibson S, Franklyn JA, Gough SC. No association of an interleukin 4 gene promoter polymorphism with Graves’ disease in the United Kingdom. J Clin Endocrinol Metab 2001;86:3861–3. 7. Gordon C, Salmon M. Update on systemic lupus erythematosus: autoantibodies and apoptosis. Clin Med 2001;1:10–4. 8. Shai R, Quismorio FP Jr, Li L, Kwon OJ, Morrison J, Wallace DJ, et al. Genome-wide screen for systemic lupus erythematosus susceptibility genes in multiplex families. Hum Mol Genet 1999;8: 639–44. DOI 10.1002/art.10484 Antioxidant treatment decreases the titer of circulating anticardiolipin antibodies: comment on the article by Sambo et al To the Editor: Sambo et al described an enhanced oxidative stress in scleroderma and the maintenance of scleroderma fibroblast phenotype by the constitutive up-regulation of reactive oxygen species generation through the NADPH oxidase complex pathway (1). Previous reports indicated that an enhanced oxidative stress could also be linked to the formation of anticardiolipin antibodies (aCL) in autoimmune diseases. In particular, an interesting hypothesis suggested that aCL bind exclusively to the peroxidized molecule, indicating that these antibodies recognize epitopes derived from phospholipid oxidation (2). Furthermore, a close association between antiphospholipid antibodies (aPL) and in vivo markers of lipid peroxidation (3) has been demonstrated. To explore this hypothesis in aPL-positive patients, we undertook an interventional study to assess whether antioxidant treatment is able to affect the serum titer of aCL. We studied 14 consecutive outpatients (12 women and 2 men; age 24–49 years) who were positive for aCL, with a titer of 14–120 IgG phospholipid units (GPL units) or IgM phospholipid units (MPL units); among these patients, 8 were also positive for lupus anticoagulant (LAC). Six of 14 aPL-positive patients had primary antiphospholipid syndrome (primary APS) with a history of arterial and/or venous thrombosis in the previous 12–25 months. The remaining 8 patients had systemic lupus erythematosus, and 1 of these patients had had a thromboembolic stroke 19 months previously. At the time of the study, all patients with a previous episode of arterial thrombosis were treated with aspirin (100 mg/day). The 3 patients who had had an episode of venous thrombosis in the previous 15–25 months discontinued the anticoagulant treatment spontaneously ⬃6 months before their inclusion into the study. The study was approved by the Internal Review Board and Ethics Committee. Patients positive for aPL were randomly treated with (group 1; n ⫽ 7) or without (group 2; n ⫽ 7) antioxidant supplementation (vitamin E at 900 IU/day, vitamin C at 2,000 mg/day). Clinical and laboratory parameters were evaluated before and after 4–6 weeks of treatment. An enzyme-linked immunosorbent assay validated in an international workshop was used for measurement of aCL. Patients were considered positive for IgG or IgM aCL when the serum concentration was ⬎10 GPL units or ⬎10 MPL units, respectively (4). Patients were considered positive for aPL if aCL and/or LAC were detected on two separate occasions at least 2 months apart. Levels of LAC, fibrinogen, prothrombin fragment 1⫹2 (F1⫹2), serum tumor necrosis factor ␣ (TNF␣), and plasma vitamin E and vitamin C were assayed as previously described (5). Previous study showed that the sample would have to consist of at least 6 patients in each group (␣ ⫽ 0.05 and 1 ⫺ ␤ ⫽ 0.90) (6), assuming that antioxidant treatment reduced F1⫹2, a marker of thrombin generation in vivo, by 40% (4). We postulated that the same sample size was necessary to observe at least 40% reduction in aCL circulating titer. The statistical analysis was performed by appropriate t-test. The linear regression test was used to study the different correlations. The effect of treatment was analyzed by two-way repeated-measures analysis of variance (ANOVA). Data were presented as the mean ⫾ SD and median and range. Only 2-tailed probabilities were used for testing statistical significance. P values less than 0.05 were considered significant. The two treatment groups were homogeneous for age, sex, disease activity, previous thrombosis, primary APS, and corticosteroid and aspirin treatment, as well as for circulating levels of aCL, F1⫹2, and plasma vitamins E and C (Table 1). In order to assess the oxidative stress in the whole aPL group, LETTERS 3111 Table 1. Clinical and laboratory characteristics of patients enrolled (group 1) or not (group 2) for vitamin E and vitamin C supplementation* Group 1 (n ⫽ 7) Before treatment After treatment Age in years, mean ⫾ SD 35 ⫾ 1 – Men 1 (14) – Primary APS 3 (43) – Active disease 2 (29) 2 (29) Thrombosis 4 (57) 4 (57) Corticosteroid treatment, prednisone 2 (29) 2 (29) 5–15 mg/day or methylprednisolone 4–24 mg/day Aspirin treatment, 100 mg/day 2 (29) 2 (29) Fibrinogen in mg/dl, mean ⫾ SD 261 ⫾ 40.9 258 ⫾ 42 TNF in pg/ml, median (range) 186.2 (94.6–290.4) 161.8 (84.9–280.4) aCL, median (range)§ 32 (16–120) 8 (3–100) F1⫹2 in moles, mean ⫾ SD 1.99 ⫾ 0.44 1.19 ⫾ 3.4 Vitamin E in moles/liter, mean ⫾ 14.5 ⫾ 3.4 39.7 ⫾ 7.7 SD Vitamin C in moles/liter, mean ⫾ 16.1 ⫾ 7.0 26.3 ⫾ 9.3 SD P† P, difference between groups after treatment‡ – – – ⬎0.05 ⬎0.05 ⬎0.05 – – – – – – ⬎0.05 ⬎0.05 ⬎0.05 ⬎0.05 ⬎0.05 ⬎0.05 – ⬎0.05 ⬎0.05 ⬍0.05 ⬍0.05 ⬍0.05 ⬎0.05 ⬍0.05 Group 2 (n ⫽ 7) P† – – – ⬎0.05 ⬎0.05 ⬎0.05 Before treatment After treatment 36 ⫾ 11 1 (14) 3 (43) 1 (14) 3 (43) 1 (14) – – – 1 (14) 3 (43) 1 (14) ⬎0.05 2 (29) 2 (29) ⬎0.05 263 ⫾ 39 258 ⫾ 40 ⬎0.05 189.2 (96.4–248.6) 177.6 (85.7–282) 0.016 50 (14–100) 50 (9–100) 0.016 1.96 ⫾ 0.48 2.02 ⫾ 0.33 0.016 14.9 ⫾ 4.3 16.6 ⫾ 3.8 0.016 17.8 ⫾ 8.9 17.2 ⫾ 8.6 * Except where indicated otherwise, values are the number (%) of patients. APS ⫽ antiphospholipid syndrome; TNF ⫽ tumor necrosis factor; aCL ⫽ anticardiolipin antibodies; F1⫹2 ⫽ prothrombin fragment 1⫹2. † Wilcoxon signed rank test. ‡ In percentages of change from baseline, by Mann-Whitney U test. § In IgG phospholipid units or IgM phospholipid units. we compared plasma levels of vitamins E and C of aPLpositive patients with those of 20 healthy subjects (18 women and 2 men; age 22–51 years) matched for age and sex. Compared with controls, aPL ⫽ positive patients had significantly lower levels of vitamin E (mean ⫾ SD 14.7 ⫾ 6.8 moles/liter versus 38.9 ⫾ 4 moles/liter; P ⬍ 0.0001) and vitamin C (16.9 ⫾ 7.8 moles/liter versus 40 ⫾ 16.7 moles/ liter; P ⬍ 0.0001). Vitamin E (r ⫽ ⫺0.63, P ⬍ 0.02) and vitamin C (r ⫽ ⫺0.65, P ⫽ 0.01) were significantly inversely correlated with circulating levels of F1⫹2 in aPL ⫽ positive patients. In subjects not assigned to antioxidant treatment, no changes in values of clinical and laboratory variables were observed at the end of treatment. Conversely, patients assigned to antioxidants showed significant decreases in aCL titer and in levels of F1⫹2 and significant increases in plasma levels of both vitamin E and vitamin C (Table 1). In all patients treated with antioxidants, we observed a mean decrease in aCL titer of 61% (range 16–84%). The ANOVA confirmed significant decreases in aCL titer (P ⫽ 0.01) and F1⫹2 plasma levels (P ⫽ 0.0001) and significant increases in levels of vitamin E (P ⫽ 0.0001) and vitamin C (P ⫽ 0.0001) after antioxidant treatment. We can reasonably exclude the possibility that change in disease activity could account for the decrease in aCL titer, since levels of TNF␣ and fibrinogen (two markers of disease activity) as well as clinical characteristics were not modified at the end of treatment. The decrease in aCL titer could therefore be considered an effect of antioxidant treatment, thus supporting the hypothesis that oxidative stress plays a central role in the formation of aCL. However, further study is necessary to assess the intrinsic mechanism by which oxidative stress gives rise to formation of new phospholipid epitopes in vivo. Antioxidant treatment may be of clinical value in antiphospholipid syndrome (APS), by virtue of a potential antithrombotic effect. In fact, consistent with previous findings, we observed a reduced rate of thrombin generation that can be interpreted to be a consequence of reduced lipid peroxidation and/or reduced aCL titer (3,5). Should findings of this pilot study be confirmed in a larger study, reducing oxidative stress may represent a new strategy for the treatment of APS. D. Ferro, MD L. Iuliano, MD F. Violi, MD G. Valesini, MD F. Conti, MD “La Sapienza” University of Rome Rome, Italy 1. Sambo P, Baroni SS, Luchetti M, Paroncini P, Dusi S, Orlandini G, et al. Oxidative stress in scleroderma: maintenance of scleroderma fibroblast phenotype by the constitutive up-regulation of reactive oxygen species generation through the NADPH oxidase complex pathway. Arthritis Rheum 2001;44:2653–64. 2. Horkko S, Miller E, Dudl E, Reaven P, Curtiss LK, Zvaifler NJ, et al. Antiphospholipid antibodies are directed against epitopes of oxidized phospholipids: recognition of cardiolipin by monoclonal antibodies to epitopes of oxidized low density lipoprotein. J Clin Invest 1996;98:815–25. 3. Iuliano L, Praticò D, Ferro D, Pittoni V, Valesini G, Lawson J, et al. Enhanced lipid peroxidation in patients positive for antiphospholipid antibodies. Blood 1997;90:3931–5. 4. Harris EN, Gharavi AE, Patel SP, Hughes GRV. Evaluation of the anticardiolipin antibody test: report of an international workshop held 4 April 1986. Clin Exp Immunol 1987;68:214–9. 5. Praticò D, Ferro D, Iuliano L, Rokach J, Conti F, Valesini G, et al. Ongoing prothrombotic state in patients with antiphospholipid 3112 antibodies: a role for increased lipid peroxidation. Blood 1999;93: 3401–7. 6. Ferro D, Parrotto S, Basili S, Alessandri C, Violi F. Simvastatin inhibits the monocyte expression of proinflammatory cytokines in patients with hypercholesterolemia. J Am Coll Cardiol 2000;36: 427–31. DOI 10.1002/art.10483 Role of T cells in the pathogenesis of osteoarthritis To the Editor: We read with interest the review by Pelletier et al (1). This is a timely article suggesting that osteoarthritis (OA) is an inflammatory disease and that the traditional view by rheumatologists that OA is a noninflammatory disease should be reconsidered and the disease should be reclassified. In their review, Pelletier et al (1) cited macrophages as the exclusive source of inflammation in OA. However, the role of T cells in the inflammatory process has not been considered. We (2–7) and others (8–10) have provided evidence that T cell infiltrates are frequently detected in the synovial membrane (SM) of patients with OA (2,8–10). These infiltrates are often angiocentric (4) and are associated with activation of local vascular endothelial cells, as suggested by the increase in expression of E-selectin (10). In patients with advanced OA, T cell infiltrates in the SM exhibit a nodular pattern in 37% (9) to 65% (2) of the patients and express early (CD69), intermediate (CD25), and late (CD45RO, HLA–DR) activation antigens (2). Additionally, T cell cytokine transcripts of the Th1 type interferon-␥ (IFN␥) and interleukin-2 (IL-2) and IL-10 were found in the SM of patients with OA, whereas IL-4 and IL-5 were not detected (2). There were no statistical differences in the levels of IFN␥ and IL-2 transcripts in the SM between patients with rheumatoid arthritis (RA) and OA, when normalized for T cell number equivalents (2). However, when the levels of IFN␥ transcripts were normalized for total cell number equivalents, they were lower in OA than in RA. The presence of substantial proportions of T cells expressing early, intermediate, and late activation antigens and of the Th1 cytokine pattern (2) in chronic SM lesions of patients with OA strongly suggests that T cells at least contribute to chronic inflammation in these patients. This Th1 response may be driven by macrophages. Macrophages and synovial lining cells express IL-12, a cytokine that drives the Th1 immune response (3). OA synovial fluid exhibits increased levels of macrophage inflammatory protein 1␤ (11), a ligand for the chemokine receptor CCR5, present on Th1 cells (12). Although these findings may be explained by a nonspecific activation of T cells, we have demonstrated (6,7) the presence of oligoclonal populations of T cells in the SM of 4 out of 5 patients with advanced OA. Amplification of ␤-chain T cell receptor (TCR) transcripts from the SM of patients with OA by either nonpalindromic adaptor polymerase chain reaction (PCR) or V␤-specific PCR (13), followed by cloning and sequencing of the amplified transcripts, revealed substantial proportions of identical ␤-chain TCR transcripts, suggesting the presence of oligoclonal populations of T cells. These results strongly suggest that T cells have undergone antigen-driven proliferation and clonal expansion in situ in the SM of patients with OA, in LETTERS response to as-yet-unidentified antigens. These antigen(s) are not known, but one study suggested a self-reactive immune response to chondrocyte membrane components (14). Like other conditions of chronic T cell activation (15), such as RA (16), systemic lupus erythematosus (17), and tumor-infiltrating lymphocytes (18), T cells in the SM of patients with OA show decreased expression of CD3- chain transcripts and protein (5). The inflammation in OA may not be confined within the joints. One study described perivascular lymphocytic infiltrates in muscle biopsies of 18% of patients with OA (19). Activated T cells, through cell contact– dependent interaction or through soluble mediators (20), can stimulate monocytes to produce cytokines (21). In rheumatoid synovitis, T cells were found to be largely responsible for the production of metalloproteinase (22). All these results taken together strongly suggest that a T cell immune response occurs in OA (6). Substantial evidence has been accumulated suggesting that OA is an inflammatory disease. The traditional view that OA is a noninflammatory disease must be revised. We believe that it is difficult to explain the chronic inflammation that is observed in the SM of patients with OA without a role for T cells and a role for putative antigen(s) in the initiation and propagation of the disease. Lazaros I. Sakkas, MD, PhD Temple University School of Medicine Philadelphia, PA and Thessaly University School of Medicine Larisa, Greece Chris D. Platsoucas, PhD Temple University School of Medicine Philadelphia, PA 1. Pelletier JP, Martel-Pelletier J, Abramson SB. Osteoarthritis, an inflammatory disease: potential implication for the selection of new therapeutic targets. Arthritis Rheum 2001;44:1237–47. 2. Sakkas LI, Scanzello C, Johanson N, Burkholder J, Mitra A, Salgame P, et al. T cells and T-cell cytokine transcripts in the synovial membrane in patients with osteoarthritis. Clin Diagn Lab Immunol 1998;5:430–7. 3. Sakkas LI, Johanson NA, Scanzello CR, Platsoucas CD. Interleukin-12 is expressed by infiltrating macrophages and synovial lining cells in rheumatoid arthritis and osteoarthritis. Cell Immunol 1998;188:105–10. 4. Sakkas LI, Scanzello CR, Katsetos CD, Johanson NA, Platsoucas CD. Angiocentric inflammation in the synovial membrane (SM) in osteoarthritis. Rheumatology 2000;39 Suppl:117. 5. Sakkas LI, Koussidis GA, Avgerinos ED, Platsoucas CD. Decreased expression of CD3zeta chain in the synovial membrane in osteoarthritis. Rheumatology 1999;39 Suppl:217. 6. Scanzello CR, Sakkas LI, Johanson NA, Platsoucas CD. Oligoclonal populations of T-cells infiltrate the synovial membrane (SM) of patients with osteoarthritis (OA) [abstract]. Arthritis Rheum 1999;43 Suppl 9:S257. 7. Scanzello C, Sakkas LI, Johanson N, Platsoucas CD. Clonally expanded T cells in the synovial membrane of patients with osteoarthritis. Scand J Immunol 2001;54 Suppl 1:59. 8. Linblad S, Hedfors E. Arthroscopic and immunohistologic characterization of knee joint synovitis in osteoarthritis. Arthritis Rheum 1987;30:1081–8. 9. Smith MD, Triantafillou S, Parker A, Youssef PP, Coleman M. Synovial membrane inflammation and cytokine production in patients with early osteoarthritis. J Rheumatol 1997;24:365–71. LETTERS 10. Koch AE, Turkiewicz W, Harlow LA, Pope RM. Soluble Eselectin in arthritis. Clin Immunol Immunopathol 1993;69:29–35. 11. Koch AE, Kunkel SL, Shah MR, Fu R, Mazarakis DD, Haines GK, et al. Macrophage inflammatory protein-1␤: a C-C chemokine in osteoarthritis. Clin Immunol Immunopathol 1995;77: 307–14. 12. Loetscher P, Uguccioni M, Bordoli L, Baggiolini M, Moser B, Chizzolini C, et al. CCR5 is characteristic of Th1 lymphocytes. Nature 1998;391:344–5. 13. Slachta CA, Jeevanandam V, Goldman B, Lin WL, Platsoucas CD. Coronary arteries from human cardiac allografts with chronic rejection contain oligoclonal T cells: persistence of identical clonally expanded TCR transcripts from the early post-transplantation period (endomyocardial biopsies) to chronic rejection (coronary arteries). J Immunol 2000;165:3469–83. 14. Alsalameh S, Mollenhauer J, Hain N, Stock KP, Kalden JR, Burmester GR. Cellular immune response toward human articular chondrocytes: T cell reactivities against chondrocyte and fibroblast membranes in destructive joint diseases. Arthritis Rheum 1990;33: 1477–86. 15. Krishnan S, Warke VG, Nambiar MP, Wong HK, Tsokos GC, Farber DL. Generation and biochemical analysis of human effector CD4 T cells: alterations in tyrosine phosphorylation and loss of CD3 expression. Blood 2001;97:3851–9. 16. Matsuda M, Ulfgren AK, Lenkei R, Petersson M, Ochoa AC, Lindblad S, et al. Decreased expression of signal-transducing CD3 chains in T cells from the joints and peripheral blood of rheumatoid arthritis patients. Scand J Immunol 1998;47:254–62. 17. Liossis SN, Ding XZ, Dennis GJ, Tsokos GC. Altered pattern of TCR/CD3-mediated protein-tyrosyl phosphorylation in T cells from patients with systemic lupus erythematosus. J Clin Invest 1998;101:1448–57. 18. Finke JH, Zea AH, Stanley J, Longo DL, Mizoguchi H, Tubbs RR, et al. Loss of T-cell receptor chain and p56lck in T-cells infiltrating human renal cell carcinoma. Cancer Res 1993;53: 5613–6. 19. Voskuyl AE, van Duinen SG, Zwinderman AH, Breedveld FC, Hazes JM. The diagnostic value of perivascular infiltrates in muscle biopsy specimens for the assessment of rheumatoid vasculitis. Ann Rheum Dis 1998;57:114–7. 20. Aarvak T, Chabaud M, Miossec P, Natvig JB. IL-17 is produced by some proinflammatory Th1/Th0 cells but not by Th2 cells. J Immunol 1999;162:1246–51. 21. Sebbag M, Parry SL, Brennan FM, Feldmann M. Cytokine stimulation of T lymphocytes regulates their capacity to induce monocyte production of tumor necrosis factor-␣, but not interleukin-10: possible relevance to pathophysiology of rheumatoid arthritis. Eur J Immunol 1997;27:624–32. 22. Klimiuk PA, Yang H, Goronzy JJ, Weyand CM. Production of cytokines and metalloproteinases in rheumatoid synovitis is T cell dependent. Clin Immunol 1999;90:65–78. DOI 10.1002/art.10528 Benefit or risk of aspirin treatment of giant cell arteritis: comment on the article by Weyand et al To the Editor: Weyand et al (1) document the efficacy and perhaps even superiority of acetylsalicylic acid (aspirin) for normalizing major immunomodulators in giant cell arteritis. Although proinflammatory cytokines were suppressed, past clinical experience must also be respected. Suppression of immunomodulation and, specifically, suppression of the erythrocyte sedimentation rate are major criteria for monitoring activity of 3113 disease and tailoring the corticosteroid treatment regimen. Aspirin does alter those parameters, but past clinical experience illustrates one major lapse: contrary to the experience with corticosteroids, aspirin does not prevent blindness (2–6). Weyand et al present an intriguing perspective on a potential mechanism of action of aspirin, but caution is required in applying this information to clinical management of patients with giant cell arteritis. Bruce M. Rothschild, MD Arthritis Center of Northeast Ohio Northeastern Ohio Universities College of Medicine Youngstown, OH 1. Weyand CM, Kaiser M, Yang H, Younge B, Goronzy JJ. Therapeutic effects of acetylsalicylic acid in giant cell arteritis. Arthritis Rheum 2002;46:457–66. 2. Chavany JA, Taptas JN. A propos d’un cas d’arterite temporale. Presse Med 1948;56:835–45. 3. Harrison CV. Giant cell or temporal arteritis: a review. J Clin Pathol 1948;1:1–24. 4. Kalliomaki JL, Lauren PA. Development of temporal arteritis in a patient with rheumatoid arthritis during treatment with indomethacin. Acta Rheum Scand 1965;11:131–6. 5. Fernandez-Herlihy L. Polymyalgia rheumatica. Semin Arthritis Rheum 1971;1:236–45. 6. Hunder GG, Disney TF, Ward LE. Polymyalgia rheumatica. Mayo Clin Proc 1969;44:849–75. DOI 10.1002/art.10523 Reply To the Editor: We agree with Dr. Rothschild’s cautionary note that glucocorticoids are the gold standard in the treatment of giant cell arteritis (GCA), and it was certainly not the intention of our experimental study to advocate replacing corticosteroids with aspirin. However, we also feel the need to optimize current treatment with glucocorticoids (1–3); to accomplish that goal, targeting of relevant pathomechanisms will be necessary. One of the critical pathways is the tissue production of interferon-␥ (INF␥) (4,5), which is relatively spared by corticosteroids but is targeted by acetylsalicylic acid (3). Tissue ischemia in GCA results from aggressive intimal hyperplasia, which cannot be reversed by any immunosuppressive or antiinflammatory interventions (6). It is correct that the frequency of visual ischemic symptoms and blindness has decreased since the introduction of glucocorticoids for treatment. However, it is also common experience that even in patients receiving corticosteroids, especially the low doses used to manage polymyalgia rheumatica, sporadic cases of fullblown arteritis and blindness can occur. The anecdotal cases cited by Dr. Rothschild, therefore, have only limited value. Controlled studies on the use of aspirin are not yet available. More importantly, most of the patients referenced by Dr. Rothschild had polymyalgia rheumatica (7–10) and were receiving nonsteroidal antiinflammatory agents, such as indomethacin and phenylbutazon, not aspirin. Our experiments clearly show that the action of aspirin on IFN␥ production is not related to its cyclooxygenase inhibitory activity. Indeed, indomethacin could not suppress tissue IFN␥. 3114 LETTERS Our molecular experiments demonstrated that corticosteroids and aspirin have distinct targets. These 2 agents should complement each other, not replace each other. In fact, each of them could have unique benefits in terms of inhibiting the different components of the GCA syndrome. Aspirin is minimally effective in inhibiting transcription of nuclear factor–dependent cytokines, such as interleukin-1 (IL-1) and IL-6. It may, therefore, not have an immediate effect on the production of these monokines in the vessel wall. It is also likely much less effective than corticosteroids in treating the acute-phase response and the systemic component of GCA and polymyalgia rheumatica. However, aspirin reduces IFN␥ production in the vessel wall and may, therefore, be able to suppress the progression of intimal hyperplasia. In short-term treatment of GCA, the major benefit of aspirin may relate to its ability as a platelet aggregation inhibitor. In long-term treatment, aspirin may have a role as a steroid-sparing agent, but it will not replace the need for corticosteroids. Naturally, such treatment recommendations cannot be based solely on experimental models but need to be tested in clinical studies. Cornelia M. Weyand, MD Jörg J. Goronzy, MD Mayo Clinic Rochester, MN 1. Weyand CM, Fulbright JW, Hunder GG, Evans JM, Goronzy JJ. Treatment of giant cell arteritis: interleukin-6 as a biologic marker of disease activity. Arthritis Rheum 2000;43:1041–8. 2. Weyand CM, Fulbright JW, Evans JM, Hunder GG, Goronzy JJ. Corticosteroid requirements in polymyalgia rheumatica. Arch Intern Med 1999;159:577–84. 3. Brack A, Rittner HL, Younge BR, Kaltschmidt C, Weyand CM, Goronzy JJ. Glucocorticoid-mediated repression of cytokine gene transcription in human arteritis-SCID chimeras. J Clin Invest 1997;99:2842–50. 4. Weyand CM, Goronzy JJ. Pathogenic principles in giant cell arteritis. Int J Cardiol 2000;75 Suppl 1:S9–S15; discussion S17–9. 5. Weyand CM, Goronzy JJ. Arterial wall injury in giant cell arteritis. Arthritis Rheum 1999;42:844–53. 6. Kaiser M, Weyand CM, Bjornsson J, Goronzy JJ. Platelet-derived growth factor, intimal hyperplasia, and ischemic complications in giant cell arteritis. Arthritis Rheum 1998;41:623–33. 7. Harrison CV. Giant cell or temporal arteritis: a review. J Clin Pathol 1948;1:1–24. 8. Kalliomaki JL, Lauren PA. Development of temporal arteritis in patients with rheumatoid arthritis during treatment with indomethicin. Acta Rheum Scand 1965;11:131–6. 9. Fernandez-Herlihy L. Polymyalgia rheumatica. Semin Arthritis Rheum 1971;1:236–45. 10. Hunder GG, Disney TF, Ward LE. Polymyalgia rheumatica. Mayo Clin Proc 1969;44:849–75.