Dear Editor,

We read with interest the retrospective case-control study by de Jong et al.(1) The authors found an intriguing association between statin use and development of rheumatoid arthritis (RA). Some comments regarding the methodology and authors statements, may be appropriate.

Evaluating treatment effects in observational studies can be problematic as prognostic factors influencing treatment decisions can cause so-called "confounding by indication". The baseline characteristics show that the prevalence of cardiovascular co-morbidity is considerably higher in cases versus controls. Since statin therapy is standard care in people with a high cardiovascular risk, this could explain more statin use in the RA group. Moreover, one study showed that the lipid profile of subjects who later developed RA is more atherogenic than of subjects who did not develop RA.(2) As the authors have correctly mentioned in the discussion, data on hyperlipidemia is limited. We therefore disagree with the statement that statins may promote the development of RA. To our opinion, this association could be based on the underlying predisposition to develop hyperlipidemia and cardiovascular disease, which is already present long before individuals develop RA.

Another point of concern is the selection of the study population. The investigators selected all patients with the ICPC code L88 (arthritis) if they were referred to a rheumatologist or had had a prescription of at least one DMARD or two or more prescriptions of corticosteroids.
In the Netherlands many patients are referred to the rheumatologist and only 10-20% has indeed RA. Furthermore, corticosteroids and DMARDs are prescribed for many (rheumatic) diseases other than RA. This could mean that patients with polymyalgia rheumatica, psoriatic arthritis or even systemic lupus erythematosis (SLE) are included in the cases group. In part, this problem could have been solved by verification of the diagnosis in the medical record. Furthermore, the controls do not seem to be matched for general practitioner (GP). Some GP's are more keen on cardiovascular risk management and prescribe statins more frequently. Would multilevel logistic regression analyses not have been more appropriate to correct for this sampling bias?

As to the underlying mechanisms behind the association, the authors suggest that statins increase the production of autoantibodies by promoting a shift in the TH1/TH2 balance. However, Kanda et al. demonstrated a reduction of the Th1/Th2 and CD4/CD8 ratios, rheumatoid factor, erythrocyte sedimentation rate and C-reactive protein levels after statin use, implying an anti-inflammatory effect and down-regulation of T-cell mediated immune responses(3). To date there are more studies showing an anti-inflammatory (4) and beneficial effect on onset and activity of Th1 mediated diseases like RA(5) and multiple sclerosis(6) than case reports showing a possible triggering effect of statins on auto-immune diseases like autoimmune hepatitis(7) and SLE(8).

Despite the possibility of confounding by indication, this study raises important questions regarding the role of statins in the development of RA. We definitely concur more evidence is needed to confirm this association before statins are withheld, since cardiovascular disease is a prominent health problem especially among individuals with prodromal RA.

References

1. de Jong HJI, Klungel OH, van Dijk L, et al. Use of statins is associated with an increased risk of rheumatoid arthritis. Ann Rheum Dis 2011 Oct 6. epub ahead of print.

2. van Halm VP, Nielen MM, Nurmohamed MT, et al. Lipids and inflammation: serial measurements of the lipid profile of blood donors who later developed rheumatoid arthritis. Ann Rheum Dis 2007;66:184-8.

3. Kanda H, Yokota K, Kohno C, et al. Effects of low-dosage simvastatin on rheumatoid arthritis through reduction of Th1/Th2 and CD4/CD8 ratios. Mod Rheumatol 2007;17:364-8.

4. McCarey DW, Sattar N, McInnes IB. Do the pleiotropic effects of statins in the vasculature predict a role in inflammatory diseases? Art Res Ther 2005;7:55-61.

5. Jick SS, Choi H, Li L et al. Hyperlipidaemia, statin use and the risk of developing rheumatoid arthritis. Ann Rheum Dis 2008;68:546-551.

6. Willey JZ, Elkind MSV. 3-hydroxy-3-methylglutaryl-coenzyme A reductase inhibitors in the treatment of central nervous system diseases. Arch Neurol 2010;67(9):1062-1067.

7. Alla V, Abraham J, Siddiqui J, et al. Autoimmune hepatitis triggered by statins. J Clin Gastroenterol 2006;40(80):757-61.

8. Noel B. Lupus erythematosus and other autoimmune diseases related to statin therapy: a systematic review. J Eur Acad Dermatol Venereol 2007;21:17-24.

Conflict of Interest:

None declared

Wang-Dong Xu, Yu-Jing Zhang, Hai-Feng Pan, Dong-Qing Ye*

Department of Epidemiology and Biostatistics, School of Public Health, Anhui Medical University, 81 Meishan Road, Hefei, Anhui, 230032, PR China;

*Correspondence: Dong-Qing Ye, M.D, Department of Epidemiology and Biostatistics, School of Public Health, Anhui Medical University, 81 Meishan Road, Hefei, Anhui, 230032, PR China,

E-mail: ydq@ahmu.edu.cn; Tel.: +86 551 5167726; Fax: +86 551 5161171.

Dear editor,

McMahon, et al1 have reported that high plasma leptin levels confer increased risk of atherosclerosis in systemic lupus erythematosus (SLE) patients, and are associated with inflammatory oxidised lipids. In the study, leptin levels are significantly higher in patients than in controls, which are also higher in the 43 SLE patients with plaque than without plaque and associate with inflammatory biomarkers of atherosclerosis such as piHDL (proinflammatory high-density lipoprotein), Lp(a) and OxPL/apoB100, furthermore, age and hypertension, and so on.

In fact, relationship between leptin and SLE have been studied increasingly both in humans and animal models, but different results exist2-10. Hahn, et al4 developed a mouse model of multigenic lupus exposed to environmental factors known to accelerate atherosclerosis in humans--high-fat diet with or without injections of the leptin, where the lupus-prone strain (BWF1 mice), addition of leptin increased proinflammatory high-density lipoprotein scores, indicating the presence of piHDL as well as atherosclerosis, accelerated proteinuria. Moreover, several investigations have confirmed the higher leptin levles are significantly associated with SLE compared to controls6,8,10, especially the serum leptin levels with age, hypertension3,5,8,9, but Chung, et al5 found no significant relationship between leptin or adiponection levels and coronary calcification in SLE pantients and controls. On contrary, lower leptin levels in SLE patients4 or no statistically significant differences between leptin levels in SLE patients and the controls7 are discovered as well.

As is known, leptin belongs to the type I cytokine superfamily and has an important role in regulating immune functions. Leptin protects T lymphocytes from apoptosis and regulates T-cell proliferation and activation, and influences cytokine production from T lymphocytes, generally switching the phenotype toward a Th1 response 11-13. Consequently, in order to better understand the link between SLE, atherosclerosis and leptin, further to clarify the associaiton of leptin and SLE based on large samples not only in humans, but in animals are still needed.

References

1. McMahon M, Skaggs BJ, Sahakian L, et al. High plasma leptin levels confer increased risk of atherosclerosis in women with systemic lupus erythematosus, and are associated with inflammatory oxidised lipids. Ann Rheum Dis 2011;70:1619-24.

2. Hahn BH, Lourenco EV, McMahon M, et al. Pro-inflammatory high- density lipoproteins and atherosclerosis are induced in lupus-prone mice by a high-fat diet and leptin. Lupus 2010;19:913-7.

3. Kim HA, Choi GS, Jeon JY, et al. Leptin and ghrelin in Korean systemic lupus erythematosus. Lupus 2010;19:170-4.

4. De Sanctis JB, Zabaleta M, Bianco NE, et al. Serum adipokine levels in patients with systemic lupus erythematosus. Autoimmunity 2009;42:272-4.

5. Chung CP, Long AG, Solus JF, et al. Adipocytokines in systemic lupus erythematosus: relationship to inflammation, insulin resistance and coronary atherosclerosis. Lupus 2009;18:799-806.

6. Vadacca M, Margiotta D, Rigon A, et al. Adipokines and systemic lupus erythematosus: relationship with metabolic syndrome and cardiovascular disease risk factors. J Rheumatol 2009;36:295-7.

7. Wisowska M, Rok M, StepieK, et al. Serum leptin in systemic lupus erythematosus. Rheumatol Int 2008;28:467-73.

8. Ryan MJ, McLemore GR Jr, Hendrix ST. Insulin resistance and obesity in a mouse model of systemic lupus erythematosus. Hypertension 2006;48:988-93.

9. Sada KE, Yamasaki Y, Maruyama M, et al. Altered levels of adipocytokines in association with insulin resistance in patients with systemic lupus erythematosus. J Rheumatol 2006;33:1545-52.

10. Garcia-Gonzalez A, Gonzalez-Lopez L, Valera-Gonzalez IC, et al. Serum leptin levels in women with systemic lupus erythematosus. Rheumatol Int 2002;22:138-41.

11. Fantuzzi G. Adipose tissue, adipokines, and inflammation. J Allergy Clin Immunol 2005;115:911-9.

12. Zhang Y, Proenca R, Maffei M, et al. Positional cloning of the mouse obese gene and its human homologue. Nature 1994;372:425-32.

13. Busso N, So A, Chobaz-Peclat V, et al. Leptin signaling deficiency impairs humoral and cellular immune responses and attenuates experimental arthritis. J Immunol 2002;168:875-82.

Conflict of Interest:

None declared

Dear Editor,

Sieper et al. propose a set of early referral criteria for ankylosing spondylitis (AS) using HLAB27 as an central test.[1] The supporting data were presented in a previous paper.[2] The HLAB27 data were taken from six study populations. In two groups the control population were either symptom free blood donors or no clinical data was known.[3, 4] There are three published trials using back pain controls with ankylosing spondylitis (AS) patients.[5-7] When reviewed these produce lower figures for sensitivity and specificity than are quoted in the paper by Rudwaleit et al. A further trial with back pain controls has been seen only in abstract form and the patient numbers quoted do not match those from the abstract.[8] If using only the data from studies with back pain controls the recalculated post test probability is 19% so that the utility of HLAB27 as a screening and diagnostic tool is greatly reduced .

The authors have used a history of inflammatory back pain[9] as the entry point for both their diagnostic and referral algorithms. However in the most recent paper the initial questions are reduced from five to two. The full set are suggested only if the HLAB27 is difficult to perform. This recommendation of HLAB27 as the primary test in determining referral is of concern. While it may ensure most patients are seen early by a rheumatologist it will certainly overwhelm many of the assessing clinics. Even in moderately well funded health systems the level of service provision is already far below recommendation.[10]

Some years ago HLAB27 positive mechanical back pain was a regular cause for referral to rheumatology clinics [11] and we would not wish to return to this situation. Further, HLAB27 testing may not be routinely available in primary care in many areas. Given this we believe that a well taught clinical history is still the best (albeit poor) indicator of inflammatory spinal disease in primary care.

References

1. Sieper, J. and M. Rudwaleit, Early referral recommendations for ankylosing spondylitis (including pre-radiographic and radiographic forms) in primary care. Ann Rheum Dis, 2005. 64(5): p. 659-63.

2. Rudwaleit, M., et al., How to diagnose axial spondyloarthritis early. Ann Rheum Dis, 2004. 63(5): p. 535-43.

3. Schlosstein, L., et al., High association of an HL-A antigen, W27, with ankylosing spondylitis. N Engl J Med, 1973. 288(14): p. 704-6.

4. Brewerton, D.A., et al., Ankylosing spondylitis and HL-A 27. Lancet, 1973. 1(7809): p. 904-7.

5. Sadowska-Wroblewska, M., et al., Clinical symptoms and signs useful in the early diagnosis of ankylosing spondylitis. Clin Rheumatol, 1983. 2(1): p. 37-43.

6. Mau, W., et al., Clinical features and prognosis of patients with possible ankylosing spondylitis. Results of a 10-year followup. J Rheumatol, 1988. 15(7): p. 1109-14.

7. Braun, J., et al., Prevalence of spondylarthropathies in HLA-B27 positive and negative blood donors. Arthritis Rheum, 1998. 41(1): p. 58- 67.

8. Rudwaleit, M., et al., Clinical parameters in the differentiation of inflammatory back pain from non-inflammatory back pain. Ann Rheum Dis, 2002. 61 ((suppl 1)): p. 57.

9. Calin, A., et al., Clinical history as a screening test for ankylosing spondylitis. Jama, 1977. 237(24): p. 2613-4.

10. Harrison, A., Provision of rheumatology services in New Zealand. N Z Med J, 2004. 117(1192): p. U846.

11. Helliwell, P.a.W., V., Seronegative spondyloarthropathies, in Clinical Rheumatology International Practice and Research: Epidemiological, Sociological and Environmental Aspects of Rheumatology, J.A.D. Anderson, Editor. 1987, Balliere Tindall: London. p. 491-524.

Dear Editor,

Felson et al make a good case for choosing continuous, rather than dichotomous, variables as the primary outcome measures in randomised controlled trials in rheumatoid arthritis but outcome measure also need to be meaningful to the clinical community. In TICORA we chose two ‘co-primary’ end-points: firstly, we employed the mean change in DAS because this was the most sensitive outcome measure available at the time; we also used the achievement of a “good” response according to EULAR criteria because we believed that differences in response rates would be more readily understood. Interestingly, our understanding of the clinical relevance of the outcome measures has been borne out: the results of the trial are (almost) never quoted as showing that the mean fall in DAS was 1.6 units (95% CI, 1.1 to 2.1) greater in the intensive arm - commentators almost invariably refer instead to the higher response rates. Consequently, I would urge some caution on the clinical trials community not to rely on sensitive (but intuitively difficult to understand) outcome measures to the exclusion of response rates. This is acknowledged by Felson et al when they recognise that such measures can be secondary outcome measures but it may still be important that some trials (depending on their purpose) are large enough to be able to detect clinically significant differences in response rates.

There has been a growing tendency for the therapeutic research agenda to become dominated by the pharmaceutical industry in their pursuit of marketing authorisation for new products. Felson et al point out that using sensitive outcome measures may avoid the need for ‘mega-trials’. Importantly, smaller and cheaper trials will help to safeguard the role of independent, investigator-initiated research which has been so fruitful in recent years. The second impact of designing smaller trials is, however, less helpful because it demands a degree of expertise in interpreting trial results. TICORA was a small trial, and so the magnitude of the benefits could not be assessed with any precision. The odds ratio for an ACR 70 response with intensive therapy was 11, but the 95% confidence interval for that odds ratio was 4.5 to 27. Rather than describe the trial as an ‘outlier’ the authors should perhaps have commented that the results should be interpreted carefully in light of the confidence intervals described in the paper.

I would add one further observation about dichotomous variables when the cohort average is close to a cut-off point. After 12 months, 44% of the TICORA intensive group were in DAS remission (data on file). At this point, the median DAS in the intensive group was 1.7, just above the cut-off for DAS remission. At 18 months, the median DAS had fallen to 1.3, just below the cut-off. This modest improvement in DAS was probably of limited clinical significance yet it resulted in a substantial increase in the remission rate (from 44% to 65%).

In conclusion, I would endorse the authors’ recommendation to use continuous measures of disease activity as primary outcome measures. This should ensure that smaller trials will be possible, and in turn this may help to safeguard the role of independent investigator-initiated research. However, trialists should continue to make sure that their measures are comprehensible and relevant, and the clinical community must be careful to interpret the results of trials thoughtfully.

References

1. D Felson, B Zhang and J Siegel. Trials in rheumatoid arthritis: choosing the right outcome measure when minimal disease is achievable. Ann Rheum Dis 2008;57:580-582.

2. C Grigor, H Capell, A Stirling et al. Effect of a treatment strategy of tight control for rheumatoid arthritis (the TICORA study): a single-blind randomised controlled trial. Lancet 2004;364:263-269.

3. D Porter. Targeting persistent disease activity in early RA: a commentary on the TICORA trial. International Journal of Advances in Rheumatology. 2005;3:2-6.