Dear Editor,

We read with interest the analysis of pro- and anti-inflammatory cytokines in pregnant women with chronic inflammatory arthritis, such as rheumatoid arthritis (RA), juvenile idiopathic arthritis (JIA) and ankylosing spondylitis, recently described by Oestensen et al [1].

Besides direct measurement of cytokine serum levels, they analysed the levels of soluble CD30 (sCD30), a surface protein released by CD30+ T cells which have been associated with the production of Th2-type anti-inflammatory cytokines (in particular IL-4 and, in part, IL-10) [2,3]. High levels of sCD30, that could represent, therefore, an indirect marker of counter-regulatory cell activity, have been described in several chronic inflammatory rheumatic disorders, including RA [3,4]. In their study, however, Oestensen et al. did not find any differences in sCD30 serum levels between RA and control non-pregnant women. Moreover, in contrast to the expected finding of increased sCD30 levels during pregnancy, that usually provides beneficial effect on RA disease activity, a progressive decrease of serum levels in the course of pregnancy was observed, although median of sCD30 from all sera of RA pregnant women demonstrated higher levels than those from non-pregnant RA and control subjects. These apparently conflicting data may be explained by some observations. High serum levels of sCD30 in RA, particularly in patients with circulating rheumatoid factor, have been confirmed in different studies, but the sCD30 concentration reported is very wide [4-7]. Indeed, the median value we found in a large series of 148 RA patients was 79.7 U/ml, significantly different from the values of age- and sex-matched healthy controls (19.1 U/ml, p<_0.001 but="but" ranging="ranging" from="from" _9.5="_9.5" to="to" _515="_515" u="u" ml.="ml." p="p"/> Longitudinal evaluations performed in several of these patients usually showed steady levels over time in subjects with stable disease activity, but showed significant changes after starting remission-inducing treatments, such as anti-TNFa-blockade (unpublished data). It is important to note, however, that these changes are essentially observed in subjects with medium-high levels of sCD30, whereas low basal values remain persistently stable irrespective of disease activity.

In the Oestensen’s study, median of sCD30 values in non-pregnant women is rather low and this may be due to both low number and different selection of the patients, for example prevalent inclusion of patients with seronegative RA, mild disease or in remission. In addition, the pooling of samples from the three trimesters of pregnancy may be misleading and, in our opinion, the very low number of pre-pregnant women does not allow to draw any conclusion. Finally, the inclusion in RA group of patients with JIA may be not correct due to possible different roles exerted by CD30+ T cells in the JIC subgroups [6]. The persistently low levels of sCD30 in the longitudinal evaluation are in agreement with the observation of stable values of sCD30 when they are around the normal range and the slight decrease may be due in fact to enhancement of plasma volume due to pregnancy, as suggested by the Authors. This finding is in line with our analysis in two RA women followed-up before and during pregnancy, whose sCD30 values remained firmly below 25 U/ml also after delivery (Table 1). On the contrary, an increase of sCD30 values was observed during pregnancy in other 4 RA patients with basal sCD30 levels ranging from 60 to 348 U/ml. The increase was already observed in the first trimester and persisted until the delivery. The lack of correlation of sCD30 circulating levels and disease activity scores in the Oestensen’ RA cohort is not surprising, since an inverse correlation between an inflammation marker, such as C-reactive protein, and sCD30 has been found only in patients with early disease and high sCD30 serum levels before starting disease modifying anti-rheumatic drugs. [8]

Although we are unable to explain the reasons of the absence of significant levels of sCD30 in a subset of RA patients, the high levels found in 4 of our 6 pregnant patients support the postulated counter-regulatory role of CD30+ T cells in RA. In addition, these findings point out the complex cytokine network characterizing RA synovitis, where spontaneous or physiological events, such as pregnancy, and/or treatment-induced prevalence of pro- or anti-inflammatory activity lead to disease flare or remission. [9] In this setting, we agree with Oestensen’s statement that simultaneous evaluation of markers not only of Th2-, but also of Th1-type cytokine-producing lymphocytes actively involved in RA synovitis, such as CD26+ cells,10,11 may help to better define the imbalance present in rheumatoid synovial microenvironment.

References

(1). Oestensen M, Foerger F, Nelson JL, et al. Pregnancy in patients with rheumatic diseases : Anti.inflammatory cytokines increase in pregnancy and decrease post-partum. Ann Rheum Dis Published Online First November 11, 2004. doi:10:1136/ard.2004.029538.

(2). Gerli R, Pitzalis C, Bistoni O, et al. CD30+ T cells in rheumatoid synovitis: mechanisms of recruitment and functional role. J Immunol 2000;164:4399-407.

(3). Gerli R, Lunardi C, Vinante F, et al. Role of CD30+ T cells in rheumatoid arthritis: a counter-regulatory paradigm for Th1-driven diseases. Trends Immunol 2001;22:72-7.

(4). Gerli R, Muscat C, Bistoni O, et al. High levels of the solubile form of CD30 molecule in rheumatoid arthritis (RA) are expression of CD30+ T cells involvement in the inflamed joints. Clin Exp Immunol 1995;102:547-50.

(5). Ichikawa Y, Yoshida M, Yamada C, et al. Circulating soluble CD30 levels in primary Sjögren’s syndrome, SLE and rheumatoid arthritis. Clin Exp Rheumatol 1998;16:759-60.

(6). de Kleer M, Kamphuis SM, Rijkers GT, et al. The spontaneous remission of juvenile idiopathic arthritis is characterized by CD30+ T cells directed to human heat-shock protein 60 capable of producing the regulatory cytokine interleukin-10. Arthritis Rheum 2003;48:2001-10.

(7). Okamoto A, Yamamura M, Iwahashi M, et al. Pathophysiological functions of CD30+ CD4+ T cells in rheumatoid arthritis. Acta Med Okoyama 2003;57:267-77.

(8). Gerli R, Bistoni O, Lunardi C, et al. Solubile CD30 in early rheumatoid arthritis as a predictor of good response to second-line therapy. Rheumatology 1999;38:1282-4.

(9). Gerli R, Lunardi C, Pitzalis C. Unmasking the anti-inflammatory cytokine response in rheumatoid synovitis. Rheumatology 2002;41:1341-5.

(10). Muscat C, Bertotto A, Agea E, et al. Expression and functional role of 1F7 (CD26) antigen on peripheral blood and synovial fluid T cells in rheumatoid arthritis patients. Clin Exp Immunol 1994;98:252-6.

(11). Gerli R, Muscat C, Bertotto A, et al. CD26 surface molecule involvement in T cell activation and lymphokine synthesis in rheumatoid and other inflammatory synovitis. Clin Immunol Immunopathol 1996;80:31-7.

Table 1: Serum levels of sCD30 (U/ml) in 6 RA patients before, during and after pregnancy

Dear Editor,

We read with interest the observational, retrospective study by Fautrel and colleagues on the efficacy of anti-TNFƒÑ agents (aTNF) in refractory adult onset Still¡¦s disease (AOSD) [1].

The authors conclude that aTNF is not as effective in AOSD as in rheumatoid arthritis (RA) or the spondyloarthrotopathies and may be helpful only in some AOSD cases. This conclusion cannot be fully supported by the results presented and we would like to suggest an alternative interpretation.

Firstly, it is difficult to draw parallels between AOSD and RA studies, due to differences in the nature of these conditions, severity of disease on inclusion, response criteria used, and overall study designs. If anything, complete remission with aTNF appears commoner in this study of AOSD than in multiple RA trials. Secondly, besides the retrospective design, there are other limitations that would also suggest that any conclusions should be viewed with caution. These include the significant heterogeneity of concomitant treatment, the exclusively clinician-based, subjective, assessment of response, possible slight overrepresentation of females and inclusion of 3 patients that may have had juvenile Still¡¦s disease [2] rather than AOSD.

More importantly, there may be a differential response to the two aTNF agents used. Excluding patients with juvenile Still¡¦s disease, Etanercept was used as the first agent in 9 AOSD patients: 2 failed to respond, 6 achieved partial remission, and 1 achieved complete remission (with concomitant prednizolone 80 mg/day!). In contrast, in the 9 AOSD patients who received infliximab as the first agent, there were 5 partial remissions, 4 complete remissions and no failures (chi-squared test for trend=3.846, P=0.0499). Infliximab was given in one additional AOSD patient after failure of Etanercept and resulted in partial response, while 2 other patients had also Infliximab-induced partial responses after a similar Etanercept-induced response. It is interesting that the only 3 patients who failed to respond to Infliximab are those who could be classified as juvenile Still¡¦s disease.

Differential efficacy of aTNF agents is well-described in Crohn¡¦s disease [3], may also occur in Adamantiades-Behcet¡¦s disease [4, 5] and, we suggest, AOSD. Allowing for reporting bias, all of the AOSD cases reported thus far appear to have had benefit from Infliximab [1, 6-10], even cases refractory to pulse steroid therapy. We suggest that, on current evidence, Infliximab is a very reasonable treatment choice in refractory AOSD. The possibility of a differential response between Infliximab and Etanercept in AOSD requires further investigation.

References

(1). Fautrel B, Sibilia J, Mariette X, Combe B. Tumour necrosis factor (alpha) blocking agents in refractory adult Still¡¦s disease: an observational study of 20 cases. Ann Rheum Dis 2005; 64:262-6

(2). Yamaguchi M, Ohta A, Tsunematsu T, Kasukawa R, Mizushima Y, Kashiwagi H et al. J Rheumatol 1992; 19: 424-30

(3). van Deventer SJH. Transmembrane TNF-alpha, induction of apoptosis, and the efficacy of TNF-targeting therapies in Crohn's disease. Gastroenterology 2001; 121:1242-6

(4). Sfikakis PP. Behcet's disease: a new target for anti-tumour necrosis factor treatment. Ann Rheum Dis 2002; 61(suppl 2):ii51-3

(5). Sfikakis PP, Kaklamanis PH, Elezoglou A, Katsilambros N, Theodossiadis PG, Papaefthimiou S, et al. Infliximab for recurrent, sight- threatening ocular inflammation in Adamantiades-Behcet disease. Ann Intern Med 2004; 140: 404-6

(6). Cavagna L, Caporali R, Epis O, Bobbio-Pallavicini F, Montecucco C. Infliximab in the treatment of adult Still's disease refractory to conventional therapy. Clin Exp Rheum 2001; 19: 329-32

(7). Caramaschi P, Biasi D, Carletto A, Bambara LM. A case of adult onset Still's disease treated with infliximab. Clin Exp Rheumatol 2002; 20: 113

(8). Dechant C, Schauenberg P, Antoni CE, Kraetsch HG, Kalden JR, Manger B. Longterm outcome of TNF blockade in adult-onset Still's disease. Dtsch Med Wochenschr 2004; 129: 1308-12

(9). Bonilla Hernan MG, Cobo Ibanez T, de Miguel Mendieta E, Martin- Mola E. Infliximab (anti-TNF alpha) treatment in patients with adult Still¡¦s disease. Experience in 2 cases. An Med Interna 2004; 21: 23-6

(10). Kokkinos A, Iliopoulos A, Greka P, Efthymiou A, Katsilambros N, Sfikakis PP. Successful treatment of refractory adult-onset Still's disease with infliximab. A prospective, non-comparative series of four patients. Clin Rheumatol 2004; 23: 45-9.

Dear Editor,

Fautrel et al [1] recently reported the cases of 20 adults with Still's disease who were treated with TNF alpha blocking agents. Most patients only achieved partial remission and there were only 2 cases of sustained remission of more than one year, which contrasts with the much higher response rates observed in patients with rheumatoid arthritis.

In this paper, the authors never refer to the pediatric experience. However, they indicate that the disease had a childhood onset in 4 of their patients and they insist on the scantiness of available data in adults. Although the disease is also rare in children, several studies had been published reporting the experience of etanercept treatment in more than one hundred patients with systemic-onset juvenile rheumatoid arthritis (JRA)/juvenile idiopathic arthritis (JIA), i.e. childhood-onset Still's disease. [2-5] Moreover, we and others showed that these patients were significantly less responsive to anti-TNF alpha therapy than patients with other forms of JIA. [4-5]

When dealing with the treatment of rare diseases, pediatricians and rheumatologists have to consider each others' experience.

References

(1). Fautrel B, Sibilia J, Mariette X, Combe B. Tumour necrosis factor a blocking agents in refractory adult Still's disease: an observational study of 20 cases. Ann Rheum Dis. 2005;64:262-266.

(2). Lovell DJ, Giannini EH, Reiff A, Cawkwell GD, Silverman ED, Nocton JJ, et al. Etanercept in children with polyarticular juvenile rheumatoid arthritis. Pediatric Rheumatology Collaborative Study Group. N Engl J Med. 2000;342:763-9.

(3). Lovell DJ, Giannini EH, Reiff A, Jones OY, Schneider R, Olson JC, et al. Long-term efficacy and safety of etanercept in children with polyarticular-course juvenile rheumatoid arthritis: interim results from an ongoing multicenter, open-label, extended-treatment trial. Arthritis Rheum. 2003;48:218-26.

(4). Quartier P, Taupin P, Bourdeaut F, Lemelle I, Pillet P, Bost M, et al. Efficacy of etanercept for the treatment of juvenile idiopathic arthritis according to the onset type. Arthritis Rheum. 2003;48:1093-101.

(5). Horneff G, Schmeling H, Biedermann T, Foeldvari I, Ganser G, Girschick HJ, et al. The German etanercept registry for treatment of juvenile idiopathic arthritis. Ann Rheum Dis. 2004;63:1638-44.

Dear Editor,

Tubach and colleagues [1] propose an 8-item shortened version of the WOMAC physical function subscale which they consider promising for clinical use and large scale surveys and clinical trials.

The authors report that the shortened measure’s test retest reliability and responsiveness are similar to those of the complete physical function subscale; the internal consistency of the shortened measure (0.84) is less than that of the full length measure (0.93). I have several questions concerning the methods:
1. Was the entire WOMAC, including pain and stiffness subscales, administered?
2. Were responses to the shortened version abstracted from the lengthened version?
3. If the answer to the previous question is “yes,” was the difference in sample size for the responsiveness analysis (shortened version n=1169; full length n=1048) a result of missing item values?
4. Would it be possible to provide descriptive statistics (mean, standard deviation) for the shortened version?
5. Could the authors present the standard error of measurement associated with the internal consistency estimates?

The answer to the last question would be helpful in labeling the measurement error at an instant in time and in ascertaining the consequence of using the shortened version to make confident—in a reliability sense—decisions concerning an individual patient’s functional status at an instant in time.

Although the WOMAC has been the principal self-report functional status measure for patients with osteoarthritis of the hip or knee for 2- decades, there is consistent evidence in the literature refuting the measure’s factorial validity [2-5] The WOMAC purports three subscales— pain, stiffness, physical function—yet the pain and function subscales have been repeatedly shown to overlap [2-5]. That pain and function are not completely distinct is to be expected; however, the OMERACT III group deemed them distinct enough to warrant independent assessment [6]. At issue is the extent to which the overlap in responses to the pain and function subscales is real or biased due to a common item phrasing in pain and function items [4]. It would be informative to know whether the proposed shortened function scale represented a factor distinct from pain in the study sample.

As a matter of interest I ran a factor analysis on the pain and shortened function scale proposed by the authors on a sample of 310 patients with osteoarthritis of the hip or knee reported previously in the peer reviewed literature [4]. I ran two analyses with oblique rotation, one used the eigenvalue greater than 1 rule and the other restricted the analysis to two factors. The eigenvalue greater than 1 rule identified three factors. Neither analysis grouped the items by purported subscales of pain and function.

The consequence of poor factorial validity is that one may not be assessing what she/he is believed to be assessing. To examine the consequence of the apparently poor factorial validity of the proposed shortened version of the WOMAC function scale, I examined the measure’s ability to detect deterioration in a group of patients undergoing total hip or knee arthroplasty. This sample is described in a previous paper [7]. The rationale being that physical function deteriorates between preoperative and postoperative assessment performed within 16 days of surgery. I repeated the analysis reported previously in a peer- reviewed paper [7] (n=104) and found the following for the proposed shortened physical function subscale: mean pre-operation 15.7 (sd=5.2); mean post-operation (median assessment interval was 8 days, 1st, 3rd quartiles were 7, 10 days) 15.9 (sd=5.6); and the standardized response mean was –0.02 (95% CL: -0.23, 0.17). That the confidence interval for the standardized response mean includes the value zero indicates the measure was not able to detect deterioration in this group.

Tubach and colleagues acknowledge that further evaluation of their shortened version is required. I believe an important next step is to ascertain whether the measure possesses factorial validity.

References

(1). Tubach F, Baron G, Logeart I, Dougados M, Bellamy N, Ravaud P. Using patients' and rheumatologists' opinions to specify a short form of the WOMAC function subscale. Ann Rheum Dis 2005;64:75-79.

(2). Faucher M, Poiraudeau S, Lefevre-Colau MM, Rannou F, Fermanian J, Revel M. Algo-functional assessment of knee osteoarthritis: comparison of the test-retest reliability and construct validity of the WOMAC and Lequesne indexes. Osteoarthritis Cartilage 2002;10:602 -10.

(3). Thumboo J, Chew LH, Soh CH. Validation of the Western Ontario and Mcmaster University osteoarthritis index in Asians with osteoarthritis in Singapore. Osteoarthritis Cartilage 2001;9:440-6.

(4). Stratford PW, Kennedy DM. Does parallel item content on WOMAC's Pain and Function Subscales limits its ability to detect change in functional status? BMC Musculoskelet Disord 2004;5:17.

(5). Guermazi M, Poiraudeau S, Yahia M, Mezganni M, Fermanian J, Habib Elleuch M, et al. Translation, adaptation and validation of the Western Ontario and McMaster Universities osteoarthritis index (WOMAC) for an Arab population: the Sfax modified WOMAC. Osteoarthritis Cartilage 2004;12:459-68.

(6). Bellamy N, Kirwan J, Boers M, Brooks P, Strand V, Tugwell P, et al. Recommendations for a core set of outcome measures for future phase III clinical trials in knee, hip, and hand osteoarthritis. Consensus development at OMERACT III. J Rheumatol 1997;24:799-802.

(7). Stratford PW, Kennedy DM, Hanna SE. Condition-specific Western Ontario McMaster Osteoarthritis Index was not superior to region- specific Lower Extremity Functional Scale at detecting change. J Clin Epidemiol 2004;57:1025-1032.