Browse the corpus

aluation over time. The first aim of this study was to determine whether treatment with infliximab modulates specific biomarkers in patients with AS, and when such modulation may occur. The second aim was to evaluate associations between changes in biomarker levels and disease activity and inflammation detected by MRI. MATERIALS AND METHODS The details of the ASSERT (Ankylosing Spondylitis Study for the Evaluation of Recombinant Infliximab Therapy) study have been published previously.22 Sera from patients who received placebo or 5 mg/kg infliximab were collected for biomarker analysis at weeks 0, 2, and 24. IL-6, VEGF and interferon (IFN)-γ were evaluated as markers of inflammation. Enzyme-linked immunoabsorbent assay kits for IL-6 and VEGF were purchased from R&D Systems (Minneapolis, MN, USA), and those for IFN-γ were purchased from Biosource, Europe S.A. The Tina-quant kit from Roche (Indianapolis, IN, USA) was used to determine CRP levels. All assays were validated by scientists at Clinical Pharmacology and Experimental Medicine, Centocor Research and Development, Inc. before use on study samples. Biomarker levels that were below the lower limit of quantification (LLOQ) were considered to be undetectable.

Ankylosing spondylitis (AS) is associated with persistent inflammation of the sacroiliac joints and spine, which can lead to bone erosions, syndesmophytes or complete ankylosis of the spine. There is evidence that the inflammation in AS is at least partly mediated by tumour necrosis factor (TNF)-α and interleukin (IL)-6, as high levels of these cytokines have been found in biopsies from the sacroiliac joints of patients with AS.1 2 Elevated serum levels of TNF-α and vascular endothelial growth factor (VEGF) have been shown to be correlated with disease activity and C-reactive protein (CRP) levels.3 4 VEGF is an important regulator of angiogenesis, which is key to the inflammatory process. Polymorphisms in the VEGF genes were reported to be associated with disease severity in patients with AS.5 6 IL-6 is a multifunctional cytokine that regulates immune response, haematopoiesis, acute phase response and inflammation. Dysregulation of IL-6 production is implicated in the pathology of several disease processes,7–9 and increased IL-6 levels have been associated with rheumatoid arthritis, systemic-onset juvenile chronic arthritis, osteoporosis and psoriasis.10–12 IL-6 is critically involved in experimentally induced autoimmune disease, such as antigen-induced arthritis, and experimental allergic encephalomyelitis.13 14 Collectively, these data from both clinical studies and animal models suggest that IL-6 plays a critical role in the pathogenesis of immune-mediated diseases,15 supporting the face validity of the use of biomarkers to predict treatment response.

tigen-induced arthritis, and experimental allergic encephalomyelitis.13 14 Collectively, these data from both clinical studies and animal models suggest that IL-6 plays a critical role in the pathogenesis of immune-mediated diseases,15 supporting the face validity of the use of biomarkers to predict treatment response. The course of spinal inflammation associated with AS can be demonstrated by magnetic resonance imaging (MRI).16 17 While a significant reduction in spinal inflammation has been shown with MRI after treatment with anti-TNF-α agents,18–21 a direct link between clinical disease activity and spinal inflammation as seen on MRI and osteoproliferative changes has not been well characterised. In the current study, biomarkers known to be important in the inflammatory and angiogenic processes that occur in AS were selected for evaluation over time. The first aim of this study was to determine whether treatment with infliximab modulates specific biomarkers in patients with AS, and when such modulation may occur. The second aim was to evaluate associations between changes in biomarker levels and disease activity and inflammation detected by MRI.

polis, IN, USA) was used to determine CRP levels. All assays were validated by scientists at Clinical Pharmacology and Experimental Medicine, Centocor Research and Development, Inc. before use on study samples. Biomarker levels that were below the lower limit of quantification (LLOQ) were considered to be undetectable. Disease activity was assessed using the Bath Ankylosing Spondylitis Disease Activity Index (BASDAI), which has a possible score of 0–10 with a higher score indicating greater disease activity.23 Clinical response was assessed by determining the number of patients who achieved a 50% improvement in the BASDAI score (BASDAI 50) and the number of patients who achieved 20% improvement in the ASsessment in Ankylosing Spondylitis working group criteria (ASAS 20).24 MRI was conducted at baseline and week 24 as previously reported.20 Pre- and post-gadolinium T1 and short τ inversion recovery magnetic resonance images of the spine were acquired at baseline and week 24. The activity score for each vertebral unit ranged from 0 to 6, and the total activity score for the spine ranged from 0 to 138 (23 vertebral units from C2 to S1).19 25–27

reported.20 Pre- and post-gadolinium T1 and short τ inversion recovery magnetic resonance images of the spine were acquired at baseline and week 24. The activity score for each vertebral unit ranged from 0 to 6, and the total activity score for the spine ranged from 0 to 138 (23 vertebral units from C2 to S1).19 25–27 Statistical analyses To validate the use of biomarkers in predicting the treatment response in patients with AS who received infliximab, we evaluated the face validity, discriminant capacity, predictive validity and sensitivity to change. Face validity was determined by assessing the relationship between biomarker levels and disease activity at baseline. Discriminant capacity and sensitivity to change was assessed by demonstrating the effect of infliximab treatment on biomarker levels relative to placebo after 2 and 24 weeks of follow-up. Predictive validity was evaluated by determining the association between changes from baseline in disease activity (BASDAI and MRI scores) and biomarker levels at baseline, week 2 and week 24. The median percentage change from baseline for each biomarker was determined at weeks 2 and 24. Statistical comparisons were made between the placebo and infliximab groups using an analysis of variance on the van der Waerden scores. Biomarker data for patients with sample baseline values less than the LLOQ were excluded from analyses.

ntage change from baseline for each biomarker was determined at weeks 2 and 24. Statistical comparisons were made between the placebo and infliximab groups using an analysis of variance on the van der Waerden scores. Biomarker data for patients with sample baseline values less than the LLOQ were excluded from analyses. Univariate Spearman Rank correlations were computed to determine the relationship between baseline levels of IL-6, VEGF and CRP. Correlation analyses were also used to examine relationships between (1) the change from baseline to week 24 in MRI activity score and the biomarker levels (baseline and changes from baseline to week 2 and week 24), and (2) the associations between change from baseline to week 24 in BASDAI and biomarker levels (baseline and changes from baseline to week 2 and week 24).

tionships between (1) the change from baseline to week 24 in MRI activity score and the biomarker levels (baseline and changes from baseline to week 2 and week 24), and (2) the associations between change from baseline to week 24 in BASDAI and biomarker levels (baseline and changes from baseline to week 2 and week 24). To further characterise the relationship between biomarker levels, disease activity and inflammation at baseline, patients were categorised into quantiles of baseline CRP and IL-6, and descriptive statistics of baseline MRI activity and BASDAI scores were calculated for each category. For CRP, patients were divided into tertile categories (<0.9 mg/dl, ⩾0.9 and <2.4 mg/dl, or ⩾2.4 mg/dl) because all patients had detectable levels of CRP at baseline (table 1). Nearly half of the patients were assigned the same baseline IL-6 value (3.13 pg/ml, which was one-half of the LLOQ of 6.25 pg/ml) because their levels were below the LLOQ (table 1). Therefore, patients were divided into categories of IL-6 levels based on the median IL-6 value at baseline (<7.38 pg/ml or >7.38 pg/ml). The percentage of patients who achieved an ASAS 20 or BASDAI 50 response at week 24 was also calculated for each category of baseline CRP and IL-6.

vels were below the LLOQ (table 1). Therefore, patients were divided into categories of IL-6 levels based on the median IL-6 value at baseline (<7.38 pg/ml or >7.38 pg/ml). The percentage of patients who achieved an ASAS 20 or BASDAI 50 response at week 24 was also calculated for each category of baseline CRP and IL-6. Table 1 Baseline characteristics Assessment Placebo n = 78 5 mg/kg Infliximab n = 201 Men, no. (%) 68 (87.2) 157 (78.1) Disease duration, years n 76 201 Mean (SD) 11.9 (8.0) 10.1 (8.7) Median (IQ range) 13.2 (3.7, 17.9) 7.7 (3.3, 14.9) HLA-B27 positive n 78 200 No. (%) 69 (88.5) 173 (86.5) Inflammation (average morning stiffness on a visual analogue scale 0–10 cm) n 78 201 Mean (SD) 6.9 (1.9) 6.9 (2.3) Median (IQ range) 7.0 (5.7, 8.3) 7.3 (5.4, 8.5) BASDAI score n 78 201 Mean (SD) 6.2 (1.6) 6.5 (1.5) Median (IQ range) 6.5 (5.2, 7.1) 6.6 (5.3, 7.6) C-reactive protein (mg/dl) n 78 201 No. (%) with values ⩾LLOQ 78 (100) 201 (100) Mean (SD) 2.4 (2.8) 2.4 (2.7) Median (IQ range) 1.7 (0.7, 3.3) 1.5 (0.7, 3.2) Interleukin-6 (pg/ml) n 67 187 No. (%) with values ⩾LLOQ 32 (47.8) 105 (56.1) Mean (SD) 11.4 (17.2) 13.2 (20.0) Median (IQ range) 3.1 (3.1, 14.8) 7.7 (3.1, 14.2) Vascular endothelial growth factor (pg/ml) n 75 193 No. (%) with values ⩾LLOQ 75 (100.0) 191 (99.0) Mean (SD) 556.2 (385.6) 520.4 (361.5) Median (IQ range) 473.5 (289.5, 738.0) 421.5 (279.9, 664.3) Interferon-γ (pg/ml) n 62 165 No. (%) with values ⩾LLOQ 3 (4.8) 9 (5.5) Mean (SD) 0.7 (1.3) 0.6 (0.3) Median (IQ range) 0.5 (0.5, 0.5) 0.5 (0.5, 0.5) BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; IQ, interquartile; LLOQ, lower limit of quantification.

n (IQ range) 473.5 (289.5, 738.0) 421.5 (279.9, 664.3) Interferon-γ (pg/ml) n 62 165 No. (%) with values ⩾LLOQ 3 (4.8) 9 (5.5) Mean (SD) 0.7 (1.3) 0.6 (0.3) Median (IQ range) 0.5 (0.5, 0.5) 0.5 (0.5, 0.5) BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; IQ, interquartile; LLOQ, lower limit of quantification. A multiple linear regression analysis was performed to explore baseline levels of IL-6, CRP and VEGF as potential indicators of the change from baseline to week 24 in MRI activity scores and the change from baseline to week 24 in BASDAI scores. Similar analyses were performed to evaluate the percentage changes in these biomarkers from baseline to week 2 and from baseline to week 24. Statistical analyses were performed using the SAS system, version 8.2 (SAS Institute, Cary, NC, USA). P-values have been provided for exploratory purposes and were not adjusted for multiplicity. Statistical tests were two-sided, and p<0.05 was considered significant.

A multiple linear regression analysis was performed to explore baseline levels of IL-6, CRP and VEGF as potential indicators of the change from baseline to week 24 in MRI activity scores and the change from baseline to week 24 in BASDAI scores. Similar analyses were performed to evaluate the percentage changes in these biomarkers from baseline to week 2 and from baseline to week 24. Statistical analyses were performed using the SAS system, version 8.2 (SAS Institute, Cary, NC, USA). P-values have been provided for exploratory purposes and were not adjusted for multiplicity. Statistical tests were two-sided, and p<0.05 was considered significant. RESULTS Baseline characteristics and biomarker levels The baseline characteristics of patients in the ASSERT trial are summarised in table 1. The study population was typical of patients with well-established, active AS. As reported previously,22 the majority of the patients in ASSERT were men, most of whom tested positive for the HLA-B27 allele. The median assessment of average morning stiffness on a visual analogue scale was 7.0 in the placebo group and 7.3 in the infliximab group. Although there appeared to be a disparity between the groups in the median disease duration (13.2 years in the placebo group and 7.7 years in the infliximab group), the means were similar (11.9 and 10.1 years, respectively).

ffness on a visual analogue scale was 7.0 in the placebo group and 7.3 in the infliximab group. Although there appeared to be a disparity between the groups in the median disease duration (13.2 years in the placebo group and 7.7 years in the infliximab group), the means were similar (11.9 and 10.1 years, respectively). The treatment groups were also generally comparable for all biomarkers of inflammation at baseline. The median IL-6 level in the placebo group (3.1 pg/ml) was lower than that of the infliximab group (7.7 pg/ml), but the means were comparable (11.4 vs 13.2, respectively). Approximately half of the patients had IL-6 levels that were above the LLOQ for the assay (47.8% and 56.1% of placebo and infliximab groups, respectively). Only 12 patients had IFN-γ levels that were above the LLOQ for the assay (4.8% and 5.5%, respectively). Correlations between individual biomarker levels at baseline and after treatment with infliximab or placebo In all patients, Spearman Rank correlation coefficients among baseline biomarker levels showed a strong and statistically significant relationship between IL-6 and CRP levels (0.698, p<0.001). Moderate correlations between CRP and VEGF (0.362, p<0.001) and IL-6 and VEGF (0.264, p<0.001) were also observed.

b or placebo In all patients, Spearman Rank correlation coefficients among baseline biomarker levels showed a strong and statistically significant relationship between IL-6 and CRP levels (0.698, p<0.001). Moderate correlations between CRP and VEGF (0.362, p<0.001) and IL-6 and VEGF (0.264, p<0.001) were also observed. Correlations among changes in biomarker levels at week 24 were also evaluated. In the infliximab group, there were significant relationships between change from baseline in IL-6 and CRP (r = 0.689, p<0.001), IL-6 and VEGF (r = 0.445, p<0.001), and VEGF and CRP (r = 0.565, p<0.001) at week 24. In the placebo group, only a significant correlation between IL-6 and CRP was observed (r = 0.452, p<0.001). Face validity: baseline biomarker levels and disease activity, peripheral involvement and extra-articular manifestations At baseline, Spearman correlation coefficients between BASDAI scores and IL-6 levels (0.134, p = 0.0327) and BASDAI scores and CRP levels (0.167, p = 0.005) were similar, as were those between MRI activity scores and IL-6 levels (0.240, p = 0.0001) and MRI activity scores and CRP levels (0.282, p<0.0001).

anifestations At baseline, Spearman correlation coefficients between BASDAI scores and IL-6 levels (0.134, p = 0.0327) and BASDAI scores and CRP levels (0.167, p = 0.005) were similar, as were those between MRI activity scores and IL-6 levels (0.240, p = 0.0001) and MRI activity scores and CRP levels (0.282, p<0.0001). We also evaluated baseline MRI activity and BASDAI scores for subgroups of baseline IL-6 and CRP (fig 1). Patients with IL-6 levels greater than the median (7.38 pg/ml, hereafter referred to as patients with elevated IL-6) had significantly greater MRI activity scores than those with IL-6 levels at or below the median (p<0.001). Overall, 75.0% of patients with elevated IL-6 had a baseline MRI activity score >1 compared with 56.3% of patients with low IL-6 (p = 0.002). Similarly, patients with CRP levels >3 times the upper limit of normal (normal range 0–0.5 mg/dl) had significantly greater MRI activity scores compared with those with CRP levels ⩽3 times the upper limit of normal (p<0.001). The trends between baseline IL-6 or CRP levels and BASDAI scores were similar to those observed for MRI activity scores; however, the magnitude of the differences between the medians of the groups was not as great for BASDAI as it was for the MRI activity scores (p = 0.0648 for IL-6 and p = 0.006 for CRP) (fig 1).

01). The trends between baseline IL-6 or CRP levels and BASDAI scores were similar to those observed for MRI activity scores; however, the magnitude of the differences between the medians of the groups was not as great for BASDAI as it was for the MRI activity scores (p = 0.0648 for IL-6 and p = 0.006 for CRP) (fig 1). Figure 1 Baseline magnetic resonance imaging (MRI) activity and BASDAI (Bath Ankylosing Spondylitis Disease Activity Index) scores for quantiles of baseline interleukin (IL)-6 and C-reactive protein (CRP). Horizontal solid lines are medians, and horizontal dotted lines are means. The boxes show the interquartile ranges. The error bars show the standard deviations. P-values show the difference between quantiles of each biomarker using an analysis of variance on the van der Waerden normal scores. Patients with elevated IL-6 levels at baseline were also more likely to have peripheral joint involvement than those with low IL-6 levels (50.0% vs. 28.1%, respectively, p<0.001) and had a greater mean (SD) number of peripheral swollen joints (2.02 (3.44) joints versus 1.20 (3.30) joints, p<0.001). Moreover, 78.9% of patients with elevated IL-6 levels had CRP levels >3 times the upper limit of normal compared with 17.2% of patients with low IL-6 levels (p<0.001). The baseline mean (SD) CRP level for patients with elevated IL-6 levels was 3.79 (3.27) mg/dl compared with 1.04 (0.81) mg/dl for patients with low IL-6 levels (p<0.001). There were no statistically significant differences in age, disease duration, baseline inflammation (visual analogue scale), or the presence of uveitis between patients with low and high IL-6 levels (data not shown).

vels was 3.79 (3.27) mg/dl compared with 1.04 (0.81) mg/dl for patients with low IL-6 levels (p<0.001). There were no statistically significant differences in age, disease duration, baseline inflammation (visual analogue scale), or the presence of uveitis between patients with low and high IL-6 levels (data not shown). Discriminant capacity and sensitivity to change: reduction in biomarker levels after treatment with infliximab or placebo Percentage changes in biomarker levels from baseline to week 2 and from baseline to week 24 are shown for infliximab- and placebo-treated patients in fig 2. Significantly greater reductions in IL-6, VEGF and CRP levels were observed at both week 2 and week 24 for patients in the infliximab group compared with the placebo group (p<0.001). There was no difference between the treatment groups in the change in IFN-γ levels.

or infliximab- and placebo-treated patients in fig 2. Significantly greater reductions in IL-6, VEGF and CRP levels were observed at both week 2 and week 24 for patients in the infliximab group compared with the placebo group (p<0.001). There was no difference between the treatment groups in the change in IFN-γ levels. Figure 2 Median percentage change from baseline in (A) interferon-γ, (B) interleukin-6, (C) vascular endothelial growth factor and (D) C-reactive protein. Predictive validity: correlations between biomarker levels and changes in disease activity and spinal inflammation Univariate correlations between biomarker levels and changes in MRI activity and BASDAI scores are summarised in table 2. In the infliximab group, baseline IL-6 and CRP levels were significantly correlated inversely with changes from baseline to week 24 in MRI activity scores (p⩽0.001) and BASDAI scores (p⩽0.001), with high baseline IL-6 and CRP levels associated with greater reductions in the two indices. In addition, percentage changes from baseline to week 2 and 24 in IL-6, VEGF and CRP levels were significantly correlated with changes in MRI activity scores (p<0.05) and BASDAI scores (p⩽0.001) from baseline to week 24. Thus, decreases from baseline to week 2 and 24 in IL-6, CRP and VEGF levels were significantly correlated with improvement in MRI activity and BASDAI scores. Correlations with changes in IFN-γ levels were not evaluated because of the small number of patients who had detectable IFN-γ levels at baseline.

line to week 24. Thus, decreases from baseline to week 2 and 24 in IL-6, CRP and VEGF levels were significantly correlated with improvement in MRI activity and BASDAI scores. Correlations with changes in IFN-γ levels were not evaluated because of the small number of patients who had detectable IFN-γ levels at baseline. Table 2 Univariate correlations between biomarker levels (baseline and percentage change from baseline to week 2 and week 24) and change from baseline to week 24 in disease activity (BASDAI) and the inflammation detected by MRI (MRI activity score) in patients with ankylosing spondylitis who received infliximab 5 mg/kg (n = 201) or placebo (n = 78) Biomarker Change in MRI activity score Change in BASDAI Placebo Infliximab Placebo Infliximab Baseline IL-6 −0.158 −0.205*** 0.062 −0.258*** VEGF 0.067 −0.074 0.140 −0.079 CRP −0.047 −0.291*** 0.169 −0.322*** Percentage change from baseline to week 2 IL-6 0.080 0.260*** 0.066 0.297*** VEGF 0.110 0.243*** 0.105 0.260*** CRP 0.006 0.296*** −0.017 0.347*** Percentage change from baseline to week 24 IL-6 −0.086 0.215** 0.178 0.340*** VEGF −0.101 0.170* 0.148 0.330*** CRP −0.090 0.243*** 0.147 0.414*** *p<0.05. **p⩽0.01. ***p⩽0.001. BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; CRP, C-reactive protein; IL-6, interleukin-6; MRI, magnetic resonance imaging, VEGF, vascular endothelial growth factor.

Table 2 Univariate correlations between biomarker levels (baseline and percentage change from baseline to week 2 and week 24) and change from baseline to week 24 in disease activity (BASDAI) and the inflammation detected by MRI (MRI activity score) in patients with ankylosing spondylitis who received infliximab 5 mg/kg (n = 201) or placebo (n = 78) Biomarker Change in MRI activity score Change in BASDAI Placebo Infliximab Placebo Infliximab Baseline IL-6 −0.158 −0.205*** 0.062 −0.258*** VEGF 0.067 −0.074 0.140 −0.079 CRP −0.047 −0.291*** 0.169 −0.322*** Percentage change from baseline to week 2 IL-6 0.080 0.260*** 0.066 0.297*** VEGF 0.110 0.243*** 0.105 0.260*** CRP 0.006 0.296*** −0.017 0.347*** Percentage change from baseline to week 24 IL-6 −0.086 0.215** 0.178 0.340*** VEGF −0.101 0.170* 0.148 0.330*** CRP −0.090 0.243*** 0.147 0.414*** *p<0.05. **p⩽0.01. ***p⩽0.001. BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; CRP, C-reactive protein; IL-6, interleukin-6; MRI, magnetic resonance imaging, VEGF, vascular endothelial growth factor. Predictive validity: baseline biomarker levels and ASAS 20 and BASDAI 50 response Figure 3 shows the treatment response of patient groups categorised according to their baseline biomarker levels. Among patients who received infliximab, a greater percentage of those with elevated IL-6 at baseline (>7.38 pg/ml) demonstrated a clinical response as measured by ASAS 20 (A, p<0.0001) and BASDAI 50 (C, p = 0.002) compared with those with low IL-6. Similarly, a greater percentage of infliximab-treated patients with CRP >3 times the upper limit of normal (1.5 mg/dl) achieved an ASAS 20 (B, p<0.001) and BASDAI 50 (D, p = 0.002) responses compared with the percentage with CRP ⩽3 times the upper limit of normal.

AI 50 (C, p = 0.002) compared with those with low IL-6. Similarly, a greater percentage of infliximab-treated patients with CRP >3 times the upper limit of normal (1.5 mg/dl) achieved an ASAS 20 (B, p<0.001) and BASDAI 50 (D, p = 0.002) responses compared with the percentage with CRP ⩽3 times the upper limit of normal. Figure 3 Comparison of the percentage of ASAS 20 (20% improvement in the ASsessment in Ankylosing Spondylitis working group criteria) (A,B) and BASDAI (Bath Ankylosing Spondylitis Disease Activity Index) 50 (C,D) responders between categories of baseline interleukin (IL)-6 and C-reactive protein (CRP) divided at the medians. Predictive validity: multiple regression associations between biomarker levels and changes in disease activity and spinal inflammation Multiple linear regression analyses were conducted to evaluate whether one individual marker (IL-6, VEGF and CRP) was more strongly associated with change in MRI activity or BASDAI scores from baseline to week 24 (table 3). Early reductions in IL-6 from baseline to week 2 were significantly associated with improvement from baseline to week 24 in MRI activity and BASDAI. Reductions from baseline to week 24 in CRP were significantly associated with reductions in MRI activity scores at week 24; however, both baseline levels and the decrease in CRP from baseline to week 24 were associated with reductions in BASDAI. Further, reductions in VEGF levels at week 24 were also significantly associated with change from baseline to week 24 in BASDAI. Thus, in the infliximab group, early decreases in IL-6 and later decreases in CRP were associated with improvement in MRI activity scores. Further, early decreases in IL-6 and baseline CRP levels were associated with improvement in BASDAI scores.

significantly associated with change from baseline to week 24 in BASDAI. Thus, in the infliximab group, early decreases in IL-6 and later decreases in CRP were associated with improvement in MRI activity scores. Further, early decreases in IL-6 and baseline CRP levels were associated with improvement in BASDAI scores. Table 3 Linear regression modelling of the associations between biomarker levels (baseline and percentage change from baseline to week 2 or week 24) and change from baseline to week 24 in disease activity (BASDAI) and the spinal inflammation detected by MRI (MRI activity score) Model r2 b p Value Infliximab group MRI activity score: change from baseline to week 24 Baseline biomarker levels no parameters significantly associated − − − Percentage change from baseline to week 2 in biomarker levels 0.011 IL-6 0.053 <0.001 Percentage change from baseline to week 24 in biomarker levels 0.042 CRP 0.009 0.009 BASDAI: change from baseline to week 24 Baseline biomarker levels 0.043 CRP −0.196 0.005 Percentage change from baseline to week 2 in biomarker levels 0.099 IL-6 0.021 <0.001 Percentage change from baseline to week 24 in biomarker levels 0.113 VEGF 0.005 0.029 CRP 0.005 0.005 Placebo group No biomarker levels at baseline or changes from baseline to week 2 or 24 were significantly associated with changes in MRI activity score or BASDAI at week 24 BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; CRP, C-reactive protein; IL-6, interleukin-6; MRI, magnetic resonance imaging; VEGF, vascular endothelial growth factor.

at baseline or changes from baseline to week 2 or 24 were significantly associated with changes in MRI activity score or BASDAI at week 24 BASDAI, Bath Ankylosing Spondylitis Disease Activity Index; CRP, C-reactive protein; IL-6, interleukin-6; MRI, magnetic resonance imaging; VEGF, vascular endothelial growth factor. DISCUSSION The results of this study show that the established efficacy of infliximab for improving the clinical signs and symptoms of AS and reducing spinal inflammation demonstrated by MRI19–21 is also associated with significant changes in inflammatory biomarkers. Understanding the utility of specific biomarkers associated with anti-TNF-α therapy may be particularly important in AS because the rate of structural changes is rather slow over time,28 29 and early identification of patients who will benefit from therapy is essential. The available information on the modulation of inflammatory markers in AS after treatment with anti-TNF agents and the association with changes in imaging measures such as MRI has been limited. Data from an early randomised controlled trial30 31 and a more recent open-label study in a small number of patients32 suggest that baseline CRP levels may be useful as markers of clinical response after treatment with infliximab. The results of the current study represent, to our knowledge, the largest data set of patients with AS evaluated for inflammatory markers, clinical signs and symptoms, and the inflammation detected by MRI.

32 suggest that baseline CRP levels may be useful as markers of clinical response after treatment with infliximab. The results of the current study represent, to our knowledge, the largest data set of patients with AS evaluated for inflammatory markers, clinical signs and symptoms, and the inflammation detected by MRI. The face validity of inflammatory marker levels was already evident at baseline. Patients with high baseline IL-6 (>7.38 pg/ml) and CRP (>1.5 mg/dl) had more spinal inflammation detected by MRI compared with those with low biomarker levels. Patients with high baseline CRP levels also had greater disease activity as measured by BASDAI, and there was a trend towards greater disease activity for patients with high baseline IL-6 levels, although the difference was not statistically significant. Although there was variability in MRI activity and BASDAI score within each group, the magnitude of the difference between the groups was greater for MRI activity compared with BASDAI, suggesting that there may be a stronger relationship between biomarker levels and spinal inflammation relative to that between biomarker levels and disease activity.

in MRI activity and BASDAI score within each group, the magnitude of the difference between the groups was greater for MRI activity compared with BASDAI, suggesting that there may be a stronger relationship between biomarker levels and spinal inflammation relative to that between biomarker levels and disease activity. Following treatment with infliximab, patients showed significantly greater reductions in IL-6, VEGF and CRP compared with patients who received placebo, demonstrating the discriminant capacity and sensitivity to change of the biomarkers. These reductions were evident as early as week 2, and the levels were similarly low at week 24. In line with this result, treatment with infliximab was shown to modulate markers of angiogenesis, including VEGF, in synovial tissue early after initiation of therapy in patients with psoriatic arthritis.33 In the infliximab group, baseline IL-6 and CRP levels and reductions from baseline in IL-6, VEGF and CRP were significantly correlated with improvement in disease activity and inflammation at week 24. No significant correlations were observed in the placebo group. The significance of these correlations is noteworthy, because a substantial number of AS patients (about 50%) had undetectable IL-6 levels at baseline. This finding is consistent with the results of an earlier pilot study34 and merits further investigation because there is limited knowledge on the natural course of IL-6 levels in AS.

e of these correlations is noteworthy, because a substantial number of AS patients (about 50%) had undetectable IL-6 levels at baseline. This finding is consistent with the results of an earlier pilot study34 and merits further investigation because there is limited knowledge on the natural course of IL-6 levels in AS. The predictive validity of biomarker levels became particularly evident when the clinical response was evaluated. Among patients who received infliximab, an ASAS 20 and a BASDAI 50 response at week 24 was achieved by a greater percentage of those with elevated IL-6 or CRP levels at baseline compared with those with low IL-6 or CRP levels. The multiple regression analysis enabled us to determine which biomarker was most associated with spinal inflammation and disease activity. The results showed that early changes in IL-6 and late changes in CRP were associated with improvements in both the MRI activity and the BASDAI score. Baseline CRP levels and reductions in VEGF levels at week 24 were also associated with improvements in BASDAI score, but were not associated with changes in MRI activity score. The finding that early improvement in IL-6 was associated with improvement in both spinal inflammation and disease activity suggests that IL-6 may play an earlier and more central role in the inflammatory cascade than CRP. In this regard, it is noteworthy that IL-6 is thought to be involved in the initiation of the acute phase response and the production of CRP in the liver.35 IL-6 polymorphisms have been shown to be associated with CRP levels in other conditions.8 9

an earlier and more central role in the inflammatory cascade than CRP. In this regard, it is noteworthy that IL-6 is thought to be involved in the initiation of the acute phase response and the production of CRP in the liver.35 IL-6 polymorphisms have been shown to be associated with CRP levels in other conditions.8 9 In the current study, baseline levels of IL-6 and CRP were generally comparable between both treatment groups. Levels of serum IL-6 and CRP in previous published studies of patients with AS have been conflicting; one reporting lower levels than those in the current study,36 37 and others reporting higher levels.38–40 The reported differences in IL-6 and CRP levels are most likely related to the heterogeneity of the AS patient populations studied. In the current study, only patients with active disease who fulfilled the criteria for anti-TNF therapy were included. Moreover, IL-6 levels were strongly correlated with CRP levels at baseline in the current study, which is consistent with the results of other studies of AS and rheumatoid arthritis.38 41 Moderate correlations were observed between baseline levels of IL-6 and VEGF and between baseline levels of VEGF and CRP. Treatment with infliximab did not alter these relationships between the biomarkers. Baseline VEGF levels in the current study were similar to those previously reported for patients with early rheumatoid arthritis42 but were higher than those of other patients with spondyloarthropathies.4 In patients with rheumatoid arthritis, IL-6 has been shown to be a key cytokine in the production of VEGF,43 and VEGF may, in turn, stimulate the production of IL-6 and TNF-α in a positive feedback loop.44 Our results suggest that similar relationships between IL-6 and VEGF may play a part in the inflammation exhibited by some patients with AS. However, as already stated, a significant number of AS patients had undetectable levels of circulating IL-6.

mulate the production of IL-6 and TNF-α in a positive feedback loop.44 Our results suggest that similar relationships between IL-6 and VEGF may play a part in the inflammation exhibited by some patients with AS. However, as already stated, a significant number of AS patients had undetectable levels of circulating IL-6. Some potential limitations of this study are noteworthy. The ASSERT study was primarily designed to evaluate the safety and efficacy of infliximab in patients with AS and was not powered to evaluate correlations between changes in inflammatory markers and clinical measures. In addition, the study population was somewhat heterogeneous, which included a range of patients with moderate to severe AS. Finally, the IL-6 assay had limited sensitivity to detect low levels of this cytokine, and thus, nearly half of the patients had undetectable IL-6 levels at baseline. This limitation may have resulted in an underestimation in the strength of the correlations between changes in IL-6 levels and changes in measures of inflammation and disease activity in the infliximab group. However, despite these potential limitations, the results of this study contribute to our understanding of the importance of IL-6 in the pathogenesis of AS. The results suggest that the early reduction of IL-6 may be a critical event in the treatment response and the pathogenesis of AS. Although biomarker levels were associated with improvement in disease activity and the spinal inflammation detected by MRI, their relative contribution to the explanation of the observed treatment response needs to be examined further, especially as there is no strong correlation between BASDAI and MRI (Braun, unpublished). However, the results of this study do suggest that an early reduction of IL-6 level is important. A reduction in CRP level may have utility as a surrogate marker of response in clinical practice, especially given the strong correlation between IL-6 and CRP levels.

orrelation between BASDAI and MRI (Braun, unpublished). However, the results of this study do suggest that an early reduction of IL-6 level is important. A reduction in CRP level may have utility as a surrogate marker of response in clinical practice, especially given the strong correlation between IL-6 and CRP levels. MRI has been shown to be a useful tool for evaluating the reduction in spinal inflammation after anti-TNF therapy, but it is unclear how early the benefit can be detected. In any case, the early identification of patients who are most likely to exhibit improvement through the use of biomarkers will assist clinicians in their treatment decisions. The authors thank the patients, investigators and study personnel who made the ASSERT study possible; Eva Silvestro and Elizabeth Lee-Rykaczewski of Centocor Research and Development, Inc. for their laboratory expertise; Stephen Xu for his statistical assistance and Scott Newcomer, of Centocor, Inc. for his expertise with preparing the manuscript. Funding: This study was funded by Centocor Research and Development, Inc. Competing interests: DvdH and JB have received research funding and/or consulting fees from Centocor and/or Schering-Plough Research Institute. SV, CW, JCM, DB, JH and TG are employees of Centocor Research and Development, Inc. and own Johnson & Johnson stock.

Rheumatoid arthritis (RA) is a chronic autoimmune disease that usually leads to synovial joint damage and consequent disability. Activated T cells play a crucial part in the immunopathology of RA.1 T cell proliferation and interaction with other cell types, including synovial macrophages, fibroblasts and B cells, result in the production of proinflammatory cytokines and matrix metalloproteinases that promote synovitis and erosive loss of subchondral bone. Thus, targeting T cell activation is a rational therapeutic approach for the treatment of RA. T cell activation requires antigen recognition by the T cell receptor, referred to as signal 1, as well as a co-stimulatory signal, referred to as signal 2.2 One of the best characterised co-stimulatory pathways involves the interaction of CD28 expressed on T cells with CD80 and CD86 expressed on antigen-presenting cells. Endogenous down-regulation of CD28-mediated T cell co-stimulation occurs through T cell expression of cytotoxic T lymphocyte antigen 4 (CTLA-4). CTLA-4 binds to CD80 and CD86 with higher avidity than CD28, thus preventing co-stimulation through this pathway, and is a major down-regulatory signal for T cells.

ing cells. Endogenous down-regulation of CD28-mediated T cell co-stimulation occurs through T cell expression of cytotoxic T lymphocyte antigen 4 (CTLA-4). CTLA-4 binds to CD80 and CD86 with higher avidity than CD28, thus preventing co-stimulation through this pathway, and is a major down-regulatory signal for T cells. Abatacept is a soluble, recombinant, fully human fusion protein composed of the extracellular domain of human CTLA-4 and the Fc domain of human IgG1, which has been modified to prevent complement fixation. Abatacept is the first in a class of biological agents that target the second signal required for full T cell activation, a mechanism of action that is fundamentally different than any other current biological RA therapy. The efficacy of abatacept monotherapy has been shown in a phase IIa trial in patients with RA with an inadequate response to disease-modifying anti-rheumatic drugs (DMARDs).3 The efficacy of abatacept added to methotrexate (MTX) was demonstrated in the phase IIb and phase III trials of patients with an inadequate response to MTX or tumour necrosis factor-targeting agents.4–6

phase IIa trial in patients with RA with an inadequate response to disease-modifying anti-rheumatic drugs (DMARDs).3 The efficacy of abatacept added to methotrexate (MTX) was demonstrated in the phase IIb and phase III trials of patients with an inadequate response to MTX or tumour necrosis factor-targeting agents.4–6 Results from the 1-year phase III, randomised, double-blind trial of Abatacept in Inadequate responders to Methotrexate (AIM) trial have been previously reported.5 Abatacept added to MTX therapy resulted in clinically significant improvements in the signs and symptoms of RA, physical function and health-related quality of life in patients with an inadequate response to MTX. Further, abatacept resulted in a statistically significant slowing of the progression of structural damage after 12 months of treatment. Progressive structural damage is associated with increasing disability over time.7 8 Thus, the effects of abatacept treatment on radiographic outcomes were assessed over longer-term treatment. Here we present radiographic assessments after 2 years of abatacept treatment in an open-label extension of the AIM trial. These data demonstrate that radiographic progression is significantly inhibited over the 2 years. The maintenance of efficacy of abatacept therapy on other clinical end points will be described in full in another report.

aphic assessments after 2 years of abatacept treatment in an open-label extension of the AIM trial. These data demonstrate that radiographic progression is significantly inhibited over the 2 years. The maintenance of efficacy of abatacept therapy on other clinical end points will be described in full in another report. PATIENTS AND METHODS The study design, baseline characteristics and results of the 12-month double-blind phase of this trial have been reported previously.5 Institutional review boards or independent ethics committees approved the clinical protocol, and written informed consent to the study protocol was provided by each patient.

ND METHODS The study design, baseline characteristics and results of the 12-month double-blind phase of this trial have been reported previously.5 Institutional review boards or independent ethics committees approved the clinical protocol, and written informed consent to the study protocol was provided by each patient. Patients Patients eligible to participate in the AIM trial were at least 18 years of age, met the American Rheumatism Association criteria for RA,9 and had RA for at least 12 months. All patients were screened for tuberculosis. At randomisation, all patients had persistent, active RA despite treatment with MTX, with ⩾10 swollen joints, ⩾12 tender joints and a C-reactive protein concentration of ⩾1.0 mg/dl. Patients were required to have received MTX therapy at a minimum dose of 15 mg/week for at least 3 months, with the dose having been stable for a minimum of 28 days prior to randomisation. A 28-day washout period was required for all other DMARDs. Reduction of MTX dosage in the first 6 months was permitted only in cases of toxicity. Stable dosages of nonsteroidal anti-inflammatory drugs or corticosteroids at ⩽10 mg prednisone/day were allowed, with a stable dosage for 25 days prior to randomisation. Adjustment of MTX and/or corticosteroid dosages were permitted after month 6, as was treatment with an additional DMARD if deemed appropriate by the investigator, including hydroxychloroquine, sulfasalazine, gold or azathioprine.

0 mg prednisone/day were allowed, with a stable dosage for 25 days prior to randomisation. Adjustment of MTX and/or corticosteroid dosages were permitted after month 6, as was treatment with an additional DMARD if deemed appropriate by the investigator, including hydroxychloroquine, sulfasalazine, gold or azathioprine. Study protocol During the first year, patients were randomised to receive a fixed dose of abatacept of approximately 10 mg/kg body weight or placebo in a 2:1 ratio. The dose of abatacept was 500 mg for patients weighing less than 60 kg, 750 mg for patients weighing 60–100 kg, and 1 g for patients weighing more than 100 kg. Study medication was infused intravenously over 30 min on days 1, 15 and 29, and every 28 days subsequently. Patients completing the double-blind period were eligible to enter an open-label, long-term extension and receive abatacept therapy every 28 days at a dose of approximately 10 mg/kg, as described above. The loading dose used in the double-blind period was not utilised here. No premedication was required for the intravenous infusions.

ting the double-blind period were eligible to enter an open-label, long-term extension and receive abatacept therapy every 28 days at a dose of approximately 10 mg/kg, as described above. The loading dose used in the double-blind period was not utilised here. No premedication was required for the intravenous infusions. Radiographic evaluation Radiographic assessments of hands, wrists and feet were performed at baseline, year 1 and year 2, or upon early termination. The trial was powered for radiographic findings at year 1, and the primary end point was the change in erosion score, with secondary assessments of joint space narrowing (JSN) and total score using the Genant-modified Sharp scoring system.5 For each patient, baseline, year 1 and year 2 radiographs were all re-read at year 2 by two independent expert readers blinded to the original treatment allocation and the sequence of films.10 The Genant-modified Sharp scoring system (fig 1) assesses changes in structural damage, assigning scores for erosions of 0–3.5 (eight gradations) for 14 sites in each hand and wrist, and six sites in each foot.11 12 JSN scores of 0–4 (nine gradations) are assigned to 13 sites in each hand and six sites in each foot.12 The erosion and JSN scores are normalised to 145 for a maximum total Genant-modified Sharp score of 290.12 Inter-reader, intra-class correlation coefficient (ICC) was determined based on the baseline and year 1 readings.10 Smallest detectable difference was determined as 1.96 SD of the two independent baseline readings.13

erosion and JSN scores are normalised to 145 for a maximum total Genant-modified Sharp score of 290.12 Inter-reader, intra-class correlation coefficient (ICC) was determined based on the baseline and year 1 readings.10 Smallest detectable difference was determined as 1.96 SD of the two independent baseline readings.13 Figure 1 Genant-modified Sharp scoring system. Coloured areas indicate sites used to evaluate erosion and joint space narrowing (JSN) scores. Each site was evaluated in 0.5-unit increments as indicated. Erosion and JSN scores are normalised to 145 for a maximum total score of 290. Statistical analysis The study was powered for change from baseline in the Genant-modified Sharp erosion score and used a rank-based analysis of covariance to compare changes between treatment groups at year 1. The radiographic analyses were performed on a modified intent-to-treat population at year 2. The modified intent-to-treat population included patients who were randomised and treated, completed the 1-year double-blind period, and continued to receive at least one dose of study medication during the open-label period. Patients who discontinued abatacept treatment during the open-label period were requested to return for follow-up radiography at year 2, regardless of any subsequent new anti-rheumatic therapy prescribed by the investigator. Radiographs were taken at the time of discontinuation in patients who were unwilling or unable to return at year 2. In these patients, 2-year data were imputed using linear extrapolation of the scores of the discontinuation film.

2, regardless of any subsequent new anti-rheumatic therapy prescribed by the investigator. Radiographs were taken at the time of discontinuation in patients who were unwilling or unable to return at year 2. In these patients, 2-year data were imputed using linear extrapolation of the scores of the discontinuation film. The changes from baseline in total Genant-modified Sharp scores, erosion scores and JSN scores were charted as cumulative probability plots to visually represent the distribution of all radiographic data. The observed cumulative proportion (scores ranked from the lowest to the highest and presented as a cumulative proportion of all scores) was plotted against the actual change from baseline of each score. The proportion of observations that fall below each possible change from baseline (y-axis) can be read on the x-axis.14

observed cumulative proportion (scores ranked from the lowest to the highest and presented as a cumulative proportion of all scores) was plotted against the actual change from baseline of each score. The proportion of observations that fall below each possible change from baseline (y-axis) can be read on the x-axis.14 RESULTS Patients and study completion Of the 652 patients randomised and treated in the double-blind portion of the AIM trial, 433 received abatacept and 219 received placebo (table 1). As previously reported, baseline demographics and clinical characteristics did not differ significantly between patients in the abatacept and placebo groups.5 After the 12-month double-blind treatment period, all patients were eligible to receive open-label treatment with abatacept. More patients receiving abatacept treatment (89%) completed the 1-year double-blind phase relative to the placebo group (74%). Of the 547 patients who completed 12 months of treatment, 539 (83% of all randomised and treated patients) were treated with abatacept in the open-label period (378 initially randomised to abatacept (87%) and 161 to placebo (73.5%)). A high retention rate was maintained with open-label abatacept treatment, with 90% of the patients who entered the long-term extension completing 2 years.

f all randomised and treated patients) were treated with abatacept in the open-label period (378 initially randomised to abatacept (87%) and 161 to placebo (73.5%)). A high retention rate was maintained with open-label abatacept treatment, with 90% of the patients who entered the long-term extension completing 2 years. Table 1 Number of subjects with radiographic assessments on days 365 and 729 Abatacept n = 376 (%) Placebo n = 160 (%) Subjects included in analyses Day 365 328 (87.2) 144 (90.0) Baseline and day 365 328 (87.2) 144 (90.0) Day 729 324 (86.2) 143 (89.4) Baseline and day 729 315 (83.8) 139 (86.9) Imputation on day 729* 9 (2.4) 4 (2.5) Subjects not included in the analyses No day 365 value† 2 (0.5) 1 (0.6) No day 729 value‡ 6 (1.6) 2 (1.3) *Subjects are discontinued and qualified for imputation requirement. †Two subjects in the abatacept group had evaluative radiographic assessment outside the pre-specified visit window of day 365, and one subject in the placebo group had no radiographic assessment on day 365. ‡Subjects had evaluative radiographic assessment outside the pre-specified visit window of day 729.

requirement. †Two subjects in the abatacept group had evaluative radiographic assessment outside the pre-specified visit window of day 365, and one subject in the placebo group had no radiographic assessment on day 365. ‡Subjects had evaluative radiographic assessment outside the pre-specified visit window of day 729. Two-year radiographic data were available from 87% of patients who entered the open-label period, which comprises 72% of all randomised and treated patients (467 of 652 patients; table 1). Observed data were available for 97% of patients who completed the open-label period; 454 of 467 patients had radiographs at baseline and at year 2. Baseline and an early termination film were available for the remaining 3% of patients, and 2-year data were imputed by linear extrapolation. The data include baseline and year 2 radiographs of 324 patients treated with abatacept for 2 years (observed data in 315 patients, imputed data in nine patients). Therefore, the 2-year data represent the majority of randomised patients. As previously reported in an abstract,10 a high degree of inter-observer agreement was demonstrated by ICC for radiographic assessments at baseline (0.90) and 12 months (0.92) and for change between baseline and 12 months (0.82). The smallest detectable differences were 3.5, 2.5 and 5.1, respectively, for erosion, JSN and total score. These correspond to 1.9%, 1.3% and 1.3%, respectively, of the maximum values for each of these scores and compare favourably with values reported for other studies.15 16

ange between baseline and 12 months (0.82). The smallest detectable differences were 3.5, 2.5 and 5.1, respectively, for erosion, JSN and total score. These correspond to 1.9%, 1.3% and 1.3%, respectively, of the maximum values for each of these scores and compare favourably with values reported for other studies.15 16 Radiographic results: population-level data The progression of structural damage was significantly reduced in patients treated with abatacept over 2 years versus those treated with placebo initially. Baseline erosion and JSN scores were comparable in patients originally randomised to either treatment group. Significantly lower mean changes in radiographic progression were observed in patients treated with abatacept for 2 years, with mean changes in total score, erosion and JSN scores of 1.55, 0.84 and 0.71 units, respectively, compared with 3.17, 1.48 and 1.69 units in patients treated with placebo for 12 months prior to abatacept open-label therapy (fig 2, table 2).

ic progression were observed in patients treated with abatacept for 2 years, with mean changes in total score, erosion and JSN scores of 1.55, 0.84 and 0.71 units, respectively, compared with 3.17, 1.48 and 1.69 units in patients treated with placebo for 12 months prior to abatacept open-label therapy (fig 2, table 2). Figure 2 Mean change from baseline in total (A), erosion (B) and joint space narrowing (C) Genant-modified Sharp scores for patients at 12 months and 2 years. *Placebo patients were switched to abatacept after 12 months. All patients received background methotrexate. Table 2 Yearly mean changes in Genant-modified Sharp scores Δ Baseline to year 1 Δ Baseline to year 2 Δ Year 1 to year 2 Total score Abatacept 1.07 1.55 0.46 Placebo/abatacept 2.40 3.17 0.75 Erosion score Abatacept 0.62 0.84 0.21 Placebo/abatacept 1.44 1.69 0.25 Joint space narrowing score Abatacept 0.45 0.71 0.24 Placebo/abatacept 0.95 1.48 0.50 Baseline is day 1 of study. All randomised and treated patients who entered the open-label period and had radiographs at baseline and year 1. Baseline and year 1 radiographs were re-read at day 729. Placebo patients were switched to abatacept in the long-term extension (year 2). All patients received background methotrexate.

ne is day 1 of study. All randomised and treated patients who entered the open-label period and had radiographs at baseline and year 1. Baseline and year 1 radiographs were re-read at day 729. Placebo patients were switched to abatacept in the long-term extension (year 2). All patients received background methotrexate. Patients treated with abatacept for 2 years also had lower median changes in radiographic progression scores. The median change from baseline in total Genant-modified Sharp scores was 0.0 units at year 1 and year 2 with abatacept treatment (fig 3), versus median changes of 0.46 and 0.51 units at year 1 and year 2, respectively, in patients who received placebo prior to abatacept therapy. Further, the median change in erosion score was 0.0 units in patients treated with abatacept for 2 years, compared with 0.26 units in patients originally in the placebo group.

), versus median changes of 0.46 and 0.51 units at year 1 and year 2, respectively, in patients who received placebo prior to abatacept therapy. Further, the median change in erosion score was 0.0 units in patients treated with abatacept for 2 years, compared with 0.26 units in patients originally in the placebo group. Figure 3 Cumulative probability distribution of changes from baseline in Genant-modified total Sharp scores by treatment at year 1 (A) and year 2 (B). The solid line represents patients treated with abatacept (A,B) and the dotted line represents patients treated with placebo (A) or treated with placebo for 12 months then abatacept for 12 months (B). All patients received background methotrexate. The mean change in the total score, erosion and JSN score were lower from year 1 to year 2 (0.46, 0.21 and 0.24 units, respectively) than the mean change in these scores from baseline to year 1 (1.07, 0.62 and 0.45 units, respectively) in patients treated with abatacept for up to 2 years (table 2). The increasing effect of treatment was consistently observed for both erosions and JSN.

r from year 1 to year 2 (0.46, 0.21 and 0.24 units, respectively) than the mean change in these scores from baseline to year 1 (1.07, 0.62 and 0.45 units, respectively) in patients treated with abatacept for up to 2 years (table 2). The increasing effect of treatment was consistently observed for both erosions and JSN. Radiographic results: patient-level data Radiographic results at 2 years demonstrate the maintenance of effect of abatacept and MTX on non-progression of structural damage in patients with no progression at year 1 (table 3), defined as a change from baseline in total Sharp score of ⩽0. Overall, 50% of patients treated with abatacept did not progress from baseline to 2 years of treatment (163 of 324 patients). At 2 years, non-progression in erosions and JSN (change ⩽0) was observed from baseline in 56% and 68% of these patients, respectively. In patients treated with abatacept with no radiographic progression at year 1, 79% had no radiographic progression at year 2. Non-progression in erosions and JSN was maintained at 2 years in 83% and 87% of patients with no progression at 12 months.

observed from baseline in 56% and 68% of these patients, respectively. In patients treated with abatacept with no radiographic progression at year 1, 79% had no radiographic progression at year 2. Non-progression in erosions and JSN was maintained at 2 years in 83% and 87% of patients with no progression at 12 months. Table 3 Patients with no progression in structural damage measured by Genant-modified Sharp scores in patients on abatacept therapy for 2 years Percentage of patients with no progression in structural damage* Percentage of patients with no progression in structural damage in year 2† Outcome Baseline to year 1 (n = 328) Baseline to year 2 (n = 324) Non-progressors at end of year 1 remaining non-progressors† Progressors at end of year 1 becoming non-progressors Erosion (%) 61 56 83 49 Joint space narrowing (%) 74 68 87 52 Total (%) 56 50 79 45 All patients received background methotrexate. *Defined by a change in the total score of ⩽0 from baseline to year 1. †Defined by a change in the total score of ⩽0 from year 1 to year 2. Non-progression at year 2 was also achieved in patients with initial damage progression on abatacept therapy. Of patients treated with abatacept who demonstrated radiographic progression at year 1, 45% had no progression of structural damage at year 2 (64 of 142 patients). Further, 53% of patients originally randomised to placebo with progression at year 1 did not progress at year 2 after receiving abatacept treatment (42 of 79 patients).

treated with abatacept who demonstrated radiographic progression at year 1, 45% had no progression of structural damage at year 2 (64 of 142 patients). Further, 53% of patients originally randomised to placebo with progression at year 1 did not progress at year 2 after receiving abatacept treatment (42 of 79 patients). Cumulative probability plots showing the distribution of change from baseline in total Genant-modified Sharp scores for year 1 and year 2 are shown in fig 3. Comparison of the curves in patients treated in the abatacept group (solid line) versus patients originally randomised to placebo (dotted line) demonstrates that abatacept treatment is associated with results in decreased numbers of patients with progression of structural damage. Furthermore, the amount of structural damage in patients who progressed is lower in patients treated with abatacept for both year 1 and year 2.

lly randomised to placebo (dotted line) demonstrates that abatacept treatment is associated with results in decreased numbers of patients with progression of structural damage. Furthermore, the amount of structural damage in patients who progressed is lower in patients treated with abatacept for both year 1 and year 2. DISCUSSION Abatacept has been shown to significantly reduce disease activity in patients with RA. The previously described randomised, double-blind, placebo-controlled portion of the phase III study of abatacept in patients with RA with an inadequate response to MTX demonstrated statistically significant improvement in the signs and symptoms of disease, physical function assessed by the health-assessment questionnaire disability index, and health-related quality of life.5 In addition, abatacept positively impacted the rate of structural damage progression at 12 months, reducing it by approximately 50% compared with placebo. This placebo-controlled radiographic finding from randomised clinical trials provides an assessment of short-term effects on structural damage. However, only long-term radiographic progression has been associated with physical disability.8 Therefore, it is important to assess longer-term effects on structural damage beyond 1 year of therapy.

radiographic finding from randomised clinical trials provides an assessment of short-term effects on structural damage. However, only long-term radiographic progression has been associated with physical disability.8 Therefore, it is important to assess longer-term effects on structural damage beyond 1 year of therapy. The retention rate of patients receiving abatacept treatment was high, with 89% completing the 1-year double-blind portion of the AIM trial,5 and 90% of the patients entering the long-term extension completing year 2, suggesting the tolerability and durability of response to abatacept. Further, radiographic data were collected from a high percentage of patients, with 2-year data available from 87% of patients entering the open-label period, and observed data for 97% of patients; linear imputation based on baseline and early termination films was used in only 3% of patients.

of response to abatacept. Further, radiographic data were collected from a high percentage of patients, with 2-year data available from 87% of patients entering the open-label period, and observed data for 97% of patients; linear imputation based on baseline and early termination films was used in only 3% of patients. Of note, an increased proportion of patients randomised to the placebo group withdrew from the double-blind portion of the study (between baseline and year 1) due to lack of efficacy (18% vs 3% in the abatacept group).5 Patients initially randomised to placebo who discontinued the study during the double-blind period are not included in these calculations. As these patients with worsening disease were not included in the assessment of radiographic outcomes, the progression of structural damage in patients initially randomised to placebo may be underestimated. As is standard for the assessment of radiographic outcome, readers were blinded to the sequence of radiographs, thus eliminating bias for the expectation of benefit. Abatacept therapy resulted in inhibition of structural damage progression over time. After 2 years of treatment, a significant reduction in the progression of structural damage was observed in patients treated with abatacept for 2 years relative to placebo for 12 months plus abatacept for 12 months.8 Lower mean changes in total Genant-modified Sharp scores at both 12 months and year 2 were 1.07 vs 2.4 units, and 1.55 vs 3.17 units, respectively. Notably, abatacept had a clear benefit on both the erosion and JSN scores.

eated with abatacept for 2 years relative to placebo for 12 months plus abatacept for 12 months.8 Lower mean changes in total Genant-modified Sharp scores at both 12 months and year 2 were 1.07 vs 2.4 units, and 1.55 vs 3.17 units, respectively. Notably, abatacept had a clear benefit on both the erosion and JSN scores. The treatment effect for abatacept relative to placebo was an approximately 50% reduction in mean progression of structural damage in the first year of the study. The rate of progression of structural damage when patients had received placebo treatment for 2 years can be estimated by linear extrapolation from the 12-month data of patients randomised to placebo. Patients in the placebo group progressed >2 units in year 1 and can be projected to progress 4–5 units over 2 years using linear extrapolation. Overall, patients receiving abatacept treatment for 2 years progressed approximately 1.5 units versus the expected progression of approximately 4.5 units if placebo treatment had been continued for the second year. Thus, the effect of abatacept treatment in the second year of treatment would hypothetically reduce radiographic progression by an estimated two-thirds when compared with patients with established disease receiving MTX alone.

ession of approximately 4.5 units if placebo treatment had been continued for the second year. Thus, the effect of abatacept treatment in the second year of treatment would hypothetically reduce radiographic progression by an estimated two-thirds when compared with patients with established disease receiving MTX alone. Fewer abatacept-treated patients had progression of structural damage, and, overall, 50% of patients receiving abatacept did not have radiographic progression over 2 years. In addition, 79% of patients treated with abatacept with no progression at year 1 maintained no progression of structural damage at year 2, demonstrating a durable effect on radiographic outcomes with abatacept therapy. In addition, 45% of patients with some progression at year 1 had no progression in year 2. As the overwhelming number of patients entering the second year of treatment completed the year, we are not reporting on a selected group of responders. The increasing effect of abatacept treatment on the progression of structural damage is consistent with other efficacy data; the proportion of patients with American College of Rheumatology (ACR) 50 and 70 responses were statistically significantly increased from 6 months to 12 months in this study.5 The ACR 20, 50 and 70 responses; improvement in physical function; and improvement in both the physical and mental components of the Short-Form-36 health-related quality of life scale in patients treated with abatacept for up to 3 years will be described in full in a future report.

6 months to 12 months in this study.5 The ACR 20, 50 and 70 responses; improvement in physical function; and improvement in both the physical and mental components of the Short-Form-36 health-related quality of life scale in patients treated with abatacept for up to 3 years will be described in full in a future report. It is difficult to compare results across clinical trials with other treatments that have described radiographic outcomes. Confounding factors include the treatment and disease history of each study population, differing rates of radiographic progression for the control group in each trial, and the use of different scoring systems for the measurement of radiographic damage. The current study assessed the efficacy of abatacept in patients with established RA and long duration of disease, whereas several trials of other biological therapies, such as those targeting tumour necrosis factor, involved patients with early RA (eg, the ERA trial of etanercept17 and the PREMIER trial of adalimumab18) or MTX-naive patients (the TEMPO trial of etanercept19). The Sharp-modified Sharp and the van der Heijde-modified Sharp scoring systems utilised in the trials of tumour necrosis factor-targeting agents,16 17 20 as well as the Genant-modified Sharp scoring system utilised here and in the interleukin-1 receptor antagonist and rituximab studies,18 21 all have differences in joints assessed, scales applied and maximum achievable scores. Thus, Sharp-, van der Heijde- and Genant-modified Sharp scores cannot be directly inter-converted, although they correlate moderately well cross-sectionally and longitudinally. In addition, the definition of the progression of structural damage varied among studies, including progression defined as change ⩾0 or the smallest detectable difference among readers.16–18 20 21 The Genant-modified scoring system employed in this study showed high inter-reader agreement (0.90–0.92 cross-sectionally and 0.82 for change) and high sensitivity to change (smallest detectable difference) for erosion (3.5), JSN (2.5) and total score (5.1).10

he smallest detectable difference among readers.16–18 20 21 The Genant-modified scoring system employed in this study showed high inter-reader agreement (0.90–0.92 cross-sectionally and 0.82 for change) and high sensitivity to change (smallest detectable difference) for erosion (3.5), JSN (2.5) and total score (5.1).10 The inhibition of progression of structural damage with abatacept has clinical implications for the prevention of disability over time. The association of progression of structural damage and impaired physical function over the long term has been well correlated, and is strongest in established disease.7 8 Impacting the rate of joint destruction with continued abatacept treatment is likely to impede progressive functional disability, and thus may have increasing benefit for patients over time. Further study in this and other patient populations will assess the efficacy of abatacept therapy over longer periods of time. Funding: This study was funded by Bristol-Myers Squibb.

The inhibition of progression of structural damage with abatacept has clinical implications for the prevention of disability over time. The association of progression of structural damage and impaired physical function over the long term has been well correlated, and is strongest in established disease.7 8 Impacting the rate of joint destruction with continued abatacept treatment is likely to impede progressive functional disability, and thus may have increasing benefit for patients over time. Further study in this and other patient populations will assess the efficacy of abatacept therapy over longer periods of time. Funding: This study was funded by Bristol-Myers Squibb. Competing interests: HG has been reimbursed by Bristol-Myers Squibb, Amgen, Wyeth, Novartis, Merck, Sanofi-Aventis, Lilly, GSK, Roche and Genentech for speaking fees and honoraria. HG has also received research funding from these organisations. HG is a founder and share holder of Synarc, Inc., and serves as a member of the board of directors. RW has been reimbursed >$10 000 by Bristol-Myers Squibb and Schering-Plough for speaking fees and honoraria. J-CB is an employee of Bristol-Myers Squibb, and has shares with the company. RA is an employee of Bristol-Myers Squibb, and has shares with the company. GV is an employee of Bristol-Myers Squibb, and has shares with the company. JT is an employee of Bristol-Myers Squibb, and has shares with the company. JK has been reimbursed >$10 000 by Bristol-Myers Squibb for speaking fees and honoraria. CP is an employee and shareholder of Synarc, Inc., and is a provider of central image analysis, molecular marker assays and subject recruitment for global clinical trials for numerous pharmaceutical and biotechnology companies across a broad range of therapeutic areas.

A crucial role in the pathogenesis of systemic sclerosis (SSc) seems to be played by the interactions occurring between endothelial cells and T cells. Endothelial cells, after metabolic activation, produce an array of cytokines and growth factors and promote immune cells to adhere and migrate through the vessel wall into the extra-vascular space. This leads to fibroblasts and myoblasts activation and changes in extracellular matrix (ECM) remodelling, resulting in an excessive accumulation of ECM components that causes fibrosis of the skin and internal organs and hypertrophy/hyperplasia of the arterial wall of pulmonary and renal vessels.1 Circulating peripheral blood T cells adhere to endothelial cells by means of an intricate system of adhesion molecules whose expression increases on the cell surface following activation. Lymphocyte function-associated antigen-1 (LFA-1), very late antigen-4 (VLA-4) and L-selectin are expressed on lymphocytes, while their counter-receptors intercellular adhesion molecule-1 (ICAM-1), vascular cell adhesion molecule-1 (VCAM-1) and CD34/endoglycan are expressed on endothelial cells.2 When overexpressed on the activated endothelial layer, VCAM-1, ICAM-1 as well as P-selectin and platelet endothelial cell adhesion molecule-1 (PECAM-1) undergo shedding and their soluble forms, sVCAM-1, sICAM-1, sP-selectin and sPECAM-1, respectively, are detectable in serum and considered to be markers of endothelial cell activity or injury. There is a growing body of evidence that these endothelial activation markers are raised in patients with SSc and that their levels correlate with disease activity and can be regulated by therapy.3

n and sPECAM-1, respectively, are detectable in serum and considered to be markers of endothelial cell activity or injury. There is a growing body of evidence that these endothelial activation markers are raised in patients with SSc and that their levels correlate with disease activity and can be regulated by therapy.3 Endothelin-1 (ET-1) is a potent mitogenic factor mainly produced by endothelial cells, and exerts its biological activity by interacting with two cell membrane-bound receptors named ETA and ETB, expressed on different cells, such as endothelial cells, smooth muscle cells and fibroblasts.4 ET-1 has pleiotropic functions, being a potent vasoconstrictor, and stimulates synthesis and accumulation of ECM proteins by fibroblasts and smooth muscle cells. Since its levels have been found to be increased in SSc, it is believed to have a key role in the pathogenesis of this disease.5 In particular, ET-1 levels have been detected in lung plexiform lesions in patients with pulmonary arterial hypertension (PAH) and there is evidence from experimental models and clinical studies that ET-1 is directly involved in promoting the remodelling of vessel walls leading to an increase of peripheral vascular resistance and hence a rise in pulmonary arterial pressure.6 Recently, a dual endothelin receptor antagonist named bosentan has been shown to be effective and safe for the treatment of patients with SSc with PAH.7 In this study, the biological outcome of bosentan on endothelial activity and T cell activation has been evaluated in patients with SSc-associated PAH. Expression of LFA-1, VLA-4 and L-selectin has been investigated on peripheral blood T cells in patients with SSc-PAH at baseline and after bosentan treatment. Serum levels of soluble VCAM-1, ICAM-1, P-selectin, PECAM-1 and von Willebrand factor (vWF) were also assessed.

en evaluated in patients with SSc-associated PAH. Expression of LFA-1, VLA-4 and L-selectin has been investigated on peripheral blood T cells in patients with SSc-PAH at baseline and after bosentan treatment. Serum levels of soluble VCAM-1, ICAM-1, P-selectin, PECAM-1 and von Willebrand factor (vWF) were also assessed. METHODS Patients A total of 35 patients with SSc were consecutively recruited at the Rheumatology Unit of the University of Bari, Bari, Italy. Mean age was 51.4 years (range 27–75) and mean disease duration was 10.4 years (range 1–28). All the patients were assigned to the limited cutaneous subset, according to the Leroy subset classification.8 A total of 25 healthy donors (HD), coming from the same geographic area and attending the Rheumatology Unit as students, teachers and employers, were enrolled on the basis of their age (mean age 46.7 years, range 25–69). In the patients with SSc, high-resolution chest tomography was carried out to assess interstitial lung fibrosis, as well as respiratory function tests to evaluate pulmonary forced vital capacity (FVC) and alveolar CO diffusion (DLCO), along with full blood count, renal function tests, serum complement levels, liver enzymes, erythrocyte sedimentation rate (ESR). Of 35 patients, 10 had isolated PAH assessed by Doppler echocardiography (tricuspid gradient at rest >35 mmHg), and subsequently confirmed by right heart catheterisation; mean pulmonary artery pressure (mPAP) was 38.6 (9) mmHg, pulmonary vascular resistance was 758 (254) dyne/s/cm, pulmonary capillary wedge pressure was 13 (1.7) mmHg. Patients with SSc-PAH were in New York Heart Association (NYHA) functional classes II–III according to the World Health Organization (WHO) classification, with 6-min walking distance (6-MWD) ranging between 60 and 450 m.7 No patient had left ventricular disease at Doppler echocardiography. Further characteristics of the patients with SSc with PAH are shown in detail in table 1. Patients with SSc-PAH were treated with bosentan at the standard dosage of 62.5 mg twice daily for 4 weeks, followed by 125 mg twice daily for 50 weeks. Additional permitted drugs were intestinal prokinetics and H-2 blockers. No patients were on steroids and/or immunosuppressive agents. Clinical assessment, including 6-MWD and Doppler echocardiography, and collection of serum and heparin blood samples were performed at baseline and after 6 and 12 months of bosentan therapy.

ional permitted drugs were intestinal prokinetics and H-2 blockers. No patients were on steroids and/or immunosuppressive agents. Clinical assessment, including 6-MWD and Doppler echocardiography, and collection of serum and heparin blood samples were performed at baseline and after 6 and 12 months of bosentan therapy. Written informed consent was obtained from all patients and healthy donors according to the declaration of Helsinki and the study was approved by the Ethics Committee of the Policlinico di Bari. Table 1 Patient characteristics Parameter Value Patients with PAH 10 Age, years 56.4 (31–79) Disease duration, years 10.2 (2–24) FVC/DLCO ratio 1.6 (1.37–2.13) ANA (n) 10 SCL70 (n) 2 ACA (n) 2 Values are mean (range). ACA, anticentromere antibodies; ANA, antinuclear antibodies; DLCO, CO diffusion; FVC, forced vital capacity; PAH, pulmonary arterial hypertension. T lymphocyte phenotype Peripheral blood (PB) lymphocytes were isolated by gradient density separation on lymphoprep and cell surface antigen expression on T cells was assessed by double immunoflorescence. Briefly, 5×105 cells were incubated with 5 μl of mAb (CD3–PE, CD11a/LFA1-FITC, CD49d/VLA4-FITC, CD62L/L-selectin-FITC, Immunotech, Marseille, France) for 20′ on ice, washed twice and T cell subsets were evaluated by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, New Jersey, USA). T cell phenotype was evaluated at each timepoint on freshly isolated lymphocytes against a control sample, incubated with rat IgG1-FITC/IgG2-PE (DAKO, Glostrup, Denmark).

rance) for 20′ on ice, washed twice and T cell subsets were evaluated by flow cytometry (FACScan, Becton Dickinson, Franklin Lakes, New Jersey, USA). T cell phenotype was evaluated at each timepoint on freshly isolated lymphocytes against a control sample, incubated with rat IgG1-FITC/IgG2-PE (DAKO, Glostrup, Denmark). Soluble endothelial activation markers Serum levels of endothelial markers were assessed by commercial ELISA: sICAM-1 ELISA kit (Cellular Communication Investigation, Immunotech, Marseille, France), sVCAM-1 ELISA kit (Immuno Biological Laboratories, Hamburg, Germany), vWF ELISA kit (Gentaur, Brussels, Belgium), sPECAM-1 ELISA kit (Bender MedSystems, Wien, Austria) and sP-selectin ELISA kit (Bender MedSystems), following the manufacturers’ instructions. ELISA experiments were run simultaneously on all samples previously frozen at −30°C. Statistical analysis Values are expressed as mean 1 SD) unless otherwise indicated. Analysis of variance (ANOVA) with the post hoc least square differences (LSD) range test was used to compare the different groups at baseline, whereas a paired Student t test was applied to compare patients with SSc-PAH at the different timepoints. The Mann–Whitney U test was used to compare NYHA functional classes over time. Correlations of clinical data with endothelial activation markers and T cell phenotype were carried out using Pearson or Spearman analysis. The significance level was set at p<0.05.

compare patients with SSc-PAH at the different timepoints. The Mann–Whitney U test was used to compare NYHA functional classes over time. Correlations of clinical data with endothelial activation markers and T cell phenotype were carried out using Pearson or Spearman analysis. The significance level was set at p<0.05. RESULTS Clinical outcomes Bosentan treatment significantly improved exercise ability in patients with SSc-PAH with 6-MWD, which rose from 267 (33) m at baseline to 359 (27) m at 6 months and to 376 (40) m at 12 months (p<0.05 vs baseline). Bosentan also induced an improvement of dyspnoea, measured by NYHA, with a significant reduction of the NYHA functional class, observed after 12 months of therapy (median baseline 2.0 vs median 12 months 1.5, p<0.05). PAP values reduced during bosentan treatment, without reaching statistical significance (data not shown). Digital ulcers were present in 8 out of 10 patients and 6 patients did not experience new digital ulcers during bosentan therapy.

ed after 12 months of therapy (median baseline 2.0 vs median 12 months 1.5, p<0.05). PAP values reduced during bosentan treatment, without reaching statistical significance (data not shown). Digital ulcers were present in 8 out of 10 patients and 6 patients did not experience new digital ulcers during bosentan therapy. Adhesion molecule expression on T cells Absolute number of lymphocytes did not change upon bosentan therapy. As shown in fig 1A, PB T cells expressing LFA-1 were significantly higher in patients with SSc-PAH at baseline (mean (SD) 46.3 (6)%) than in HD (32.6 (3)%, p<0.05) or in patients with SSc without PAH (35.8 (5)%, p<0.05). They decreased after 6 months of bosentan therapy (35.5 (5)%, p<0.05 vs baseline) and fell to normal values after 12 months (32.7 (4)%, p<0.05 vs baseline). The proportion of T cells bearing VLA-4 antigen (fig 1B) was significantly reduced in the SSc-PAH group (73.4 (4)%) in comparison with HD (83.3 (6)%, p<0.05) and patients with SSc without PAH (84.8 (5)%, p<0.05), but was not regulated by bosentan treatment after 6 months (70.5 (6)%, not significant vs baseline) nor after 12 months (75.1 (2)%, not significant vs baseline). Expression of L-selectin on T cells (fig 1C) was significantly lower in patients with SSc-PAH (26.4 (8)%) than in the HD group (75.5 (8)%, p<0.01) or in patients with SSc without PAH (75.5 (2)%, p<0.01), and gradually rose following bosentan treatment at the 6 months (43.1 (10)%, p<0.05 vs baseline) and 12 months controls (63.2 (6)%, p<0.01 vs baseline). In fig 2, the representative histograms of the fluorescence intensity for CD3–L-selectin cell subset from a patient with SSc-PAH before and after bosentan therapy, a patient with SSc without PAH and a healthy donor are shown.

onths (43.1 (10)%, p<0.05 vs baseline) and 12 months controls (63.2 (6)%, p<0.01 vs baseline). In fig 2, the representative histograms of the fluorescence intensity for CD3–L-selectin cell subset from a patient with SSc-PAH before and after bosentan therapy, a patient with SSc without PAH and a healthy donor are shown. Figure 1 Percentage of peripheral blood T cells expressing lymphocyte function antigen-1 (LFA-1) (A), very late antigen-4 (VLA-4) (B) and L-selectin (C) in healthy donors (HD), patients with systemic sclerosis (SSc) without pulmonary arterial hypertension and patients with SSc with pulmonary arterial hypertension (SSc-PAH) at baseline (t 0), after 6 months (t 6) and 12 months (t 12) of bosentan therapy. Values are mean (SD). Figure 2 Fluorescence intensity for CD3–L-selectin cell subset from a patient with SSc-PAH at baseline and after 12 months of bosentan therapy (upper panels), from a patient with SSc without PAH (middle panels) and from a healthy donor (lower panels). Negative control histograms (a fluorescent non-binding mAb) are shown on the left. The longitudinal axis shows the percentage of lymphocytes positive for CD3 and L-selectin and the horizontal axis the mean channel fluorescence. Serum endothelial markers Serum levels of soluble vWF (fig 3A) were increased in patients with SSc-PAH (1043 (97) mU/ml) and patients with SSc without PAH (1061 (86) mU/ml) in comparison with the HD group (770 (54) mU/ml, p<0.05) and did not change after 12 months of bosentan therapy (980 (89) mU/ml, not significant vs baseline). Serum levels of sP-selectin (fig 3B) were significantly higher in the SSc-PAH group (367 (42) ng/ml) and in patients with SSc without PAH (362 (61) ng/ml) than in HD (132 (12) ng/ml, p<0.01), and significantly decreased after 12 months of bosentan therapy (211 (31) ng/ml, p<0.05, vs baseline). Soluble PECAM serum levels were similar in patients with SSc-PAH (48.7 (3) ng/ml), in patients with SSc without PAH (46.7 (4) ng/ml) and in HD (41.1 (2) ng/ml) and were not modulated by bosentan treatment (37.7 (4) ng/ml). Serum levels of sVCAM-1 (fig 3C) were significantly higher in patients with SSc without PAH (300 (16) ng/ml) than in the HD group (261 (9) ng/ml, p<0.05), and were also increased in patients with SSc-PAH (303 (3) ng/ml) but without reaching statistical significance. However, sVCAM-1 levels significantly dropped after 12 weeks of bosentan therapy (229 (15) ng/ml, p<0.01 vs baseline).

ients with SSc without PAH (300 (16) ng/ml) than in the HD group (261 (9) ng/ml, p<0.05), and were also increased in patients with SSc-PAH (303 (3) ng/ml) but without reaching statistical significance. However, sVCAM-1 levels significantly dropped after 12 weeks of bosentan therapy (229 (15) ng/ml, p<0.01 vs baseline). Furthermore, sICAM-1 serum levels (fig 3D) were significantly higher in patients with SSc-PAH (2536 (647) pg/ml) and patients with SSc without PAH (3293 (1006) pg/ml) than in HD (588 (48) pg/ml, p<0.01), and declined to normal levels following bosentan treatment (696 (98) pg/ml, p<0.05 vs baseline). Finally, soluble PECAM levels (fig 3E) were increased in patients with SSc-PAH (46.7 (4) ng/ml) in comparison with HD (41 (2) ng/ml), but the difference was not statistically significant and significantly reduced after bosentan treatment (37 (3) ng/ml; p<0.05).

g bosentan treatment (696 (98) pg/ml, p<0.05 vs baseline). Finally, soluble PECAM levels (fig 3E) were increased in patients with SSc-PAH (46.7 (4) ng/ml) in comparison with HD (41 (2) ng/ml), but the difference was not statistically significant and significantly reduced after bosentan treatment (37 (3) ng/ml; p<0.05). Figure 3 Serum levels of soluble von Willebrand Factor (vWF) (A), P-selectin (B), vascular cell adhesion molecule 1 (VCAM-1) (C), intercellular adhesion molecule-1 (ICAM-1) (D) and platelet endothelial cell adhesion molecule-1 (PECAM-1) (E) in healthy donors (HD), patients with systemic sclerosis (SSc) without pulmonary arterial hypertension and patients with SSc with pulmonary arterial hypertension (SSc-PAH) at baseline (t 0) and after 12 months (t 12) of bosentan therapy. Values are mean SD). No significant correlations were found between adhesion molecules, either soluble or expressed on T cells, and clinical parameters such as modified Rodnan skin score, 6-MWD, NYHA classes, new digital ulcers, FVC and DLCO. Furthermore, no correlations between adhesion molecule changes over time and treatment response were detected.

nt correlations were found between adhesion molecules, either soluble or expressed on T cells, and clinical parameters such as modified Rodnan skin score, 6-MWD, NYHA classes, new digital ulcers, FVC and DLCO. Furthermore, no correlations between adhesion molecule changes over time and treatment response were detected. DISCUSSION The pathophysiology of SSc is multifaceted and different types of cells, such as fibroblasts, immune cells, inflammatory cells, endothelial cells, smooth muscle cells, pericytes, myointimal cells, are involved. Each disease picture is unique because the pathological process may lead to different tissue damage depending on the organ involved.1 In recent years, the study of pulmonary vasculature involvement has posed a strong challenge since PAH dramatically affects the clinical outcome and survival of patients with SSc, especially those classified within the limited cutaneous subset.9

cess may lead to different tissue damage depending on the organ involved.1 In recent years, the study of pulmonary vasculature involvement has posed a strong challenge since PAH dramatically affects the clinical outcome and survival of patients with SSc, especially those classified within the limited cutaneous subset.9 A widely accepted hypothesis is that, maybe on a genetic basis and following unknown external stimuli, endothelial cells and lymphocytes become activated and stimulate fibroblasts and smooth muscle cells to proliferate and produce collagen and other ECM components.1 10 Endothelial damage would let leucocytes migrate through the vessels into the extra vascular tissue and promote the activation of the innate immunity. This local inflammatory response would be self limiting without the transition to the adoptive immunity with activation of T cells that sustain a chronic inflammatory state and stimulate fibrosis by directly interacting with fibroblasts and producing fibrogenic mediators, such as interleukin (IL)4, IL13 and transforming growth factor (TGF)-β1.1 Focusing on the interplays between endothelial cells and lymphocytes may be crucial to unravel the early events occurring in SSc pathogenesis. Circulating T cells interact with endothelial layer through an intricate system of receptors and their ligands, generally known as “adhesion molecules”.2 L-selectin and P-selectin, oligosaccharide proteins belonging to the family of selectins, are expressed on T lymphocytes and endothelial cells, respectively, and reciprocally interact to allow loose adhesion of circulating lymphocytes. If a firm adhesion, which preludes migration through the vessels into the extra-vascular space, occurs, endothelial cells and T cells undergo activation and modulate the expression of their surface receptors.11 ICAM-1 and VCAM-1 are upregulated on activated endothelial cells and interact with LFA-1 and VLA-4, respectively, expressed on activated T cells. When overexpressed on the cell surface, some of these proteins are shed in biological fluids as a regulatory mechanism, and the soluble forms, sVCAM-1, sICAM-1 and sP-selectin, can be dosed and used as markers of endothelial cell activation. Many authors have shown that serum activation markers are increased in patients with SSc, and have attempted to correlate them to disease activity.3 12–19

iological fluids as a regulatory mechanism, and the soluble forms, sVCAM-1, sICAM-1 and sP-selectin, can be dosed and used as markers of endothelial cell activation. Many authors have shown that serum activation markers are increased in patients with SSc, and have attempted to correlate them to disease activity.3 12–19 Among the array of cytokines involved, ET-1, mainly released by endothelial cells, has been shown to play a physiological role in controlling vascular tone by regulating vasodilatation and vasoconstriction. In SSc, changes of crosstalk between ET-1 and its receptors due to a different distribution of ETA and ETB receptors break down this balance, causing vasoconstriction to predominate. Besides vascular tone alterations, ET-1 stimulates smooth muscle cells and myofibroblasts and promote vasoproliferative changes leading to increased pulmonary vasculature resistances and subsequent PAH.5 The importance of ET-1 in the pathogenesis of PAH is further corroborated by the beneficial effects of the dual endothelin receptors inhibitor, bosentan, in the treatment of PAH.7 20

ells and myofibroblasts and promote vasoproliferative changes leading to increased pulmonary vasculature resistances and subsequent PAH.5 The importance of ET-1 in the pathogenesis of PAH is further corroborated by the beneficial effects of the dual endothelin receptors inhibitor, bosentan, in the treatment of PAH.7 20 In this study we evaluated the expression of LFA-1, VLA-4 and L-selectin antigens on circulating T cells and serum levels of soluble vWF, ICAM-1, VCAM-1 and P-selectin in patients with SSc with associated PAH, at baseline and after 12 months of bosentan therapy, as compared to patients with SSc without PAH and healthy subjects. Our results confirm that bosentan treatment improves exercise ability and dyspnoea in patients with SSc-PAH,7 and provide further evidence that endothelial activation occurs in SSc, since soluble endothelial activation markers were significantly higher in the whole SSc group than in controls. Interestingly, whereas vWF antigen levels did not change after bosentan therapy, serum levels of sICAM-1, sVCAM-1, sPECAM and sP-selectin fell to normal values following blocking of the ETA and ETB receptors. These findings are apparently in contrast with those recently reported by Sfikakis et al who did not detect changes in sICAM-1 levels in patients with SSc following bosentan treatment.21 However, the latter was a short-term study in which sICAM-1 was assessed in 10 patients with SSc treated with bosentan at lower dosage for 4 months.

are apparently in contrast with those recently reported by Sfikakis et al who did not detect changes in sICAM-1 levels in patients with SSc following bosentan treatment.21 However, the latter was a short-term study in which sICAM-1 was assessed in 10 patients with SSc treated with bosentan at lower dosage for 4 months. The behaviour of T cell subsets bearing adhesion molecules was noteworthy. T cells expressing LFA-1, the counter-receptor of ICAM-1, were significantly and selectively enhanced in the SSc-PAH group at baseline, and LFA-1 expression was downregulated by bosentan therapy. The reduction of VLA-4 on T cells from patients with SSc-PAH was unforeseen as its expression is expected to increase on activated T cells and the meaning of these findings is unknown. VLA-4 has pleiotropic functions and it is involved in T cells/endothelial cells interactions as well as in T cells/stromal cells interplay. The finding that VLA-4 was selectively reduced on peripheral T cells from patients with SSc-PAH but not in patients with SSc without PAH and HD, suggests that this molecule may be directly involved in SSc-PAH pathogenesis. However, since bosentan did not modify VLA-4 expression on T cells, it is conceivable that ET-1 does not regulate VLA-4. The pattern of L-selectin was remarkable in that its expression on T cells was lower in patients with SSc-PAH than in healthy donors or patients with SSc without PAH, and strikingly increased to normal values after bosentan treatment. These data suggest that bosentan, a dual endothelin receptor antagonist approved for the therapy of PAH, is able to restore T cell functions in patients with SSc with PAH in vivo. We can speculate that blocking both ET-1 receptors can antagonise the detrimental effects of the imbalance of the ET-1 system in SSc. How inhibition of ET-1 receptors may regulate LFA-1 and L-selectin expression on circulating T cells can only be a matter of speculation at the moment. This may be indirectly mediated by the microvasculature, which expresses ETA and ETB receptors, and changes of adhesion molecules on endothelial cells might modify the expression of their respective ligands on circulating T cells. Additionally, it may be assumed that T cells may bear ET-1 receptors on their membrane and that the changes of LFA-1 and L-selectin expression we detected in our patients with SSc-PAH are due to downstream signals following direct stimulation by ET-1, subsequently inhibited by bosentan.

ligands on circulating T cells. Additionally, it may be assumed that T cells may bear ET-1 receptors on their membrane and that the changes of LFA-1 and L-selectin expression we detected in our patients with SSc-PAH are due to downstream signals following direct stimulation by ET-1, subsequently inhibited by bosentan. This hypothesis needs to be verified by further studies in vitro addressed to show the expression of ET-1 receptors by T cells and the role of ET-1 in modulating T cell phenotype. However, an analogous mechanism has already been shown on human neutrophils. ET-1 downregulated the expression of L-selectin and upregulated the expression of LFA-2 on neutrophils and these effects were selectively prevented by a specific ETA receptor antagonist.22 These data seem to be consistent with our results and coherent with the sequential two-step action of these molecules. Resting leucocytes highly express L-selectin and loosely adhere to endothelial cells, the so-called “tethering”, and roll on the endothelial layer. After activation, endothelial cells upregulate ICAM-1 and leucocytes downregulate L-selectin and simultaneously increase LFA-1 expression. LFA-1/ICAM-1 interactions would lead leucocytes to firmly adhere and migrate through the vessels. Furthermore, LFA-1/ICAM-1 ligation may be involved in T cells/fibroblasts interactions. In SSc-PAH, ET-1 could stimulate these events on T cells, which become activated, with high LFA-1 and low L-selectin expression, and migrate in the extra-vascular space, while bosentan down-modulates these processes and prevents further priming of circulating naïve T cells. However, it should be taken into account that dermal fibroblasts may be a further source of increased serum levels of adhesion molecules. ICAM-1 expression on surface fibroblasts is modulated by ET-1 and inflammatory cytokines, is involved in cell–cell and cell–matrix interactions and plays a key role in regulating inflammatory cells binding to dermal tissue.23 24

dermal fibroblasts may be a further source of increased serum levels of adhesion molecules. ICAM-1 expression on surface fibroblasts is modulated by ET-1 and inflammatory cytokines, is involved in cell–cell and cell–matrix interactions and plays a key role in regulating inflammatory cells binding to dermal tissue.23 24 In conclusion, this study provides evidence that ET-1 can induce changes in the T cell/endothelium interplay in SSc-associated PHA and that blocking ET-1 by the administration of bosentan can restore these interactions. Funding: This work was supported by grants from the University of Bari, Italy. Competing interests: None declared. Ethics approval: The study was approved by the Ethics Committee of the Policlinico di Bari. Patient consent: Written informed consent was obtained from all patients and healthy donors according to the declaration of Helsinki.

Abatacept is a selective T-cell co-stimulation modulator, that modulates the CD80/CD86:CD28 co-stimulatory signal required for full T-cell activation.1 The mechanism of action of abatacept is fundamentally different to that of other biological disease-modifying antirheumatic drugs (DMARDs) for the treatment of rheumatoid arthritis (RA). The efficacy of abatacept has previously been demonstrated in patients with RA and an inadequate response to methotrexate (MTX)2 and anti-tumour necrosis factor (TNF) agents,3 respectively. The ATTEST (for “Abatacept or infliximab vs placebo, a Trial for Tolerability, Efficacy and Safety in Treating rheumatoid arthritis”) trial was designed to obtain data on the magnitude of the treatment effect in RA of abatacept or infliximab (an established inhibitor of TNF for RA) vs placebo, and to obtain relative efficacy and safety data on these two biological treatments in a single study. The study utilised a double-blind, randomised, placebo-controlled design for the first 6 months to validate efficacy responses, and the study duration allowed for the opportunity to directly compare the safety profile of the active biologic treatment groups over 1 year.

hese two biological treatments in a single study. The study utilised a double-blind, randomised, placebo-controlled design for the first 6 months to validate efficacy responses, and the study duration allowed for the opportunity to directly compare the safety profile of the active biologic treatment groups over 1 year. METHODS Patients Eligible patients met the American College of Rheumatology (ACR) criteria for RA, were at least 18 years of age, had RA for at least 1 year,4 and had an inadequate response to MTX, as demonstrated by ongoing active disease (at randomisation ⩾10 swollen joints, ⩾12 tender joints, and C-reactive protein (CRP) levels ⩾1 mg/dl using a high sensitivity assay (upper limit of the normal range, 0.5)). All patients had received MTX ⩾15 mg/week for ⩾3 months prior to randomisation (stable for at least 28 days) and washed out all DMARDs (⩾28 days prior) except for MTX. No prior experience of abatacept or anti-TNF therapy was permitted. All patients were screened for tuberculosis (TB) by purified protein derivative (PPD) testing and chest x ray. The protocol used for TB screening was the same as that employed in the “Anti-TNF Trial in rheumatoid arthritis with Concomitant Therapy” (ATTRACT) trial.5

ience of abatacept or anti-TNF therapy was permitted. All patients were screened for tuberculosis (TB) by purified protein derivative (PPD) testing and chest x ray. The protocol used for TB screening was the same as that employed in the “Anti-TNF Trial in rheumatoid arthritis with Concomitant Therapy” (ATTRACT) trial.5 Concomitant medications were permitted between days 1–197: oral corticosteroids (⩽10 mg of prednisone or equivalent daily (stable for ⩾25 out of 28 days prior to randomisation)), and/or stable non-steroidal anti-inflammatory drugs (NSAIDs) including acetyl salicylic acid, and analgesics not containing aspirin or NSAIDs. No MTX dose adjustments were permitted except in the occurrence of adverse events (AEs). Between days 198–365, dose modification was permitted for MTX (⩽25 mg weekly) and oral corticosteroids (⩽10 mg prednisone or equivalent daily); hydroxychloroquine, sulfasalazine, gold, or azathioprine were also permitted. Premedication prior to infusions of study drug was left at the discretion of the investigator (not required by protocol). Study design ATTEST was a randomised, double-blind, double-dummy, placebo- and active (infliximab)-controlled, 12-month global trial. Adult patients with active RA and an inadequate response to MTX were randomised by centre in a 3:3:2 ratio to 6 months of abatacept (approximating 10 mg/kg), infliximab (3 mg/kg), or placebo treatment by intravenous (IV) infusion, on a background of MTX. Assessors, physicians and patients were blinded to the treatment group assignment for 1 year.

A and an inadequate response to MTX were randomised by centre in a 3:3:2 ratio to 6 months of abatacept (approximating 10 mg/kg), infliximab (3 mg/kg), or placebo treatment by intravenous (IV) infusion, on a background of MTX. Assessors, physicians and patients were blinded to the treatment group assignment for 1 year. The study was approved by the institutional review boards and independent ethics committees at participating sites, and was carried out in accordance with the ethical principles of the Declaration of Helsinki. All patients provided written informed consent before randomisation. Treatment with placebo was limited to days 1–197 to provide internal validity to the trial design and the clinical response rates of the two active treatment groups. On day 198, placebo-treated patients were reallocated to abatacept (with blinding maintained). Patients initially randomised to abatacept or infliximab continued their treatment. Clinical efficacy and safety assessments are presented for the abatacept, placebo and infliximab groups up to day 197, and for the abatacept and infliximab groups up to day 365 (excluding patients in the placebo group who were reallocated to abatacept at day 198).

d to abatacept or infliximab continued their treatment. Clinical efficacy and safety assessments are presented for the abatacept, placebo and infliximab groups up to day 197, and for the abatacept and infliximab groups up to day 365 (excluding patients in the placebo group who were reallocated to abatacept at day 198). Study drugs Abatacept was dosed according to weight: patients weighing less than 60 kg, 60–100 kg, or more than 100 kg received 500 mg, 750 mg, or 1000 mg of abatacept, respectively. Infliximab was dosed at 3 mg/kg for all patients. Abatacept was administered by IV infusion on days 1, 15 and 29, and every 28 days thereafter, up to and including day 337 (with normal saline received on day 43). Infliximab was administered on days 1, 15, 43 and 85, and every 56 days thereafter (normal saline was received at the remaining visit days). Two IV bags were infused simultaneously to ensure blinding to treatment group assignment, one over 30 min (abatacept or placebo) and one over 2 h (infliximab or placebo).

day 43). Infliximab was administered on days 1, 15, 43 and 85, and every 56 days thereafter (normal saline was received at the remaining visit days). Two IV bags were infused simultaneously to ensure blinding to treatment group assignment, one over 30 min (abatacept or placebo) and one over 2 h (infliximab or placebo). Objectives The primary endpoint was to evaluate a reduction in disease activity, measured by Disease Activity Score 28 (based on erythrocyte sedimentation rate levels; DAS28 (ESR)) with abatacept vs placebo at 6 months. Secondary endpoints included mean reduction in DAS28 (ESR) with infliximab vs placebo at 6 months. Additional secondary endpoints at 6 months and 1 year included: mean reduction in DAS28 (ESR) with abatacept vs infliximab; DAS28 (ESR) European League Against Rheumatism (EULAR) responses;6 low disease activity score (LDAS; DAS28 (ESR) ⩽3.2); DAS28 (ESR)-defined remission (DAS28 (ESR) <2.6); ACR 20, 50 and 70 responses;7 Health Assessment Questionnaire Disability Index (HAQ-DI) response rates (⩾0.3 improvement from baseline); and mean changes in the physical and mental component summary (PCS and MCS, respectively) scores, and eight subscales of the Short Form-36 (SF-36). Tertiary endpoints included comparative safety at 1 year between the abatacept and infliximab groups.

bility Index (HAQ-DI) response rates (⩾0.3 improvement from baseline); and mean changes in the physical and mental component summary (PCS and MCS, respectively) scores, and eight subscales of the Short Form-36 (SF-36). Tertiary endpoints included comparative safety at 1 year between the abatacept and infliximab groups. Clinical assessments Disease activity was assessed using the validated DAS28 (ESR).8 The EULAR criteria were used to assess good responses (endpoint DAS28 (ESR) of ⩽3.2 and an improvement from baseline of ⩾1.2).6 Signs and symptoms of RA were evaluated using ACR 20, 50 and 70 response rates based on the ACR criteria.7 Physical function was assessed using the HAQ-DI.9 Health-related quality of life (HRQoL) was assessed using the SF-36.10 Safety All patients who received at least one dose of study drug were evaluated for safety, including all reported AEs, serious AEs (SAEs), and clinically significant changes in vital signs, physical exams, and clinical laboratory test abnormalities. Sample size calculation The sample size and power were calculated to detect a treatment difference in the primary analysis of a mean change from baseline in DAS28 (ESR) for the abatacept vs placebo groups at day 197. Prospectively, this study was not powered with a superiority or non-inferiority design to compare the two active arms.

culation The sample size and power were calculated to detect a treatment difference in the primary analysis of a mean change from baseline in DAS28 (ESR) for the abatacept vs placebo groups at day 197. Prospectively, this study was not powered with a superiority or non-inferiority design to compare the two active arms. Statistical analyses All patients who received at least one dose of study medication were assessed for efficacy and safety (intent-to-treat population). At day 197, the abatacept or infliximab groups were compared with the placebo group by analyses of covariance (ANCOVA) for mean changes from baseline in DAS28 (ESR) and in the SF-36 (PCS and MCS). The model included the change as the dependent variable, with treatment group as a main effect and the baseline score as an additional covariate. The proportion of patients with ACR 20, 50 and 70 responses, LDAS, DAS28-defined remission, a good EULAR response and a clinical meaningful HAQ-DI response was calculated. The χ2 test was performed to evaluate the differences (and 95% CIs) between the abatacept or infliximab groups and placebo. At day 365, the reference group was changed to infliximab.

R 20, 50 and 70 responses, LDAS, DAS28-defined remission, a good EULAR response and a clinical meaningful HAQ-DI response was calculated. The χ2 test was performed to evaluate the differences (and 95% CIs) between the abatacept or infliximab groups and placebo. At day 365, the reference group was changed to infliximab. Patients who discontinued the study prematurely were considered as non-responders subsequent to the time of discontinuation for ACR 20, 50 and 70 responses, good EULAR responses and clinically meaningful HAQ-DI responses. For all continuous measurements (mean changes in DAS28, SF-36 and the HAQ-DI score), LDAS and DAS28-defined remission the last observations prior to the discontinuation were carried forward (LOCF). To access the effect of antirheumatic medications on the abatacept and infliximab treatment groups, a predefined sensitivity analysis was conducted on the data with the last DAS28 (ESR) score just prior to the initiation of the additional DMARD, or any increase in MTX or corticosteroid use during treatment days 198–365 carried forward. Safety was assessed at each visit. Summary statistics were tabulated by treatment group at days 197 and 365.

Patients who discontinued the study prematurely were considered as non-responders subsequent to the time of discontinuation for ACR 20, 50 and 70 responses, good EULAR responses and clinically meaningful HAQ-DI responses. For all continuous measurements (mean changes in DAS28, SF-36 and the HAQ-DI score), LDAS and DAS28-defined remission the last observations prior to the discontinuation were carried forward (LOCF). To access the effect of antirheumatic medications on the abatacept and infliximab treatment groups, a predefined sensitivity analysis was conducted on the data with the last DAS28 (ESR) score just prior to the initiation of the additional DMARD, or any increase in MTX or corticosteroid use during treatment days 198–365 carried forward. Safety was assessed at each visit. Summary statistics were tabulated by treatment group at days 197 and 365. RESULTS Baseline demographics and clinical characteristics A total of 431 patients were randomised and treated with abatacept (n = 156), placebo (n = 110) or infliximab (n = 165) (fig 1). Baseline demographics and clinical characteristics were similar between groups (table 1). At randomisation, patients had active disease despite background MTX (mean DAS28 (ESR) of 6.8–6.9, tender joint counts 30.3–31.7, swollen joint counts 20.1–21.3 and mean HAQ-DI of 1.7–1.8). The mean dose of MTX was 16.3–16.6 mg and mean treatment duration was 18.3–23.7 months.

ics were similar between groups (table 1). At randomisation, patients had active disease despite background MTX (mean DAS28 (ESR) of 6.8–6.9, tender joint counts 30.3–31.7, swollen joint counts 20.1–21.3 and mean HAQ-DI of 1.7–1.8). The mean dose of MTX was 16.3–16.6 mg and mean treatment duration was 18.3–23.7 months. Figure 1 Patient disposition over 1 year. The ATTEST trial was a 12-month global trial conducted at 86 sites in the US (20 sites), Europe (18 sites (5 in Poland, 4 in Spain, 4 in Sweden, 2 in Russia, 2 in Denmark and 1 in Switzerland)), Canada (11 sites), Australia (6 sites), Mexico (10 sites), Argentina (5 sites), Brazil (8 sites), Peru (5 sites) and South Africa (3 sites). Patients were randomised in a 3:3:2 ratio to 6 months of abatacept (approximating 10 mg/kg), infliximab (3 mg/kg), or placebo treatment. During days 198–365, efficacy and safety data are not presented for the placebo group following reallocation to abatacept. Table 1 Baseline demographics and clinical characteristics Demographic/characteristic Abatacept + MTX (n = 156) Placebo + MTX (n = 110) Infliximab + MTX (n = 165) Age, years (SD) 49.0 (12.5) 49.4 (11.5) 49.1 (12.0) Gender, % female 83.3 87.3 82.4 Race, % Caucasian 80.8 76.4 80.6 Geographic origin: North America, n (%) 16 (10.3) 10 (9.1) 15 (9.1) South America, n (%) 93 (59.6) 66 (60.0) 96 (58.2) Europe, n (%) 39 (25.0) 29 (26.4) 39 (23.6) Rest of the world, n (%) 8 (5.1) 5 (4.5) 15 (9.1) Disease duration, years (SD) 7.9 (8.5) 8.4 (8.6) 7.3 (6.2) Tender joints, n (SD) 31.6 (13.9) 30.3 (11.7) 31.7 (14.5) Swollen joints, n (SD) 21.3 (8.6) 20.1 (7.0) 20.3 (8.0) Erythrocyte sedimentation rate, mm/h (SD) 49.4 (31.2) 47.0 (32.6) 47.8 (30.4) C-reactive protein levels, mg/dl (SD) 3.1 (2.7) 2.7 (2.6) 3.3 (3.2) DAS28 (ESR), n (SD) 6.9 (1.0) 6.8 (1.0) 6.8 (0.9) HAQ-DI, 0–3 (SD) 1.8 (0.6) 1.8 (0.7) 1.7 (0.7) Rheumatoid factor positive, n (%) 136 (87.2) 85 (77.3) 140 (84.8) Concomitant medications Total patients on concomitant medications, n (%) 156 (100) 110 (100) 165 (100) MTX, n (%) 156 (100) 110 (100) 164 (99.4) Dose, mg/week (SD) 16.5 (3.7) 16.6 (3.7) 16.3 (3.6) Duration, months (SD) 18.3 (20.0) 23.7 (25.6) 23.6 (26.8) Corticosteroids, n (%) 118 (75.6) 77 (70.0) 118 (71.5) NSAIDs, n (%) 133 (85.3) 93 (84.5) 142 (86.1) MTX, methotrexate; DAS28 (ESR), Disease Activity Score 28 (based on erythrocyte sedimentation rate levels); HAQ-DI, Health Assessment Questionnaire Disability Index; NSAID, non-steroidal anti-inflammatory drug.

6 (26.8) Corticosteroids, n (%) 118 (75.6) 77 (70.0) 118 (71.5) NSAIDs, n (%) 133 (85.3) 93 (84.5) 142 (86.1) MTX, methotrexate; DAS28 (ESR), Disease Activity Score 28 (based on erythrocyte sedimentation rate levels); HAQ-DI, Health Assessment Questionnaire Disability Index; NSAID, non-steroidal anti-inflammatory drug. Use of additional medications Concomitant medications and NSAIDs were used by a similar proportion of patients across treatment groups at randomisation (table 1). Between days 198–365, when the protocol permitted adjustments to background medications, 12.8% and 17.6% of abatacept and infliximab-treated patients, respectively, added a DMARD, or increased their dose of MTX/corticosteroids from baseline. Discontinuations During the first 6 months, discontinuations occurred in 5.8, 2.7 and 7.9% of the abatacept, placebo and infliximab groups, respectively. Between days 198–365, 5.1 and 6.7%, of the abatacept and infliximab groups, respectively, discontinued. Discontinuation due to AEs and SAEs were highest in the infliximab group in both periods (fig 1). Similar numbers of patients from the abatacept and infliximab groups completed 1 year of treatment (fig 1).

pectively. Between days 198–365, 5.1 and 6.7%, of the abatacept and infliximab groups, respectively, discontinued. Discontinuation due to AEs and SAEs were highest in the infliximab group in both periods (fig 1). Similar numbers of patients from the abatacept and infliximab groups completed 1 year of treatment (fig 1). Clinical efficacy DAS28 At day 197, the reduction in DAS28 (ESR), was significantly greater with abatacept vs placebo (–2.53 vs –1.48, p<0.001; fig 2A). A greater proportion of patients also experienced a good EULAR response (20.0 vs 10.8%), LDAS (20.7 vs 10.8%) and were in DAS28 (ESR)-defined remission (11.3 vs 2.9%), for abatacept vs placebo, respectively (fig 2B). Reductions in DAS28 (ESR) were also significantly greater in the infliximab vs placebo groups at day 197 (–2.25 vs –1.48, p<0.001; fig 2A), with a higher proportion of patients experiencing a good EULAR response (22.9 vs 10.8%), LDAS (25.6 vs 10.8%) and were in DAS28 (ESR)-defined remission (12.8 vs 2.9%), respectively (fig 2B). At day 197, the relative efficacy of abatacept and infliximab as assessed by the DAS28 (ESR) was similar.

1.48, p<0.001; fig 2A), with a higher proportion of patients experiencing a good EULAR response (22.9 vs 10.8%), LDAS (25.6 vs 10.8%) and were in DAS28 (ESR)-defined remission (12.8 vs 2.9%), respectively (fig 2B). At day 197, the relative efficacy of abatacept and infliximab as assessed by the DAS28 (ESR) was similar. Figure 2 Disease Activity Score 28 (DAS28) based on erythrocyte sedimentation rates (ESR). A. DAS28 (ESR) mean changes from baseline at days 197 and 365. Error bars represent standard error of the mean. B. European League Against Rheumatism (EULAR) good responses, low DAS (LDAS; DAS28 ⩽3.2) and DAS28 (ESR)-defined remission at day 197 and at day 365. Data are presented for the intent-to-treat population with a last-observation carried forward analysis for mean changes in DAS28 (ESR), LDAS and DAS28 (ESR)-defined remission. Good EULAR responses were presented for the intent-to-treat population with patients who discontinued the study prematurely considered as non-responders subsequent to the time of discontinuation. Error bars show standard error of the mean. *Adjustment based on covariance with treatment as factor and baseline as covariant. At day 365, a greater reduction in DAS28 (ESR) was observed with abatacept than with infliximab (fig 2A; –2.88 vs –2.25; estimate of difference (95% CI) = –0.62 (–0.96, –0.29)). Also, the proportion of patients achieving a good EULAR response (32.0 vs 18.5%, estimate of difference (95% CI) = 13.5% (3.6, 23.3)), LDAS (35.3 vs 22.4%, estimate of difference (95% CI) = 12.9 (2.1, 23.7)), and DAS28 (ESR)-defined remission (18.7 vs 12.2%, estimate of difference (95% CI) = 18.7 (–2.2, 15.2)) were higher with abatacept compared with infliximab (fig 2B).

AR response (32.0 vs 18.5%, estimate of difference (95% CI) = 13.5% (3.6, 23.3)), LDAS (35.3 vs 22.4%, estimate of difference (95% CI) = 12.9 (2.1, 23.7)), and DAS28 (ESR)-defined remission (18.7 vs 12.2%, estimate of difference (95% CI) = 18.7 (–2.2, 15.2)) were higher with abatacept compared with infliximab (fig 2B). When disease activity was assessed using a sensitivity analysis (LOCF prior to increase in antirheumatic medications was performed), consistent improvements were demonstrated with abatacept: DAS28 (ESR) mean changes were significantly greater with abatacept vs infliximab at day 365 (–2.8 vs –2.2, respectively (difference from infliximab of –0.7, 95% CI of –1.0, –0.3)).

ity analysis (LOCF prior to increase in antirheumatic medications was performed), consistent improvements were demonstrated with abatacept: DAS28 (ESR) mean changes were significantly greater with abatacept vs infliximab at day 365 (–2.8 vs –2.2, respectively (difference from infliximab of –0.7, 95% CI of –1.0, –0.3)). American College of Rheumatology response rates At day 197, ACR 20, 50 and 70 responses were significantly greater with abatacept vs placebo (ACR 20: 66.7 vs 41.8%, p<0.001; ACR 50: 40.4 vs 20.0%, p<0.001; and ACR 70: 20.5 vs 9.1%, p = 0.019). ACR 20, 50 and 70 responses were also significantly higher in the infliximab group vs placebo (ACR 20: 59.4 vs 41.8%, p = 0.006; ACR 50: 37.0 vs 20.0%, p = 0.004; and ACR 70: 24.2 vs 9.1%, p = 0.002). The onset of response, as assessed by ACR 20 response rates, was generally more rapid for infliximab compared with abatacept (fig 3), however, by day 85, responses were similar (fig 3). Abatacept and infliximab demonstrated similar responses at day 197. During the second 6 months of the trial, the responses associated with abatacept were maintained, while those observed with infliximab were not. At day 365, ACR 20 responses were higher with abatacept than with infliximab (ACR 20: 72.4 vs 55.8%, difference of 16.7, 95% CI = 5.5, 27.8). In addition, the percentages of ACR 50 and 70 responders were numerically higher with abatacept vs infliximab treatment (with overlapping 95% CIs for the estimate of difference for ACR 50: 45.5 vs 36.4%, estimate of difference (95% CI) = 9.1 (–2.2, 20.5); ACR 70: 26.3 vs 20.6%, estimate of difference (95% CI) = 5.7 (–4.2, 15.6), respectively).

ges of ACR 50 and 70 responders were numerically higher with abatacept vs infliximab treatment (with overlapping 95% CIs for the estimate of difference for ACR 50: 45.5 vs 36.4%, estimate of difference (95% CI) = 9.1 (–2.2, 20.5); ACR 70: 26.3 vs 20.6%, estimate of difference (95% CI) = 5.7 (–4.2, 15.6), respectively). Figure 3 American College of Rheumatology (ACR) responses over 1 year. Proportion of patients with ACR 20, 50 and 70 responses was assessed on each visit day. Data are presented for the intent-to-treat population with a last-observation carried forward analysis. *Infliximab was administered on days 1, 15, 43, 85 and then every 56 days thereafter. Abatacept dosing occured at each visit day presentation following the assesment of efficacy. Physical function At 6 months, significantly more patients in the abatacept group than in the placebo group demonstrated a clinically meaningful improvement in physical function (HAQ-DI responses: 61.5 vs 40.9%, respectively, p = 0.001). Similarly, significantly more patients in the infliximab group vs the placebo group achieved a HAQ-DI response (58.8 vs 40.9%, p = 0.005). At day 365, HAQ-DI responses were maintained in the abatacept and infliximab groups (57.7 and 52.7%, respectively, estimate of difference (95% CI) = 5.0 (–6.5, 16.5)).

, p = 0.001). Similarly, significantly more patients in the infliximab group vs the placebo group achieved a HAQ-DI response (58.8 vs 40.9%, p = 0.005). At day 365, HAQ-DI responses were maintained in the abatacept and infliximab groups (57.7 and 52.7%, respectively, estimate of difference (95% CI) = 5.0 (–6.5, 16.5)). Health-related quality of life Patients in the abatacept group experienced statistically significantly greater improvements from baseline in the PCS (p<0.001) and MCS (p = 0.004) of the SF-36, and in each of the eight individual subscales compared with placebo, following 6 months of treatment (fig 4A). Patients in the infliximab group also experienced significantly greater improvements from baseline in the PCS (p = 0.002) and MCS (p = 0.027), and all eight subscales of the SF-36 at day 197 compared with patients in the placebo group. At day 365, greater improvements from baseline in the PCS were observed with abatacept vs infliximab (difference of 1.93, 95% CI = 0.02, 3.84). Improvements in the MCS (difference of 1.92, 95% CI = –0.30, 4.15) and in all eight subscales were also numerically higher with abatacept vs infliximab (fig 4B).

placebo group. At day 365, greater improvements from baseline in the PCS were observed with abatacept vs infliximab (difference of 1.93, 95% CI = 0.02, 3.84). Improvements in the MCS (difference of 1.92, 95% CI = –0.30, 4.15) and in all eight subscales were also numerically higher with abatacept vs infliximab (fig 4B). Figure 4 Health-related quality of life. A. Mean change in physical and mental component summary (PCS and MCS, respectively) scores and the individual subscales of the Short Form-36 (SF-36) from day 1 to day 197. B. Mean change in PCS and MCS scores and the individual subscales of the SF-36 from day 1 to day 365; Data are presented for the intent-to-to treat population with a last observation carried forward analysis. Safety A summary of safety for all patients who received at least one dose of study medication is presented in table 2 (excluding the original placebo group between days 197–365).

the SF-36 from day 1 to day 365; Data are presented for the intent-to-to treat population with a last observation carried forward analysis. Safety A summary of safety for all patients who received at least one dose of study medication is presented in table 2 (excluding the original placebo group between days 197–365). Table 2 Summary of safety to day 197 and day 365 Days 1–197 Days 1–365 Abatacept + MTX (n = 156) Placebo + MTX (n = 110) Infliximab + MTX (n = 165) Abatacept + MTX (n = 156) Infliximab + MTX (n = 165) n (%)* n (%)* n (%)* n (%)* n (%)* Deaths 1 (0.6) 0 1 (0.6) 1 (0.6) 2 (1.2) SAEs 8 (5.1) 13 (11.8) 19 (11.5) 15 (9.6) 30 (18.2) Related SAEs 3 (1.9) 3 (2.7) 8 (4.8) 5 (3.2) 14 (8.5) Discontinuations due to SAEs 2 (1.3) 0 4 (2.4) 4 (2.6) 6 (3.6) AEs 129 (82.7) 92 (83.6) 140 (84.8) 139 (89.1) 154 (93.3) Related AEs 64 (41.0) 46 (41.8) 74 (44.8) 72 (46.2) 96 (58.2) Discontinuations due to AEs 3 (1.9) 1 (0.9) 8 (4.8) 5 (3.2) 12 (7.3) Serious infections 2 (1.3) 3 (2.7) 7 (4.2) 3 (1.9) 14 (8.5) Autoimmune symptoms and disorders 1 (0.6) 1 (0.9) 1 (0.6) 2 (1.3) 1 (0.6) Malignant neoplasms 1 (0.6) 1 (0.9) 2 (1.2) 1 (0.6) 2 (1.2) *More than one AE or SAE could be reported in each patient; percentages represent the proportion of patients who reported ⩾1 event. AE, adverse event; MTX, methotrexate; SAE, serious adverse event.

Table 2 Summary of safety to day 197 and day 365 Days 1–197 Days 1–365 Abatacept + MTX (n = 156) Placebo + MTX (n = 110) Infliximab + MTX (n = 165) Abatacept + MTX (n = 156) Infliximab + MTX (n = 165) n (%)* n (%)* n (%)* n (%)* n (%)* Deaths 1 (0.6) 0 1 (0.6) 1 (0.6) 2 (1.2) SAEs 8 (5.1) 13 (11.8) 19 (11.5) 15 (9.6) 30 (18.2) Related SAEs 3 (1.9) 3 (2.7) 8 (4.8) 5 (3.2) 14 (8.5) Discontinuations due to SAEs 2 (1.3) 0 4 (2.4) 4 (2.6) 6 (3.6) AEs 129 (82.7) 92 (83.6) 140 (84.8) 139 (89.1) 154 (93.3) Related AEs 64 (41.0) 46 (41.8) 74 (44.8) 72 (46.2) 96 (58.2) Discontinuations due to AEs 3 (1.9) 1 (0.9) 8 (4.8) 5 (3.2) 12 (7.3) Serious infections 2 (1.3) 3 (2.7) 7 (4.2) 3 (1.9) 14 (8.5) Autoimmune symptoms and disorders 1 (0.6) 1 (0.9) 1 (0.6) 2 (1.3) 1 (0.6) Malignant neoplasms 1 (0.6) 1 (0.9) 2 (1.2) 1 (0.6) 2 (1.2) *More than one AE or SAE could be reported in each patient; percentages represent the proportion of patients who reported ⩾1 event. AE, adverse event; MTX, methotrexate; SAE, serious adverse event. Summary of safety for days 1–197 During days 1–197, AEs were reported by a similar proportion of patients in the abatacept (82.7%), placebo (83.6%) and infliximab (84.8%) groups. Two deaths were reported: one patient in the abatacept group due to a cerebrovascular accident, and one patient in the infliximab group due to fibrosarcoma. Overall, the frequency of AEs, related AEs and discontinuations due to AEs or SAEs was similar for the abatacept and placebo groups. Between days 1–197, SAEs were lower with abatacept vs placebo (5.1 vs 11.8%), the difference was largely attributed to a higher frequency of gastrointestinal disorders, bacterial infections and musculoskeletal disorders in the placebo group. For the infliximab vs placebo groups during the same period, the frequency of AEs (84.8 vs 83.6%), related AEs (44.8 vs 41.8%) and SAEs (11.5 vs 11.8%) were similar; however, a higher proportion of patients in the infliximab group compared with the placebo group reported related SAEs (4.8 vs 2.7%), discontinued due to AEs (4.8 vs 0.9%), and discontinued due to SAEs (2.4 vs 0%). The higher frequency of SAEs in the infliximab vs placebo groups was largely due to an increase in serious infections (4.2 vs 2.7%, respectively).

in the infliximab group compared with the placebo group reported related SAEs (4.8 vs 2.7%), discontinued due to AEs (4.8 vs 0.9%), and discontinued due to SAEs (2.4 vs 0%). The higher frequency of SAEs in the infliximab vs placebo groups was largely due to an increase in serious infections (4.2 vs 2.7%, respectively). The most frequently reported AEs were primarily infections and infestations, and most were of mild to moderate intensity. Acute infusional AEs (within 3 h of the start of dosing) were reported in 5.1, 10.0 and 18.2% of the abatacept, placebo and infliximab groups, respectively (table 3). Infections and infestations were reported in 48.1, 51.8 and 52.1% of the abatacept, placebo and infliximab groups, respectively. Table 3 Frequently occurring acute infusional events (⩾2.0% of patients in any group) to day 197 and day 365 Acute infusional events Days 1–197 Days 1–365 Abatacept + MTX (n = 156) Placebo + MTX (n = 110) Infliximab + MTX (n = 165) Abatacept + MTX (n = 156) Infliximab + MTX (n = 165) n (%) n (%) n (%) n (%) n (%) Total patients with AEs 8 (5.1) 11 (10.0) 30 (18.2) 11 (7.1) 41 (24.8) Hypotension 0 0 7 (4.2) 0 8 (4.8) Headache 2 (1.3) 2 (1.8) 7 (4.2) 2 (1.3) 7 (4.2) Nausea 2 (1.3) 1 (0.9) 6 (3.6) 3 (1.9) 7 (4.2) Flushing 1 (0.6) 0 4 (2.4) 1 (0.6) 5 (3.0) Dyspnea 0 0 4 (2.4) 0 5 (3.0) Urticaria 0 1 (0.9) 4 (2.4) 0 8 (4.8) Pruritus 0 0 2 (1.2) 0 5 (3.0) Dizziness 0 0 2 (1.2) 1 (0.6) 4 (2.4) AE, adverse event; MTX, methotrexate.

(4.8) Headache 2 (1.3) 2 (1.8) 7 (4.2) 2 (1.3) 7 (4.2) Nausea 2 (1.3) 1 (0.9) 6 (3.6) 3 (1.9) 7 (4.2) Flushing 1 (0.6) 0 4 (2.4) 1 (0.6) 5 (3.0) Dyspnea 0 0 4 (2.4) 0 5 (3.0) Urticaria 0 1 (0.9) 4 (2.4) 0 8 (4.8) Pruritus 0 0 2 (1.2) 0 5 (3.0) Dizziness 0 0 2 (1.2) 1 (0.6) 4 (2.4) AE, adverse event; MTX, methotrexate. Serious infections were lower in the abatacept group (1.3%) than in the placebo (2.7%) and infliximab (4.2%) groups at day 197 (table 4), with patients originating from Latin America (abatacept, 8.3%; placebo, 16.7%; infliximab, 25.0%) and Europe (abatacept, 8.3%; placebo, 8.3%; infliximab, 33.3%). Between days 1–197, two opportunistic infections occurred (a pseudomonal lung infection and a Pneumocysitis jiroveci pneumonia) in the infliximab group.

e 4), with patients originating from Latin America (abatacept, 8.3%; placebo, 16.7%; infliximab, 25.0%) and Europe (abatacept, 8.3%; placebo, 8.3%; infliximab, 33.3%). Between days 1–197, two opportunistic infections occurred (a pseudomonal lung infection and a Pneumocysitis jiroveci pneumonia) in the infliximab group. Table 4 Serious infectious events to day 197 and day 365 Days 1–197 Days 1–365 Abatacept + MTX (n = 156) Placebo + MTX (n = 110) Inflixima + MTX (n = 165) Abatacept + MTX (n = 156) Infliximab + MTX (n = 165) n (%) n (%) n (%) n (%) n (%) Total patients with SAEs 8 (5.1) 13 (11.8) 19 (11.5) 15 (9.6) 30 (18.2) Infections and infestations 2 (1.3) 3 (2.7) 7 (4.2) 3 (1.9) 14 (8.5) Pneumonia 2 (1.3) 0 2 (1.2) 2 (1.3) 3 (1.8) Sinusitis 1 (0.6) 0 0 1 (0.6) 0 Postoperative wound infection 0 1 (0.9) 0 0 1 (0.6) Soft tissue abscess 0 1 (0.9) 0 0 0 Infective bursitis 0 1 (0.9) 0 0 0 Bronchitis 0 0 1 (0.6) 0 1 (0.6) Cellulitis 0 0 1 (0.6) 0 1 (0.6) Gastroenteritis 0 0 1 (0.6) 0 1 (0.6) Herpes zoster 0 0 1 (0.6) 0 1 (0.6) Lung infection pseudomonal 0 0 1 (0.6) 0 1 (0.6) Pneumocystis jiroveci pneumonia 0 0 1 (0.6) 0 1 (0.6) Infection skin ulcer 0 0 0 1 (0.6) 0 Encephalitis herpetic 0 0 0 0 1 (0.6) Erysipelas 0 0 0 0 1 (0.6) Lobar pneumonia 0 0 0 0 1 (0.6) Peritoneal tuberculosis 0 0 0 0 1 (0.6) Pulmonary tuberculosis 0 0 0 0 1 (0.6) Septic shock 0 0 0 0 1 (0.6) MTX, methotrexate; SAE, serious adverse event.

pneumonia 0 0 1 (0.6) 0 1 (0.6) Infection skin ulcer 0 0 0 1 (0.6) 0 Encephalitis herpetic 0 0 0 0 1 (0.6) Erysipelas 0 0 0 0 1 (0.6) Lobar pneumonia 0 0 0 0 1 (0.6) Peritoneal tuberculosis 0 0 0 0 1 (0.6) Pulmonary tuberculosis 0 0 0 0 1 (0.6) Septic shock 0 0 0 0 1 (0.6) MTX, methotrexate; SAE, serious adverse event. Autoimmune symptoms or disorders were uncommon (<1%), occurring at a similar frequency across groups (table 2). Malignancies were reported for four patients, including one abatacept-treated patient (bladder cancer), one placebo-treated patient (non-melanomatous skin cancer) and two infliximab-treated patients (malignant anorectal neoplasm and fibrosarcoma in one patient for each).

at a similar frequency across groups (table 2). Malignancies were reported for four patients, including one abatacept-treated patient (bladder cancer), one placebo-treated patient (non-melanomatous skin cancer) and two infliximab-treated patients (malignant anorectal neoplasm and fibrosarcoma in one patient for each). Summary of safety for days 1–365 for abatacept and infliximab groups During the entire 12-month, double-blind period (days 1–365), SAEs, related SAEs, and discontinuations due to SAEs were lower with abatacept than infliximab (SAEs: 9.6 vs 18.2%; related SAEs: 3.2 vs 8.5%; and discontinuations due to SAEs: 2.6 vs 3.6%, respectively (table 2)). Overall, AEs, related AEs and discontinuations due to AEs were also lower with abatacept than with infliximab (AEs: 89.1 vs 93.3%; related AEs: 46.2 vs 58.2%: and discontinuations due to AEs: 3.2 vs 7.3%, respectively). An additional infliximab-treated patient with peritoneal TB died during the second 6 months of the trial due to septic shock following surgery. One patient who was randomised to the placebo group died while receiving abatacept between days 198–365 from pneumonia and sepsis. The investigator assessment deemed the death possibly related to study treatment. Acute infusional events (7.1 vs 24.8%; table 3) were lower with abatacept vs infliximab, respectively.

Summary of safety for days 1–365 for abatacept and infliximab groups During the entire 12-month, double-blind period (days 1–365), SAEs, related SAEs, and discontinuations due to SAEs were lower with abatacept than infliximab (SAEs: 9.6 vs 18.2%; related SAEs: 3.2 vs 8.5%; and discontinuations due to SAEs: 2.6 vs 3.6%, respectively (table 2)). Overall, AEs, related AEs and discontinuations due to AEs were also lower with abatacept than with infliximab (AEs: 89.1 vs 93.3%; related AEs: 46.2 vs 58.2%: and discontinuations due to AEs: 3.2 vs 7.3%, respectively). An additional infliximab-treated patient with peritoneal TB died during the second 6 months of the trial due to septic shock following surgery. One patient who was randomised to the placebo group died while receiving abatacept between days 198–365 from pneumonia and sepsis. The investigator assessment deemed the death possibly related to study treatment. Acute infusional events (7.1 vs 24.8%; table 3) were lower with abatacept vs infliximab, respectively. Up to day 365, infections and infestations were reported in 59.6 and 68.5% of the abatacept and infliximab groups, respectively. Serious infections were reported in 1.9% of abatacept-treated patients and 8.5% of infliximab-treated patients (table 4). A total of five serious opportunistic infections were reported, all occurring in the infliximab group in one patient for each (encephalitis herpetic, lung infection pseudomonal, peritoneal TB, P jiroveci pneumonia and pulmonary TB). Both cases of TB were in patients who were PPD test negative and chest x ray negative at study entry. Between days 1–365, autoimmune symptoms or disorders were uncommon (<1%) and occurred at a similar frequency in the abatacept and infliximab groups (table 2). No additional malignant neoplasms were reported in either group between days 198–365.

atients who were PPD test negative and chest x ray negative at study entry. Between days 1–365, autoimmune symptoms or disorders were uncommon (<1%) and occurred at a similar frequency in the abatacept and infliximab groups (table 2). No additional malignant neoplasms were reported in either group between days 198–365. DISCUSSION The multi-centre, double-blind, placebo-controlled ATTEST trial, examining the relative efficacy and safety of abatacept or infliximab vs placebo, confirms that both biologics are effective for the treatment of RA. Over the 12-month study, a relative difference in safety was observed, with fewer SAEs, serious infections, acute infusional events and discontinuations due to AEs in the abatacept group than the infliximab group. Following 6 months of treatment, abatacept and infliximab on a background of MTX significantly reduced the signs and symptoms of disease (DAS28 (ESR); ACR 20, 50 and 70), and improved physical function (HAQ-DI) and quality of life (SF-36) compared with placebo, in patients with an inadequate response to MTX. The relative efficacy of abatacept and infliximab at day 197 was similar, as the 95% CIs for the treatment difference for DAS28 (ESR) scores, ACR response rates, HAQ-DI responses and HRQoL improvements overlapped. However, by day 365, the 95% CIs for the treatment difference between abatacept and infliximab for the reduction in DAS28 (ESR), good EULAR responses, LDAS, the ACR 20 response rate and the HRQoL physical summary measure did not include zero, suggesting a difference favouring abatacept for these specific endpoints.

ed. However, by day 365, the 95% CIs for the treatment difference between abatacept and infliximab for the reduction in DAS28 (ESR), good EULAR responses, LDAS, the ACR 20 response rate and the HRQoL physical summary measure did not include zero, suggesting a difference favouring abatacept for these specific endpoints. While the onset of response (as assessed by ACR 20 responses), initially appeared more rapid with infliximab, similar response rates were noted with abatacept and infliximab by day 85. Between 6 months and 1 year, abatacept responses were maintained, while those with infliximab were not. The sustained efficacy observed at 1 year compared with 6 months in abatacept-treated patients is consistent with results from other abatacept trials.2 3 Similarly, the finding that patients treated with infliximab did not sustain response rates over 1 year is also consistent with previous trials.11 When the impact of adding additional therapies was assessed using a sensitivity analysis according to the last score prior to addition, reductions in disease activity tended to be consistently higher with abatacept than with infliximab.

d not sustain response rates over 1 year is also consistent with previous trials.11 When the impact of adding additional therapies was assessed using a sensitivity analysis according to the last score prior to addition, reductions in disease activity tended to be consistently higher with abatacept than with infliximab. Overall, abatacept had a relatively more acceptable safety and tolerability profile than infliximab. Over 1 year, fewer SAEs, AEs, infections, and discontinuations due to AEs or SAEs were observed with abatacept relative to infliximab. Generally, opportunistic infections appeared to be relatively uncommon in this study and were mainly observed in the infliximab group, including two events of TB, an event of P jiroveci pneumonia, an event of pseudomonal lung infection and an event of herpes encephalitis. No opportunistic infections were reported in the abatacept group. Autoimmune symptoms or disorders were uncommon (<1%) with abatacept and infliximab. All of the malignancies were reported during the first 6 months, including one in the abatacept group, one in the placebo group and two in the infliximab group.

phalitis. No opportunistic infections were reported in the abatacept group. Autoimmune symptoms or disorders were uncommon (<1%) with abatacept and infliximab. All of the malignancies were reported during the first 6 months, including one in the abatacept group, one in the placebo group and two in the infliximab group. The data presented here should be interpreted within the context of several limitations. Although collectively the data support a relatively more acceptable risk-to-benefit profile of abatacept compared with infliximab, the design of this study utilised the infliximab dose of 3 mg/kg, without dose adjustment. At the time of the study, the recommended dose of infliximab (approved labelled starting dose for RA) was 3 mg/kg; however, today regulatory agencies recognise the use of higher doses of infliximab, and physicians use them in a proportion of patients to maintain a durable response. In fact, the dose may be increased in up to 30% of patients.12 Although higher doses have been associated with an increased risk of AEs in clinical trials.5 13 Finally, this study was not designed to evaluate the effect of abatacept or infliximab vs placebo on radiographic progression.

s to maintain a durable response. In fact, the dose may be increased in up to 30% of patients.12 Although higher doses have been associated with an increased risk of AEs in clinical trials.5 13 Finally, this study was not designed to evaluate the effect of abatacept or infliximab vs placebo on radiographic progression. In conclusion, these two biologic therapies with two distinct mechanisms of action have different safety and efficacy profiles; however, abatacept and infliximab both offer clinical improvements to patients with an inadequate response to MTX. Over 1 year, abatacept exhibited a durable response, and had a relatively more acceptable safety and tolerability profile than infliximab, as evidenced by fewer SAEs, serious infections, acute infusional events and discontinuations due to AEs. This study was funded and sponsored by Bristol-Myers Squibb, Princeton, New Jersey, USA. The authors would like to thank Laura Gardiner, Medicus International, for her editorial assistance. Editorial support was funded by Bristol-Myers Squibb, Princeton, New Jersey, USA. All authors were involved in the drafting and critical revision of the article; in addition, all authors have seen and approved the final article for submission Funding: This study is based upon clinical trial results from a study sponsored by Bristol-Myers Squibb, Princeton, New Jersey, USA. The authors actively participated in the conduct of these trials, and had full access to the raw data and responsibility for the analysis and interpretation.

This study was funded and sponsored by Bristol-Myers Squibb, Princeton, New Jersey, USA. The authors would like to thank Laura Gardiner, Medicus International, for her editorial assistance. Editorial support was funded by Bristol-Myers Squibb, Princeton, New Jersey, USA. All authors were involved in the drafting and critical revision of the article; in addition, all authors have seen and approved the final article for submission Funding: This study is based upon clinical trial results from a study sponsored by Bristol-Myers Squibb, Princeton, New Jersey, USA. The authors actively participated in the conduct of these trials, and had full access to the raw data and responsibility for the analysis and interpretation. Competing interests: RA is an employee of Bristol-Myers Squibb and owns stocks. JCB is an employee of Bristol-Myers Squibb, with stock options and restricted shares. PLNC is an employee of Bristol-Myers Squibb and owns shares. MD has received research grants, consulting fees and been on the speakers’ bureau for Bristol-Myers Squibb, Abbott and Wyeth, and has also received research grants and consulting fees from Centocor and Schering Plough. MK has received research grants and consulting fees from, and is an Advisory Board Member for, Bristol-Myers Squibb. TL is an employee of Bristol-Myers Squibb and owns stocks and shares. CL is an employee of Bristol-Myers Squibb and owns stocks and shares. MS has received research grants and consulting fees from Amgen, Bristol-Myers Squibb, Centocor and Wyeth.

Tumour necrosis factor α (TNFα) blocking agents as treatment for rheumatoid arthritis (RA) were developed based on evidence that the pro-inflammatory cytokine TNFα plays an important role in the pathogenesis.1 Some patients however do not clinically respond to TNFα blockade. At present no factors have been identified that fully explain or predict the differential response. One explanation for the heterogeneous clinical response may be found in the baseline variability in TNFα expression among individual patients.2 3 Genetic studies have suggested that individuals predisposed to high TNF production could show worse responses to anti-TNFα therapy.4 5 By contrast, a recent study using an in vitro bioassay suggested that good responsiveness to anti-TNF therapy is associated with significantly higher TNFα bioactivity at baseline compared to non-responding patients.6 Taken together, it remains to be determined which baseline cytokine profile distinguishes responding from non-responding patients in vivo. Another explanation for the diversity in response may be that inflammatory mediators other than TNFα drive different pathogenetic subsets of RA.

ompared to non-responding patients.6 Taken together, it remains to be determined which baseline cytokine profile distinguishes responding from non-responding patients in vivo. Another explanation for the diversity in response may be that inflammatory mediators other than TNFα drive different pathogenetic subsets of RA. We hypothesised that the pretreatment TNFα level in the synovium might be related to clinical efficacy, where TNFα blocking therapy could be most effective in patients with high pretreatment TNFα levels, as previously suggested in a small pilot study.7 In a prospective study we obtained arthroscopic synovial tissue samples from 143 patients with RA prior to initiation of infliximab therapy. We examined the cell infiltrate as well as the expression of cytokines, adhesion molecules and growth factors to identify predictors of clinical response.

in a small pilot study.7 In a prospective study we obtained arthroscopic synovial tissue samples from 143 patients with RA prior to initiation of infliximab therapy. We examined the cell infiltrate as well as the expression of cytokines, adhesion molecules and growth factors to identify predictors of clinical response. PATIENTS AND METHODS Patients Consecutive patients with RA according to the American College of Rheumatology (ACR) criteria were enrolled in the study. All failed at least two disease-modifying antirheumatic drugs (DMARDs) including methotrexate (MTX) and had a 28-joint Disease Activity Score (DAS28) of ⩾3.2 when included in the study. Patients were on stable maximal tolerable MTX treatment (5–30 mg/week). Oral corticosteroids (⩽10 mg/day) and non-steroidal anti-inflammatory drug (NSAIDs) were allowed if stable for at least 1 month prior to baseline. Concomitant medication was kept stable throughout the study. Previous use of a TNF blocking agent was an exclusion criterion. The Medical Ethics Committee of the Academic Medical Center, University of Amsterdam approved the protocol. All patients gave written informed consent. Treatment and evaluation of clinical response All patients were treated with infliximab according to the label for RA in a dosage of 3 mg/kg intravenously at baseline, week 2, week 6 and subsequently every 8 weeks. The DAS28 was evaluated at baseline and weeks 4, 8, 12 and 16 by specially trained research nurses.

PATIENTS AND METHODS Patients Consecutive patients with RA according to the American College of Rheumatology (ACR) criteria were enrolled in the study. All failed at least two disease-modifying antirheumatic drugs (DMARDs) including methotrexate (MTX) and had a 28-joint Disease Activity Score (DAS28) of ⩾3.2 when included in the study. Patients were on stable maximal tolerable MTX treatment (5–30 mg/week). Oral corticosteroids (⩽10 mg/day) and non-steroidal anti-inflammatory drug (NSAIDs) were allowed if stable for at least 1 month prior to baseline. Concomitant medication was kept stable throughout the study. Previous use of a TNF blocking agent was an exclusion criterion. The Medical Ethics Committee of the Academic Medical Center, University of Amsterdam approved the protocol. All patients gave written informed consent. Treatment and evaluation of clinical response All patients were treated with infliximab according to the label for RA in a dosage of 3 mg/kg intravenously at baseline, week 2, week 6 and subsequently every 8 weeks. The DAS28 was evaluated at baseline and weeks 4, 8, 12 and 16 by specially trained research nurses. For the analysis the absolute change in DAS28 (ΔDAS28) at week 16 was dichotomised and defined as non-response (ΔDAS28 <1.2) vs response (ΔDAS28 ⩾1.2). The dichotomy of the ΔDAS28 (on average comparable with a 20% improvement in DAS28) was chosen because it is applied in daily clinical practice and required for prolongation of reimbursement for TNFα blocking therapy by insurance companies in The Netherlands. Response was evaluated at 16 weeks because a significant improvement is expected to occur within 3 to 4 months, after which alternative treatment should be considered.8

applied in daily clinical practice and required for prolongation of reimbursement for TNFα blocking therapy by insurance companies in The Netherlands. Response was evaluated at 16 weeks because a significant improvement is expected to occur within 3 to 4 months, after which alternative treatment should be considered.8 Arthroscopy and synovial biopsy Before the first infliximab infusion patients underwent a mini-arthroscopy under local anaesthesia to obtain synovial tissue samples from an actively inflamed knee, ankle, wrist or metacarpophalangeal joint.9 Biopsies were taken with 2 mm forceps (Storz, Tuttlingen, Germany) from six or more sites within the joint to minimise sampling error. Biopsies were immediately snap frozen en bloc in Tissue Tek OCT (Miles, Elkhart, Indiana, USA) after collection. Sections of 5 μm were cut in a cryostat and mounted on Star Frost adhesive glass slides (Knittelgläser, Braunschweig, Germany). Slides were stored at −80°C until immunohistochemical staining.

sampling error. Biopsies were immediately snap frozen en bloc in Tissue Tek OCT (Miles, Elkhart, Indiana, USA) after collection. Sections of 5 μm were cut in a cryostat and mounted on Star Frost adhesive glass slides (Knittelgläser, Braunschweig, Germany). Slides were stored at −80°C until immunohistochemical staining. Immunohistochemical analysis Synovial sections were stained using the following monoclonal antibodies to analyse the infiltrate: anti-CD55 (67:Serotec, Oxford, UK) to detect fibroblast-like synoviocytes (FLS), anti-CD68 (EBM11: DAKO, Glostrup, Denmark) to detect macrophages, anti-CD3 (SK7, Becton Dickinson (BD), California, USA) for T cells, anti-CD22 (CLB-B-ly/1,6B11, The Netherlands) for B cells and anti-CD38 (HB7, BD) for plasma cells. CD163 (Ber-MAC3; DAKO) was stained to detect resident tissue macrophages; infiltrating macrophages were evaluated by detection of myeloid related protein (MRP)8 (8-5c2, BMA Biomedicals, Augst, Switzerland) and MRP14 (S36.48, BMA Biomedicals). For the detection of cytokines, adhesion molecules and growth factors we used anti-human TNFα (52B83; Monosan, Brussels, Belgium), anti-interleukin (IL)6 (Nephrology Department, Leiden University Medical Center, Leiden, The Netherlands), anti-IL10 (23738.111, R&D, Abingdon, UK), anti-IL18 (2d3b6, MD Biosciences, Minnesota, USA), anti-IL1β (2D8, ImmunoKontact, Oxford, UK), anti-intercellular adhesion molecule (ICAM)-1 (MEM111, Sanbio, Erembodegem, Belgium), anti-vascular cell adhesion molecule (VCAM) (1G11B1, Sanbio), anti-E-selectin (BBIG-E4, R&D), anti- vascular endothelial growth factor (VEGF) (N.010313 C-1, Santa Cruz, California, USA) and anti-basic fibroblast growth factor (bFGF) (F14220, BD). Staining of cellular markers was performed using a three-step immunoperoxidase method.2 For staining of cytokines, adhesion molecules and growth factors, biotinylated tyramine was used as amplification. For control sections the primary antibody was omitted or irrelevant immunoglobulins were applied.

bFGF) (F14220, BD). Staining of cellular markers was performed using a three-step immunoperoxidase method.2 For staining of cytokines, adhesion molecules and growth factors, biotinylated tyramine was used as amplification. For control sections the primary antibody was omitted or irrelevant immunoglobulins were applied. Digital image analysis All sections were analysed at random by trained analysts blinded for clinical outcome. Images were acquired and analysed by computer-assisted image analysis using a Syndia algorithm on a Qwin-based analysis system (Leica, Cambridge, UK).10 A total of 18 high-power fields per marker were analysed. Cellular markers were expressed as positive cells/mm2 (counts/mm2). Staining of cytokines, adhesion molecules and growth factors was expressed as integrated optical density/mm2 (IOD/mm2). CD68+ macrophages and TNFα expression were analysed separately in the intimal lining layer and the sublining.

marker were analysed. Cellular markers were expressed as positive cells/mm2 (counts/mm2). Staining of cytokines, adhesion molecules and growth factors was expressed as integrated optical density/mm2 (IOD/mm2). CD68+ macrophages and TNFα expression were analysed separately in the intimal lining layer and the sublining. Statistical analysis Independent Student t tests or Mann–Whitney U tests were used to detect significant differences in baseline parameters between responders and non-responders. The χ2 test was employed to compare the percentage of erosive, IgM-rheumatoid factor (RF) and anti-cyclic citrullinated peptide (CCP) positive patients in the responder and non-responder groups. Variables with a p value <0.1 in the univariable analysis were selected as possible predictors in a stepwise backward multivariable logistic regression model. Several exchangeable prediction sets were constructed based on the correlation structure among the potential predictors with strongly correlating predictors in different sets. The model with the highest explained variance according to Nagelkerke is reported. Correlations were assessed with the Pearson product-moment or Spearman rank-order correlation coefficients, whichever was appropriate. SPSS V.11.1.4 (Chicago, Illinois, USA) was used. Because the primary objective of this study was to find a pathogenetic mechanism in synovial tissue relating to clinical response, a per-protocol analysis was performed.

he Pearson product-moment or Spearman rank-order correlation coefficients, whichever was appropriate. SPSS V.11.1.4 (Chicago, Illinois, USA) was used. Because the primary objective of this study was to find a pathogenetic mechanism in synovial tissue relating to clinical response, a per-protocol analysis was performed. RESULTS Enrolment, discontinuation and exclusion A total of 143 patients were enrolled. The population was predominantly female (74%), RF positive (71%) with a mean (SD) age of 55 (13) years and mean disease duration of 125 (108) months. Erosions were present in 110 (77%) patients. The mean DAS28 score was 5.9 (1.0) at baseline and the mean MTX dose was 18.2 (8.5) mg/week. In three patients, a flare of arthritis at week 6 (n = 1) and week 12 (n = 2) was the reason for discontinuation. These patients were analysed as non-responders. Patients (n = 18) who violated the protocol or discontinued infliximab therapy (for reasons other than non-response) were excluded. Synovial tissue exclusion Following strict quality control, 23 synovial biopsies were excluded by a blinded assessor for absence of an intimal lining layer. After excluding 18 patients from the response analysis and 23 synovial biopsies from the tissue analysis a total of 103 completely evaluable patients remained for the analysis of synovial tissue in combination with clinical response.

l biopsies were excluded by a blinded assessor for absence of an intimal lining layer. After excluding 18 patients from the response analysis and 23 synovial biopsies from the tissue analysis a total of 103 completely evaluable patients remained for the analysis of synovial tissue in combination with clinical response. Baseline patient characteristics The remaining group of 103 patients who were analysed for response in combination with synovial tissue consisted of 71 females and 32 males, of whom 75% were rheumatoid factor positive. The mean (SD) age was 55 (13) years and the mean disease duration was 125 (110) months. Erosions were present in 79 (77%) patients. The mean DAS28 score was 5.9 (1.1) at baseline and the mean MTX dose was 18.8 (8.5) mg/week. Corticosteroids were used by 28 (27%) patients with a mean dose of 8.0 (2.9) mg/day. On average, patients had failed treatment with 2.2 DMARDs before inclusion in the study. No notable changes occurred in the baseline characteristics of the analysed 103 patients compared to the baseline characteristics of all 143 enrolled patients (see table 1).

y 28 (27%) patients with a mean dose of 8.0 (2.9) mg/day. On average, patients had failed treatment with 2.2 DMARDs before inclusion in the study. No notable changes occurred in the baseline characteristics of the analysed 103 patients compared to the baseline characteristics of all 143 enrolled patients (see table 1). Table 1 Baseline patient characteristics Analysed patients (n = 103) Responders (n = 70) Non-responders (n = 33) p Value Demographics: Age (years) 55 (13) 54 (13) 56 (12) 0.40 Female (%) 71 (69) 51 (73) 20 (61) 0.21 Disease status: Disease duration (months) 125 (110) 123 (111) 130 (110) 0.80 Erosive disease (%) 79 (77) 57 (81) 22 (67) 0.10 RF positive (%) 77 (75) 57 (81) 20 (61) 0.02 Anti-CCP positive (%) 78 (76) 57 (81) 13 (39) 0.08 DAS28 5.9 (1.1) 6.0 (1.0) 5.6 (1.2) 0.07 Patients global score (0–100 mm) 60 (22) 62 (21) 56 (23) 0.23 ESR (mm/h) 34 (24) 36 (23) 29 (25) 0.16 CRP (mg/dl) 22 (28) 24 (26) 19 (30) 0.36 Drug treatments: Previous DMARDs 2.2 (1.5) 2.1 (1.4) 2.4 (1.6) 0.42 Methotrexate (mg/week) 18.8 (8.5) 19.5 (8.2) 17.1 (8.9) 0.18 Receiving corticosteroids (%) 28 (27) 22 (31) 6 (18) 0.22 Receiving NSAIDs (%) 52 (50) 35 (50) 17 (52) 0.89 Mean (SD), median and interquartile range (IQR) or percentages are shown. p Values <0.05 (two-sided) were considered significant. CCP, cyclic citrullinated peptide; CRP, C-reactive protein; DAS28, 28-joint Disease Activity Score; DMARD, disease-modifying antirheumatic drug; ESR, erythrocyte sedimentation rate; NSAID, non-steroidal anti-inflammatory drug; RF, rheumatoid factor.

Figure 2 Changes in spine and hip bone mineral density over 4 years in postmenopausal women with osteoporosis aged 50–65 years receiving either placebo or strontium ranelate. Over the 4-year follow-up, clinical, serious, drug-related adverse effects were similar in the placebo and treated groups (no clinical significance between groups regarding the incidence of nausea, diarrhoea, headache, dermatitis, eczema, venous thromboembolism events) and the overall safety profile was very similar to that already described for the whole SOTI study population over 3 and 4 years. No case of pulmonary embolism or hypersensitivity reaction was observed in this study population. DISCUSSION This study shows that strontium ranelate (2 g/day) decreases the risk of vertebral fractures over 4 years in women aged 50–65 years with severe osteoporosis. The fracture incidence increases with age, but young postmenopausal women experience fractures. In the National Osteoporosis Risk Assessment (NORA) study, 37% of the fractures occurred in women 50–64 years of age, and this population of young postmenopausal women accounted for one-fifth of the hip fractures that occurred within 1 year of baseline.9 A prevalent fracture may be a much more severe predictor of subsequent fractures than other well-known risk factors when considering younger populations at risk due to the fact that the prevalence of fractures largely increases with age. Such a prevalent fracture is more common in communities aged 70 or 80 years than in younger populations, giving less weight in terms of risk to this factor in the elderly.10 In addition, it has been shown in a general semi-urban population that the presence of vertebral fractures led to a decrease in the expected remaining life years, the decrease being greater in the younger than in the older age groups.11 This highlights the importance of diagnosing and treating appropriately as soon as possible this youngest postmenopausal population at high further risk.

Table 1 Baseline patient characteristics Analysed patients (n = 103) Responders (n = 70) Non-responders (n = 33) p Value Demographics: Age (years) 55 (13) 54 (13) 56 (12) 0.40 Female (%) 71 (69) 51 (73) 20 (61) 0.21 Disease status: Disease duration (months) 125 (110) 123 (111) 130 (110) 0.80 Erosive disease (%) 79 (77) 57 (81) 22 (67) 0.10 RF positive (%) 77 (75) 57 (81) 20 (61) 0.02 Anti-CCP positive (%) 78 (76) 57 (81) 13 (39) 0.08 DAS28 5.9 (1.1) 6.0 (1.0) 5.6 (1.2) 0.07 Patients global score (0–100 mm) 60 (22) 62 (21) 56 (23) 0.23 ESR (mm/h) 34 (24) 36 (23) 29 (25) 0.16 CRP (mg/dl) 22 (28) 24 (26) 19 (30) 0.36 Drug treatments: Previous DMARDs 2.2 (1.5) 2.1 (1.4) 2.4 (1.6) 0.42 Methotrexate (mg/week) 18.8 (8.5) 19.5 (8.2) 17.1 (8.9) 0.18 Receiving corticosteroids (%) 28 (27) 22 (31) 6 (18) 0.22 Receiving NSAIDs (%) 52 (50) 35 (50) 17 (52) 0.89 Mean (SD), median and interquartile range (IQR) or percentages are shown. p Values <0.05 (two-sided) were considered significant. CCP, cyclic citrullinated peptide; CRP, C-reactive protein; DAS28, 28-joint Disease Activity Score; DMARD, disease-modifying antirheumatic drug; ESR, erythrocyte sedimentation rate; NSAID, non-steroidal anti-inflammatory drug; RF, rheumatoid factor. Clinical response Of 103 patients, 70 (68%) were DAS28 responders and 33 (32%) non-responders at week 16. The clinical efficacy matched the expected 60–70% responders observed in previous randomised controlled trials with infliximab in patients with RA. The mean change in DAS28 was 1.84 (1.26) at week 16. For non-responders the average ΔDAS28 was 0.45 (0.58) compared to 2.52 (0.89) for responders. This difference was highly significant (p<0.001).

tched the expected 60–70% responders observed in previous randomised controlled trials with infliximab in patients with RA. The mean change in DAS28 was 1.84 (1.26) at week 16. For non-responders the average ΔDAS28 was 0.45 (0.58) compared to 2.52 (0.89) for responders. This difference was highly significant (p<0.001). All baseline patient characteristics were also tested for differences between responders and non-responders. The DAS28 at baseline was nearly significant and was included in the prediction model (p = 0.07). The same holds for anti-CCP positivity (p = 0.08). Finally, there were significantly more rheumatoid factor positive patients in the responder compared to the non-responder group (p = 0.020), (see table 1). Baseline synovial TNFα expression and the number of major TNFα producing cells are associated with clinical response (ΔDAS28 ⩾1.2) after 16 weeks of TNFα blocking therapy Scatter diagrams of the synovial predictors in individual patients in both response groups are presented in fig 1. TNFα expression levels were higher in the synovial sublining of responders compared to non-responders (p = 0.008) (fig 1A). Similarly, TNFα expression in the intimal lining layer was higher in responders compared to non-responders (p = 0.047). Table 2 shows the associations of synovial markers and response.

d in fig 1. TNFα expression levels were higher in the synovial sublining of responders compared to non-responders (p = 0.008) (fig 1A). Similarly, TNFα expression in the intimal lining layer was higher in responders compared to non-responders (p = 0.047). Table 2 shows the associations of synovial markers and response. Figure 1 A. The median synovial sublining tumour necrosis factor (TNF)α expression was higher in responders compared to non-responders (p = 0.008). B. The median number of CD3+ T cells in responders vs non-responders (p = 0.001). (C) The mean number of CD68+ sublining macrophages was also higher in responders 576 (428), than in non-responders 387 (338) (p = 0.029). (D) CD163+ macrophages in responders vs non-responders (p = 0.017). Table 2 Associations of studied synovial predictors with clinical response ΔDAS28 score p Value Responders (n = 70) ⩾1.2 Non-responders (n = 33) <1.2 Cytokines (IOD/mm2): TNFα lining 48 214 (29 116–88 971) 36 376 (23 394–54 271) 0.047 TNFα sublining 77 947 (38 710–123 535) 46 033 (19 890–78 129) 0.008 IL1β 56 278 (31 063–81 655) 54 484 (32 098–129 041) 0.814 IL6 50 537 (41 939) 38 364 (28 527) 0.135 IL10 131 106 (87 359) 115 964 (83 953) 0.415 IL18 9530 (12 086) 7009 (7470) 0.289 Cellular markers (counts/mm2 ) CD55 652 (418) 694 (389) 0.895 CD3 149 (66–385) 47 (22–163) 0.001 CD68 lining 374 (253) 295 (184) 0.130 CD68 sublining 576 (428) 387 (338) 0.029 CD163 1100 (432) 878 (433) 0.017 CD22 54 (99) 47 (111) 0.756 CD38 284 (384) 325 (556) 0.661 MRP8 139 (35–294) 53 (19–135) 0.018 MRP14 159 (34–526) 51 (18–166) 0.024 Adhesion molecules (IOD/mm2) ICAM 30 790 (33 016) 22 321 (21 706) 0.195 VCAM 80 065 (52 630) 70 701 (44 574) 0.387 E-Selectin 37 651 (40 134) 30 488 (29 972) 0.373 Growth factors (IOD/mm2 ) VEGF 18 252 (78 689) 5776 (5372) 0.374 bFGF 209 (736) 64 (163) 0.276 *Data are presented as mean (SD) or median (interquartile range), whichever appropriate. p Values<0.05 (two-sided) were considered significant.

5 (52 630) 70 701 (44 574) 0.387 E-Selectin 37 651 (40 134) 30 488 (29 972) 0.373 Growth factors (IOD/mm2 ) VEGF 18 252 (78 689) 5776 (5372) 0.374 bFGF 209 (736) 64 (163) 0.276 *Data are presented as mean (SD) or median (interquartile range), whichever appropriate. p Values<0.05 (two-sided) were considered significant. bFGF, basic fibroblast growth factor; ICAM, intercellular adhesion molecule; IF, interferon; IOD, integrated optical density; MRP, myeloid related protein; TNF, tumour necrosis factor; VCAM, vascular cell adhesion molecule; VEGF, vascular endothelial growth factor. Because macrophages are known to be the main TNFα producing cells in the synovium, the number of macrophages was studied. The mean number of CD68+ sublining macrophages was significantly higher in responders than in non-responders (p = 0.029) (fig 1C). Similarly, the number of CD163+ resident tissue macrophages as well as the number of infiltrating MRP8+ and MRP14+ macrophages was higher in responders (p = 0.017, p = 0.018 and p = 0.024, respectively) (fig 1D). Finally, the median number of CD3+ T cells was higher in responders (p = 0.001) (fig 1B).

the presence of vertebral fractures led to a decrease in the expected remaining life years, the decrease being greater in the younger than in the older age groups.11 This highlights the importance of diagnosing and treating appropriately as soon as possible this youngest postmenopausal population at high further risk. In the present study, the incidence of vertebral and non-vertebral fractures over 4 years in the placebo group was high (32.8% and 14.6%, respectively). This demonstrates in this population aged less than 65 years that the presence of prevalent fractures leads to a dramatic increase in the overall fracture risk, not limited to vertebral fractures. Over the first year, the incidence of vertebral fractures was 8.3%, which reinforces the indication of a prompt and effective treatment in this population. As strontium is a heavier element than calcium, its presence in bone could lead to an overestimation of BMD measurement. However, a strong association has been demonstrated for strontium ranelate-treated patients between the increase in total hip and/or femoral neck BMD and the subsequent decrease in fracture incidence.12

0.029) (fig 1C). Similarly, the number of CD163+ resident tissue macrophages as well as the number of infiltrating MRP8+ and MRP14+ macrophages was higher in responders (p = 0.017, p = 0.018 and p = 0.024, respectively) (fig 1D). Finally, the median number of CD3+ T cells was higher in responders (p = 0.001) (fig 1B). Baseline TNFα is the only significant synovial predictor of response The relationship between TNFα expression and clinical response was confirmed by stepwise backward multivariable logistic regression analysis. Due to multicolinearity among predictors, 10 exchangeable bivariate prediction sets were constructed consisting of (1) TNFα expression in either the intimal lining layer or synovial sublining with (2) one out of five cellular markers. This revealed that TNFα expression in the synovial sublining was the only independent, early determinant of therapy response (p = 0.011), explaining just about 10% of the variance in response to therapy (R2 = 0.099, according to Nagelkerke, odds ratio (OR) = 1.013). Likewise, TNFα expression in the intimal lining layer rather than synovial sublining explained 9% of the variance in response to therapy (p = 0.020), (R2 = 0.089, OR = 1.017). Thus, after adjustment for TNFα in a bivariate logistic regression model including either the numbers of CD68+ sublining macrophages, CD163+, MPR8+, MRP14+ macrophages, or CD3+ T cells, these cellular markers were no longer significantly associated with response. The removal of these cellular markers from the model is consistent with the notion that these cells are the main producers of TNFα.

cluding either the numbers of CD68+ sublining macrophages, CD163+, MPR8+, MRP14+ macrophages, or CD3+ T cells, these cellular markers were no longer significantly associated with response. The removal of these cellular markers from the model is consistent with the notion that these cells are the main producers of TNFα. Contribution of disease activity at baseline to the prediction of response To assess the value of a combined prediction model consisting of clinical and synovial data we added the relevant clinical variables (DAS28 score, the presence of IgM-RF or anti-CCP antibodies) to the 10 bivariate synovial models. With stepwise backward multivariate logistic regression this resulted in a significant prediction model with TNFα expression in the synovial sublining (p = 0.008, OR = 1.014) and the DAS28 (p = 0.031, OR = 1.611) with an increased explained variance of 17% (R2 = 0.172). A similar analysis except with TNFα expression in the intimal lining layer (p = 0.023, OR = 1.017) resulted in a significant model including the DAS28 (p = 0.044, OR = 1.588) and the presence of anti-CCP antibodies (p = 0.046, OR = 3.038) (R2 = 0.182).

= 1.611) with an increased explained variance of 17% (R2 = 0.172). A similar analysis except with TNFα expression in the intimal lining layer (p = 0.023, OR = 1.017) resulted in a significant model including the DAS28 (p = 0.044, OR = 1.588) and the presence of anti-CCP antibodies (p = 0.046, OR = 3.038) (R2 = 0.182). Baseline IL1β, IL6, IL18 and IL10 expression is not associated with response to TNFα blocking therapy In addition to TNFα we studied the expression of other cytokines such as IL18, which is believed to operate mainly upstream of TNFα, while IL1β and IL6 appear to act mainly downstream of TNFα. Despite the known interplay of these cytokines with TNFα we found no relationship between synovial expression levels of either of these cytokines and clinical response. We did find, however, a significant correlation between the level of IL1β expression and TNFα expression (r = 0.306, p = 0.014). The expression level of IL10 was not different between response groups (see table 2).

nd no relationship between synovial expression levels of either of these cytokines and clinical response. We did find, however, a significant correlation between the level of IL1β expression and TNFα expression (r = 0.306, p = 0.014). The expression level of IL10 was not different between response groups (see table 2). Baseline expression of adhesion molecules and angiogenic factors is not associated with response to TNFα blocking therapy In accordance with their function in leukocyte recruitment a significant correlation was found between the number of sublining macrophages on the one hand and the expression of the vascular adhesion molecules E-selectin (Spearman rho correlation coefficient 0.275–0.536, p<0.01) on the other. However, the level at which the adhesion molecules were expressed in the synovium at baseline did not differ between responders and non-responders. Furthermore, VEGF is known to mediate angiogenesis and is regulated by pro-inflammatory cytokines such as TNFα. However, no association was found between VEGF expression and response (table 2). Similarly, no difference was found in the expression of bFGF.

t baseline did not differ between responders and non-responders. Furthermore, VEGF is known to mediate angiogenesis and is regulated by pro-inflammatory cytokines such as TNFα. However, no association was found between VEGF expression and response (table 2). Similarly, no difference was found in the expression of bFGF. DISCUSSION The primary objective of this study was to investigate whether immunohistological assessment of the cell infiltrate and cytokine expression in the synovium prior to initiation of TNFα blocking therapy could predict clinical response in patients with RA. We could confirm our hypothesis that the level of synovial TNFα expression is a significant early predictor of the response. This was shown by increased TNFα expression levels in the intimal lining layer and synovial sublining of responding compared to non-responding patients. In line with these findings, there was increased infiltration by macrophages, including CD163+ resident tissue macrophages and MRP8+ and MRP14+ infiltrating macrophages, as well as T cells in responders vs non-responders. It is important to note that these cells are the main source of TNFα in the synovium of patients with RA.11

ith these findings, there was increased infiltration by macrophages, including CD163+ resident tissue macrophages and MRP8+ and MRP14+ infiltrating macrophages, as well as T cells in responders vs non-responders. It is important to note that these cells are the main source of TNFα in the synovium of patients with RA.11 Consistent with the clinical experience that the response to TNFα blockade is not a dichotomous phenomenon,12 there was no distinct threshold value in TNFα expression in the synovium of patients with RA. Multivariate logistic regression analysis of synovial markers showed that TNFα expression in the sublining could explain about 10% of the variance in response to therapy. After adjusting for disease activity at baseline this further increased to 17%. Hence, the predictive value of synovial TNF expression is statistically significant, but overall limited. This clearly indicates that variables other than synovial TNFα expression are involved as well. The results presented here show for the first time that the expression of IL1β, IL6, IL18, IL10, E-selectin, ICAM-1, VCAM-1, VEGF and bFGF is not associated with clinical response to anti-TNFα treatment. Thus, the features of synovial inflammation at baseline in responders to anti-TNF therapy compared to non-responders do not merely represent a greater amount of inflammation, but the results presented here underscore the importance of specific inflammatory pathways. Future work should expand the search for other biomarkers and molecular networks to better understand the variable response to anti-TNFα therapy.

mpared to non-responders do not merely represent a greater amount of inflammation, but the results presented here underscore the importance of specific inflammatory pathways. Future work should expand the search for other biomarkers and molecular networks to better understand the variable response to anti-TNFα therapy. The search for predictors of response is important in the context of personalised medicine, which may be an effective approach to increase the percentage of patients exhibiting a robust response to a given treatment. Previous work has clearly suggested that RA consists of different pathogenetic subsets, leading to common signs and symptoms associated with what we at present define clinically as RA.2 3 13 Thus, it is conceivable that, for instance, TNFα expression is more important in some patients than in others, and that TNF blockade could be more effective in the former. Although we are as yet not able to predict the response sufficiently to select patients who are likely to have a beneficial response to a specific treatment before initiation of therapy and to guide treatment decisions, the results of the present study do provide proof of principle that this might be achieved by further optimisation of the biomarkers or perhaps combinations with other clinical and biological variables that need to be identified. The relevance of this approach is underscored by the expanding array of biological therapies and their costs.14

esent study do provide proof of principle that this might be achieved by further optimisation of the biomarkers or perhaps combinations with other clinical and biological variables that need to be identified. The relevance of this approach is underscored by the expanding array of biological therapies and their costs.14 There is no evidence that simple measurement of plasma TNFα levels can be used to predict clinical response to TNFα blockade. As it appears logical that patients producing high levels of TNFα at the site of inflammation are more likely to benefit from TNFα blockade than those with lower TNFα levels, we focused on the synovium as the primary target of RA. The association between pretreatment TNFα expression in the synovium and clinical response to 3 mg/kg of infliximab described here in 103 patients with RA confirms and extends a trend observed in a pilot study in patients with RA where 4 out of 8 patients who met the ACR50 response criteria at 2 weeks after initiation of infliximab (10 mg/kg) therapy were those with the highest levels of TNFα expression in the synovium, as shown by immunohistochemistry.7 Of interest, the findings are also in accordance with a recent study suggesting that clinical response at 52 weeks is associated with significantly higher circulating TNFα bioactivity (measured by an in vitro bioassay) at baseline in responding compared to non-responding patients.6 Together, these studies support the notion that the initial clinical response to TNFα blockade is related to pretreatment levels of TNFα production. In patients who initially respond, but loose response over time, other mechanisms such as formation of antibodies against the drug may be operative.15

esponding patients.6 Together, these studies support the notion that the initial clinical response to TNFα blockade is related to pretreatment levels of TNFα production. In patients who initially respond, but loose response over time, other mechanisms such as formation of antibodies against the drug may be operative.15 In conclusion, the results presented here show proof of principle that the heterogeneous response to TNFα blockade is associated with TNFα expression in the inflamed synovium. Future work should expand the search for other biomarkers and molecular networks as well as combinations with clinical variables. Thus, the prediction of how a patient will respond might come in reach of the treating doctor. The authors wish to thank Professor A H Zwinderman, head of the department of Department of Clinical Epidemiology, Biostatistics and Bioinformatics at the Academic Medical Center/University of Amsterdam, for reviewing the statistical analysis. We also wish to thank Desiree Pots for assisting with the immunohistochemical staining, and research nurses Nitolanda van Rijn and Natasja Cassin for performing clinical assessments. Funding: This study was funded by a Health Care Efficiency Research Programme grant from The Netherlands Organization for Health Research and Development (ZonMw) in assignment of The Netherlands Organization for Scientific Research (NWO) (grant number 945-02-029), the Dutch Arthritis Association and the European Community’s FP6 funding (Autocure). Competing interests: PPT has served as a consultant for Abbott, Amgen, Centocor and Wyeth.

Funding: This study was funded by a Health Care Efficiency Research Programme grant from The Netherlands Organization for Health Research and Development (ZonMw) in assignment of The Netherlands Organization for Scientific Research (NWO) (grant number 945-02-029), the Dutch Arthritis Association and the European Community’s FP6 funding (Autocure). Competing interests: PPT has served as a consultant for Abbott, Amgen, Centocor and Wyeth. Ethics approval: The Medical Ethics Committee of the Academic Medical Center, University of Amsterdam approved the protocol. All patients gave written informed consent. This publication reflects only the authors’ views. The European Community is not liable for any use that may be made of the information herein.

Vertebral fractures represent 27% of all osteoporotic fractures coming to clinical attention1 and their deleterious consequences on health are now well recognised. Clinical vertebral fractures increase mortality in the elderly.2 3 Prevalent vertebral fractures increase the risk of subsequent vertebral and non-vertebral fractures,3 4 as well as the risk of hip fracture with at least a twofold excess. The risk of further fracturing has been shown to be higher among younger people compared with the elderly.4 Few data are available in clinical trials in patients younger than 65 years. Strontium ranelate is an antiosteoporotic treatment that decreases the risk of vertebral5 and non-vertebral fractures,6 including the risk of hip fractures in a high-risk population.6 The aim of the present study was to assess the efficacy of strontium ranelate in patients with osteoporosis aged 50–65 years, most presenting with a prevalent vertebral fracture, a subgroup of patients having a very high lifetime risk of fractures.

tures,6 including the risk of hip fractures in a high-risk population.6 The aim of the present study was to assess the efficacy of strontium ranelate in patients with osteoporosis aged 50–65 years, most presenting with a prevalent vertebral fracture, a subgroup of patients having a very high lifetime risk of fractures. MATERIALS AND METHODS Study subjects Data from the Spinal Osteoporosis Therapeutic Intervention (SOTI) trial were used for this study. In this randomised, double blind, placebo controlled clinical trial, 1649 postmenopausal patients aged 50 years or more were enrolled with these inclusion criteria: at least one vertebral fracture and a lumbar spine bone mineral density (BMD) of 0.840 g/cm2 or less. The vertebral fracture incidence was the main efficacy criterion of the study. The assessment of this criterion was performed over 3 years (main statistical analysis) showing an early and sustained significant reduction in the vertebral fracture risk by 49% over the first year and by 41% over 3 years.5 Furthermore, a pre-planned analysis was performed over 4 years showing a 33% reduction (relative risk (RR) 0.67; 95% CI 0.55 to 0.81, p<0.001) in the risk of sustaining a vertebral fracture over 4 years (data on file). For this post-hoc analysis investigating, blinded to treatment, the antifracture efficacy of strontium ranelate in younger postmenopausal women with osteoporosis, we selected the subset of the SOTI study population aged 50–65 years.

MATERIALS AND METHODS Study subjects Data from the Spinal Osteoporosis Therapeutic Intervention (SOTI) trial were used for this study. In this randomised, double blind, placebo controlled clinical trial, 1649 postmenopausal patients aged 50 years or more were enrolled with these inclusion criteria: at least one vertebral fracture and a lumbar spine bone mineral density (BMD) of 0.840 g/cm2 or less. The vertebral fracture incidence was the main efficacy criterion of the study. The assessment of this criterion was performed over 3 years (main statistical analysis) showing an early and sustained significant reduction in the vertebral fracture risk by 49% over the first year and by 41% over 3 years.5 Furthermore, a pre-planned analysis was performed over 4 years showing a 33% reduction (relative risk (RR) 0.67; 95% CI 0.55 to 0.81, p<0.001) in the risk of sustaining a vertebral fracture over 4 years (data on file). For this post-hoc analysis investigating, blinded to treatment, the antifracture efficacy of strontium ranelate in younger postmenopausal women with osteoporosis, we selected the subset of the SOTI study population aged 50–65 years. Treatment regimens Patients were randomly assigned to receive 2 g per day of strontium ranelate or placebo. Throughout the study period, subjects received daily calcium and vitamin D supplements at lunchtime.

For this post-hoc analysis investigating, blinded to treatment, the antifracture efficacy of strontium ranelate in younger postmenopausal women with osteoporosis, we selected the subset of the SOTI study population aged 50–65 years. Treatment regimens Patients were randomly assigned to receive 2 g per day of strontium ranelate or placebo. Throughout the study period, subjects received daily calcium and vitamin D supplements at lunchtime. Assessment of outcomes Radiographs of the spine were obtained at baseline and annually and assessed centrally. A semiquantitative visual assessment of each vertebrae from T4 to L4 was performed by the same reader throughout the study, using the semiquantitative grading scale as previously described by Genant et al.7 BMD at the lumbar spine and proximal femur was measured by dual-energy x ray absorptiometry at baseline and at 6-month intervals. For patient diagnostic categorisation, lumbar spine and femoral neck BMD T-scores were calculated using a reference population previously described.8 Femoral neck BMD T-scores were also re-calculated using the National Health and Nutrition Examination Survey III (NHANES III) reference. Statistical analysis The primary endpoint of this analysis was the incidence in patients experiencing a new vertebral fracture, estimated according to the Kaplan–Meier method, and an adjusted Cox model was used to compare groups and to estimate the RR of experiencing a new vertebral fracture and its 95% CI. Secondary endpoints included the incidence of non-vertebral fractures and spine and hip BMD changes.

experiencing a new vertebral fracture, estimated according to the Kaplan–Meier method, and an adjusted Cox model was used to compare groups and to estimate the RR of experiencing a new vertebral fracture and its 95% CI. Secondary endpoints included the incidence of non-vertebral fractures and spine and hip BMD changes. The adjusted Cox model considered the treatment effects and the following covariates: country, L2/L4 BMD at baseline and prevalent vertebral fractures. RESULTS Among the patients included in the SOTI study, 385 were aged 50–65 years, of whom 353 were eligible for the assessment of the efficacy of strontium ranelate on vertebral fractures according to the intention-to-treat principle (at least one sachet intake, at least one assessable vertebral x ray at baseline and one post-baseline). Baseline characteristics of this population are presented in table 1. There was no statistically significant difference between the patients given 2 g/day strontium ranelate (N  =  168) and those given placebo (N  =  185). Table 1 Baseline characteristics of patients Placebo Strontium ranelate p Value* N  =  185 N  =  168 Age, years 60 (3.4) 60 (3.6) NS Years since menopause 13.4 (6.1) 13.5 (6.0) NS BMI (kg/m2) 26.1 26.7 NS ⩾1 Prevalent vertebral fracture (%) 153 (82.7) 131 (78.0) NS Previous non-vertebral fracture (%) 50 (27.0) 31 (18.5) NS Bone mineral density Lumbar spine T score −3.7 (1.0) −3.5 (1.2) NS Femoral neck T score −2.5 (0.8) −2.5 (0.8) Femoral neck T score (NHANES III reference) −1.92 −1.92 Data are presented as mean values (SD).

evalent vertebral fracture (%) 153 (82.7) 131 (78.0) NS Previous non-vertebral fracture (%) 50 (27.0) 31 (18.5) NS Bone mineral density Lumbar spine T score −3.7 (1.0) −3.5 (1.2) NS Femoral neck T score −2.5 (0.8) −2.5 (0.8) Femoral neck T score (NHANES III reference) −1.92 −1.92 Data are presented as mean values (SD). *Wilcoxon test between the two groups of treatment. BMI, body mass index; NHANES, National Health and Nutrition Examination Survey III.

evalent vertebral fracture (%) 153 (82.7) 131 (78.0) NS Previous non-vertebral fracture (%) 50 (27.0) 31 (18.5) NS Bone mineral density Lumbar spine T score −3.7 (1.0) −3.5 (1.2) NS Femoral neck T score −2.5 (0.8) −2.5 (0.8) Femoral neck T score (NHANES III reference) −1.92 −1.92 Data are presented as mean values (SD). *Wilcoxon test between the two groups of treatment. BMI, body mass index; NHANES, National Health and Nutrition Examination Survey III. Over 3 years, treatment with strontium ranelate significantly reduced the risk of vertebral fracture by 43% (RR 0.57; 95% CI 0.36 to 0.92, p = 0.019), with an incidence of vertebral fractures of 16.9% in the strontium ranelate group versus 29.6% in the placebo group. This efficacy in reducing the risk of vertebral fractures was sustained over 4 years of treatment with strontium ranelate, with a reduction by 35% (RR 0.65, 95% CI 0.42 to 0.99, p = 0.049) and an incidence of vertebral fractures of 21.6% in the strontium ranelate group versus 32.8% in the placebo group. There was a trend towards a reduction in the risk of vertebral fracture over the first year, which was not statistically significant (fig 1). A significant effect of strontium ranelate compared with placebo was also observed regarding symptomatic vertebral fractures (defined as radiological fractures plus concomitant back pain or a decrease in body height by at least 1 cm) with a 54% reduction in the risk of symptomatic vertebral fracture over 3 years (RR 0.46; 95% CI 0.22 to 0.97, p = 0.033), sustained over 4 years with a 52% reduction (RR 0.48; 95% CI 0.24 to 0.95, p = 0.030). There was no difference in the incidence of non-vertebral fractures between groups over 4 years (14.5% and 14.6% in the treated and placebo groups, respectively).

ebral fracture over 3 years (RR 0.46; 95% CI 0.22 to 0.97, p = 0.033), sustained over 4 years with a 52% reduction (RR 0.48; 95% CI 0.24 to 0.95, p = 0.030). There was no difference in the incidence of non-vertebral fractures between groups over 4 years (14.5% and 14.6% in the treated and placebo groups, respectively). Figure 1 Incidence of vertebral fractures over the study period (1 (M12), 3 (M36) and 4 (M48) years data). RR, relative risk. As shown in fig 2, the BMD mean change from baseline over 3 years at the spine and hip sites was 11.8% and 4.5%, respectively, for the strontium ranelate-treated group, and −2.8% and −3.0% at these two sites in the placebo group. Over 4 years, BMD increased further in the strontium ranelate group, with a mean change of 15.8% and 7.1% at the spine and hip, whereas it remained lower than baseline values in the placebo group, with mean changes of −2.4% and −2.8% at these two sites (significant difference between groups, p<0.001 at both sites).

up. Over 4 years, BMD increased further in the strontium ranelate group, with a mean change of 15.8% and 7.1% at the spine and hip, whereas it remained lower than baseline values in the placebo group, with mean changes of −2.4% and −2.8% at these two sites (significant difference between groups, p<0.001 at both sites). Figure 2 Changes in spine and hip bone mineral density over 4 years in postmenopausal women with osteoporosis aged 50–65 years receiving either placebo or strontium ranelate. Over the 4-year follow-up, clinical, serious, drug-related adverse effects were similar in the placebo and treated groups (no clinical significance between groups regarding the incidence of nausea, diarrhoea, headache, dermatitis, eczema, venous thromboembolism events) and the overall safety profile was very similar to that already described for the whole SOTI study population over 3 and 4 years. No case of pulmonary embolism or hypersensitivity reaction was observed in this study population. DISCUSSION This study shows that strontium ranelate (2 g/day) decreases the risk of vertebral fractures over 4 years in women aged 50–65 years with severe osteoporosis.

As strontium is a heavier element than calcium, its presence in bone could lead to an overestimation of BMD measurement. However, a strong association has been demonstrated for strontium ranelate-treated patients between the increase in total hip and/or femoral neck BMD and the subsequent decrease in fracture incidence.12 Few studies are available with other antiosteoporotic drugs in women aged 50–65 years. The Women’s Health Initiative showed that treatment with oestrogens alone, or oestrogen plus progestin, reduces the incidence of fractures compared with control,13 but the study was conducted in healthy postmenopausal women. Young age was not a significant factor in affecting the efficacy of raloxifene, in a population of postmenopausal women with a prevalence of 37% of vertebral fractures.14 Several studies have shown the absence of effect of age on response to antiosteoporotic treatment, but this point was assessed either in patients older than 65 years15 or in a population with a mean age of 70 years.16 Our study indicates a significant efficacy of strontium ranelate in reducing the risk of subsequent vertebral fractures in young postmenopausal women with severe osteoporosis. These data together with previous reports confirm the efficacy of this antiosteoporotic drug at all ages. Competing interests: None.

Osteoarthritis (OA) is the most common rheumatic joint disease. It is reported to be more prevalent than all other forms of arthritis, and this prevalence seems to be increasing.1 The typical clinical manifestations are pain, stiffness and physical disability. Localised loss of hyaline articular cartilage and adjacent bone remodelling are the key structural changes of OA, and local inflammation may also contribute to the pain and joint damage.2 Knee OA is the most common form of the disease, followed by hand OA (HOA).3 The majority of people aged 55 years and over have radiographic changes of OA in at least one hand joint and approximately one-fifth of this population has symptomatic HOA.4 5 The prevalence of HOA increases with age and is higher in females than in males. 4 Recent studies also indicate that the burden of disease for patients with HOA is considerable across a variety of dimensions of health-related quality of life.6

joint and approximately one-fifth of this population has symptomatic HOA.4 5 The prevalence of HOA increases with age and is higher in females than in males. 4 Recent studies also indicate that the burden of disease for patients with HOA is considerable across a variety of dimensions of health-related quality of life.6 Non-steroidal anti-inflammatory drugs (NSAIDs), including selective cyclo-oxygenase-2 (COX-2) inhibitors are important symptom-modifying therapeutic options for patients with OA.7–10 However, NSAIDs are associated with risk of gastrointestinal adverse events and the selective COX-2 inhibitors have come under special scrutiny because of cardiovascular adverse effects.11 12 Similar concerns have recently been raised about the cardiovascular safety of non-naproxen NSAIDs,13 and the efficacy of long-term dosing of NSAIDs in knee OA has also been questioned.14 Few controlled clinical trials have addressed the efficacy of pharmacological therapies in HOA and, in particular, few controlled studies have included a placebo group.10 15 Thus, there is an obvious need to document the efficacy of existing drugs and, in particular, to identify new and effective agents for patients with HOA.16

olled clinical trials have addressed the efficacy of pharmacological therapies in HOA and, in particular, few controlled studies have included a placebo group.10 15 Thus, there is an obvious need to document the efficacy of existing drugs and, in particular, to identify new and effective agents for patients with HOA.16 Corticosteroids are a mainstay of effective anti-inflammatory therapy in many clinical settings, but the side effects associated with chronic administration have limited their use in OA to occasional intra-articular administration. It has long been a goal to develop a therapeutic agent with the anti-inflammatory and disease modifying activities of corticosteroids without their associated side effects.17 One approach to creating such a therapeutic agent is to develop a drug combination that contains a glucocorticoid and an enhancing agent that pair synergistically to generate a powerful anti-inflammatory effect. Synergistic combinations can have an effect that is greater and more selective than the sum of the activities of the individual components, and thus can provide greater therapeutic benefit with lower toxicity. Synergy is often observed in multi-target therapeutics that modulate the activity of two or more molecular targets to create a novel therapeutic action.18 19 In vivo models testing anti-inflammatory combinations containing a low dose steroid have demonstrated a synergistic interaction between the steroid and the enhancing agent that produces an effect equivalent to that of a high dose steroid alone, without indications of high dose steroid side effects.

apeutic action.18 19 In vivo models testing anti-inflammatory combinations containing a low dose steroid have demonstrated a synergistic interaction between the steroid and the enhancing agent that produces an effect equivalent to that of a high dose steroid alone, without indications of high dose steroid side effects. CRx-102 is one such novel synergistic drug candidate and is in clinical development for the treatment of immunoinflammatory diseases including rheumatoid arthritis (RA) and OA. This drug candidate comprises a combination of a low dose of prednisolone (3 mg) and 200 or 400 mg dipyridamole. According to results from pre-clinical pharmacology experiments, CRx-102 works through a novel mechanism of action by which dipyridamole selectively amplifies prednisolone’s anti-inflammatory and immunomodulatory effects without replicating steroid side effects.20–22 The objective of the current phase 2 study was to evaluate the efficacy and safety of the novel synergistic drug candidate CRx-102 compared to placebo in patients with HOA over a 6-week dosing period.

amplifies prednisolone’s anti-inflammatory and immunomodulatory effects without replicating steroid side effects.20–22 The objective of the current phase 2 study was to evaluate the efficacy and safety of the novel synergistic drug candidate CRx-102 compared to placebo in patients with HOA over a 6-week dosing period. MATERIALS AND METHODS Study population Males and females between the ages of 30 and 70 years with HOA, as defined by the American College of Rheumatology (ACR) criteria,23 were enrolled in the study. Additional inclusion criteria included presence of more than one swollen joint and more than one tender joint, a Kellgren–Lawrence (K–L) score of two or more on radiographs and self-reported hand pain that had to be at least 30 mm on the Australian/Canadian Osteoarthritis Hand Index (AUSCAN) visual analogue scales (VAS).24 25 All subjects had to sign and date an informed consent. The regional ethics committee evaluated the study, the storage and analyses of data was licensed from the data inspectorate and approval for the collection of biologic material was obtained from the Department of Health.

(AUSCAN) visual analogue scales (VAS).24 25 All subjects had to sign and date an informed consent. The regional ethics committee evaluated the study, the storage and analyses of data was licensed from the data inspectorate and approval for the collection of biologic material was obtained from the Department of Health. Subjects who were pregnant, lactating or using hormonal birth control pills as well as subjects with a history of hypersensitivity to corticosteroids and/or dipyridamole, taking bisphosphonates or who had a positive rheumatoid factor test were excluded from the study. Furthermore, subjects who had taken any corticosteroids orally, topically or intra-articularly 3 months prior to enrolment were also excluded. Other exclusion criteria included a history of asthma, HIV infection, hepatitis, currently uncontrolled diabetes, use of statins in a dose that had changed during the prior 3 months, known active infection or a surgical procedure within 30 days of study initiation. Since one of the components of CRx-102 is dipyridamole, patients on warfarin, ticlopidine, clopidogrel or aspirin of more than 81 mg daily were also excluded from entering the study. Design and medication The study was designed as a 6-week randomised, blinded, placebo-controlled four-centre parallel group study. Within 2 weeks of a screening visit, patients who fulfilled the eligibility criteria were randomly assigned to CRx-102 or placebo. Follow-up assessments were performed after 7, 14, 28 and 42 days, with a final safety visit after approximately 56 days.

mised, blinded, placebo-controlled four-centre parallel group study. Within 2 weeks of a screening visit, patients who fulfilled the eligibility criteria were randomly assigned to CRx-102 or placebo. Follow-up assessments were performed after 7, 14, 28 and 42 days, with a final safety visit after approximately 56 days. A total of 83 subjects were randomised 1:1 to a daily dose regimen of either CRx-102 or placebo. CRx-102 for days 1–7 combined 2 mg of prednisolone with 100 mg of dipyridamole at 8am and 1 mg of prednisolone with 100 mg of dipyridamole at 1pm. From days 8–42 CRx-102 combined 2 mg prednisolone with 200 mg of dipyridamole at 8 am and 1 mg of prednisolone with 200 mg dipyridamole at 1pm. Patients in the placebo group received an equal number of tablets, dosed at the same time of day as for the test compound and all containing placebo. Paracetamol was provided as rescue medication throughout the study at a daily dose of up to 4000 mg, and the usage was recorded. The use of NSAIDs in all subjects was stopped for the duration of the study starting at the initial screening visit.

A total of 83 subjects were randomised 1:1 to a daily dose regimen of either CRx-102 or placebo. CRx-102 for days 1–7 combined 2 mg of prednisolone with 100 mg of dipyridamole at 8am and 1 mg of prednisolone with 100 mg of dipyridamole at 1pm. From days 8–42 CRx-102 combined 2 mg prednisolone with 200 mg of dipyridamole at 8 am and 1 mg of prednisolone with 200 mg dipyridamole at 1pm. Patients in the placebo group received an equal number of tablets, dosed at the same time of day as for the test compound and all containing placebo. Paracetamol was provided as rescue medication throughout the study at a daily dose of up to 4000 mg, and the usage was recorded. The use of NSAIDs in all subjects was stopped for the duration of the study starting at the initial screening visit. Assessments The AUSCAN24 25 was used as the primary assessment tool. Previous studies have demonstrated that this instrument has acceptable reliability, construct validity and responsiveness. The translated Norwegian version has also satisfactory clinimetric properties.26 We chose to use the version with responses on VAS to each item. AUSCAN has five items measuring pain, one measuring stiffness and nine measuring physical function. The pain and physical functioning scores were normalised to a 0–100-point scale prior to the analyses.

rsion has also satisfactory clinimetric properties.26 We chose to use the version with responses on VAS to each item. AUSCAN has five items measuring pain, one measuring stiffness and nine measuring physical function. The pain and physical functioning scores were normalised to a 0–100-point scale prior to the analyses. Additional patient-reported measures included a joint pain VAS (question: how would you describe the intensity of your joint pain during the last 2 days?) and global assessment VAS (question: we ask you to evaluate the activity of your osteoarthritis over the last 2 days. When you take all symptoms into consideration, how will you evaluate your condition?). The patients did not have access to scores from previous visits when they were performing each subsequent assessment. The patients were clinically examined for vital signs at each visit and each individual finger joint (distal interphalangeal (DIP), proximal interphalangeal (PIP), metacarpophalangeal (MCP) and carpometacarpal (CMC)) on the right and left hand was examined for the presence of joint tenderness, soft tissue swelling, bony enlargement and limited joint motion. A score was calculated using the number of PIP and DIP joints for the presence of each of these four characteristics (ie, score range 0–18 for each). Sera were frozen and stored and later analysed with a high sensitivity technique to determine levels of C-reactive protein (CRP). Regular blood chemistry, including fasting blood glucose, was recorded. Unsolicited adverse events were also recorded.

The patients were clinically examined for vital signs at each visit and each individual finger joint (distal interphalangeal (DIP), proximal interphalangeal (PIP), metacarpophalangeal (MCP) and carpometacarpal (CMC)) on the right and left hand was examined for the presence of joint tenderness, soft tissue swelling, bony enlargement and limited joint motion. A score was calculated using the number of PIP and DIP joints for the presence of each of these four characteristics (ie, score range 0–18 for each). Sera were frozen and stored and later analysed with a high sensitivity technique to determine levels of C-reactive protein (CRP). Regular blood chemistry, including fasting blood glucose, was recorded. Unsolicited adverse events were also recorded. Analyses AUSCAN pain was the predefined primary endpoint. The sample size was calculated based on an assumed improvement of 20% in AUSCAN pain VAS in the CRx-102 group compared to a 10 % improvement in the placebo-group from baseline to day 42 with an alpha of 0.05%, to achieve 80% power assuming a 15% drop out rate using a one tailed t test for the comparison of the mean changes.

ample size was calculated based on an assumed improvement of 20% in AUSCAN pain VAS in the CRx-102 group compared to a 10 % improvement in the placebo-group from baseline to day 42 with an alpha of 0.05%, to achieve 80% power assuming a 15% drop out rate using a one tailed t test for the comparison of the mean changes. The primary analysis was conducted on the intention to treat (ITT) population that included all patients who took at least one dose of study medication. Secondarily, an analysis was performed in the per-protocol population, which was defined as all subjects who received at least one dose of study medication, had no major protocol violations and had a study drug compliance of at least 75%. The treatment effects were derived from analysis of covariance (ANCOVA) adjusting for the baseline values. These analyses tested the difference (with 95% confidence intervals) between these adjusted mean changes in the active drug compared to the placebo group from baseline to 42 days. The last non-missing post-baseline observation was carried forward (LOCF) to replace subsequent missing values. Two-sided p values for the differences in least square means adjusted for baseline are presented (the study protocol recommended use of one-sided tests, but we found it more appropriate to use two-sided tests, ie, a more conservative approach, in this work). The statistical analyses were performed by the sponsor in collaboration with the principal investigator (TKK).

quare means adjusted for baseline are presented (the study protocol recommended use of one-sided tests, but we found it more appropriate to use two-sided tests, ie, a more conservative approach, in this work). The statistical analyses were performed by the sponsor in collaboration with the principal investigator (TKK). RESULTS A total of 83 patients (77 (93%) females) with a mean (standard deviation (SD)) age of 60.4 (5.2) years were enrolled into the study. A Consolidated Standards of Reporting Trials (CONSORT) flow-chart is shown in fig 1. The study groups were comparable for demographic characteristics (table 1). The patients in the active treatment group had wider OA joint involvement (table 1), but the baseline levels of pain and physical limitations were comparable (table 2). Figure 1 Flow chart of the selection of patients for this study and patient disposition. Table 1 Demographic variables and joint involvement (mean (SD)) for continuous variables, percentages for counts) CRx-102 (n = 42) Placebo (n = 41) Age 61.1 (5.0) 59.6 (5.3) Female 93 93 Caucasian 100 100 Height, cm 166.0 (6.7) 167.7 (8.2) Weight, kg 71.1 (12.0) 74.5 (14.6) Percentage with OA joint involvement: Right MTP joint I 36 10 Left MTP joint I 33 10 Lumbar spine 24 17 Cervical spine 19 7 Right hip 17 10 Left hip 10 12 Right knee 19 7 Left knee 17 10 Other joints 17 15 Finger joints: percentage with radiographic grade 2–4 K–L score: Minimal (2) 14 12 Moderate (3) 45 32 Severe (4) 40 56 K–L, Kellgren–Lawrence; MTP, metatarsophalangeal; OA, osteoarthritis.

nt I 33 10 Lumbar spine 24 17 Cervical spine 19 7 Right hip 17 10 Left hip 10 12 Right knee 19 7 Left knee 17 10 Other joints 17 15 Finger joints: percentage with radiographic grade 2–4 K–L score: Minimal (2) 14 12 Moderate (3) 45 32 Severe (4) 40 56 K–L, Kellgren–Lawrence; MTP, metatarsophalangeal; OA, osteoarthritis. Table 2 Baseline mean (SD) values of efficacy variables, adjusted mean changes from baseline to day 42 (least squares mean (standard error of mean)) and treatment effect (mean difference (95% CI) placebo minus CRx-102) in the intention to treat population Baseline Changes Treatment effect p Value CRx-102 (n = 42) Placebo (n = 41) CRx-102 Placebo AUSCAN: Pain 57.9 (20.2) 60.9 (19.4) −14.2 (3.0) −4.0 (3.1) 10.2 (1.6 to 18.7) 0.020 Physical 62.4 (19.5) 67.8 (17.5) −8.1 (2.7) −3.6 (2.7) 4.5 (−3.2 to 12.2) 0.246 Stiffness 61.1 (18.0) 64.5 (21.2) −15.2 (3.2) −7.7 (3.3) 7.5 (−1.7 to 16.7) 0.108 VAS: Joint pain 58.3 (20.1) 62.1 (16.9) −18.6 (3.3) −6.3 (3.3) 12.3 (3.0 to 21.5) 0.010 Patient global 58.0 (19.5) 62.3 (17.9) −15.9 (3.2) −4.2 (3.3) 11.7 (2.5 to 20.8) 0.013 Lab tests: CRP mg/litre 2.5 (2.9) 2.3 (2.2) −0.2 (0.4) 0.1 (0.4) 0.3 (−0.7 to 1.4) 0.536 Joint counts: Tender joints 9.5 (4.7) 9.4 (4.6) −3.6 (0.7) −2.4 (0.7) 1.2 (−0.9 to 3.2) 0.258 Soft tissue swelling 5.5 (4.7) 5.0 (4.4) −2.4 (0.5) −1.6 (0.5) 0.8 (−0.6 to 2.2) 0.262 AUSCAN, Australian/Canadian Osteoarthritis Hand Index; CRP, C-reactive protein; VAS, visual analogue scale.

1 (0.4) 0.3 (−0.7 to 1.4) 0.536 Joint counts: Tender joints 9.5 (4.7) 9.4 (4.6) −3.6 (0.7) −2.4 (0.7) 1.2 (−0.9 to 3.2) 0.258 Soft tissue swelling 5.5 (4.7) 5.0 (4.4) −2.4 (0.5) −1.6 (0.5) 0.8 (−0.6 to 2.2) 0.262 AUSCAN, Australian/Canadian Osteoarthritis Hand Index; CRP, C-reactive protein; VAS, visual analogue scale. A significant difference (p = 0.020) in the AUSCAN pain score (the primary endpoint) at the end of the study was demonstrated in favour of CRx-102 compared to placebo in the ITT population (table 2). CRx-102 was also statistically superior to placebo at 42 days for joint pain VAS and patient global VAS (table 2). Figure 2 displays how the improvement developed over time and also that the differences between CRx-102 and placebo were discernible after 2 weeks. The comparison between CRx-102 and placebo in the per-protocol population was also consistently in favour of CRx-102 (table 3).

t pain VAS and patient global VAS (table 2). Figure 2 displays how the improvement developed over time and also that the differences between CRx-102 and placebo were discernible after 2 weeks. The comparison between CRx-102 and placebo in the per-protocol population was also consistently in favour of CRx-102 (table 3). Figure 2 Mean improvements from baseline adjusted for baseline values Australian/Canadian Osteoarthritis Hand Index (AUSCAN) pain (A), AUSCAN stiffness (B), AUSCAN physical (C), pain visual analogue scale (VAS) (D) and global VAS (E) in patients receiving CRx-102 and placebo (intention to treat population) with one-sided p values for the differences of adjusted least square means. Table 3 Baseline mean (SD) values of efficacy variables, adjusted mean changes from baseline to day 42 (least squares mean (SEM)) and treatment effect (mean difference (95% CI) placebo minus CRx-102) in the per-protocol population Baseline Changes Treatment effect p Value CRx-102 (n = 26) Placebo (n = 33) CRx-102 Placebo AUSCAN: Pain 61.9 (16.6) 63.8 (17.2) −20.5 (4.1) −6.2 (3.7) 14.3 (3.2 to 25.5) 0.012 Physical 64.9 (18.9) 70.9 (15.5) −12.9 (3.7) −5.9 (3.2) 7.0 (−2.9 to 16.8) 0.061 Stiffness 62.9 (17.4) 67.8 (19.8) −20.3 (4.4) −8.3 (3.9) 12.0 (0.2 to 23.9) 0.047 VAS: Joint pain 59.8 (19.5) 62.9 (16.7) −23.5 (4.4) −6.3 (3.9) 17.2 (5.5 to 28.9) 0.005 Patient global 61.5 (17.5) 62.5 (17.6) −23.4 (4.0) −4.6 (3.6) 18.8 (8.1 to 29.5) 0.001 Lab tests: CRP mg/litre 2.0 (1.8) 2.3 (2.2) −0.2 (0.5) 0.4 (0.4) 0.6 (−0.7 to 1.8) 0.364 Joint counts: Tender joints 9.6 (4.8) 9.8 (4.7) −5.0 (1.0) −2.6 (0.9) 2.4 (−0.3 to 5.0) 0.083 Soft tissue swelling 5.6 (4.6) 4.9 (4.3) −3.1 (0.6) −1.9 (0.5) 1.3 (−0.3 to 2.8) 0.116 AUSCAN, Australian/Canadian Osteoarthritis Hand Index; CRP, C-reactive protein; VAS, visual analogue scale.

4 (0.4) 0.6 (−0.7 to 1.8) 0.364 Joint counts: Tender joints 9.6 (4.8) 9.8 (4.7) −5.0 (1.0) −2.6 (0.9) 2.4 (−0.3 to 5.0) 0.083 Soft tissue swelling 5.6 (4.6) 4.9 (4.3) −3.1 (0.6) −1.9 (0.5) 1.3 (−0.3 to 2.8) 0.116 AUSCAN, Australian/Canadian Osteoarthritis Hand Index; CRP, C-reactive protein; VAS, visual analogue scale. The tender and swollen joint counts of the 18 PIP and DIP joints were numerically improved in the CRx-102 group compared to placebo (table 2) and the group differences approach statistical significance in the per-protocol population (table 3). The counts of joints with limited motion and bony swelling did not change during the study (data not shown). The proportions of patients reporting at least one adverse event in the CRx-102 and placebo groups were 64% and 32%, respectively. The most common adverse event in both groups was headache, which was more frequently reported in the CRx-102 group (52%) than in the placebo group (15%). A total of 21% of the patients in the CRx-102 group also reported nausea, versus none in the placebo group (table 4). No serious adverse events were reported in the CRx-102 group. Discontinuation occurred more often in the CRx-102 (n = 16, 38%) than in placebo group (n = 6, 15%) (fig 1) and was mostly due to headache. Most of the discontinuations occurred early in the study (fig 3).

nausea, versus none in the placebo group (table 4). No serious adverse events were reported in the CRx-102 group. Discontinuation occurred more often in the CRx-102 (n = 16, 38%) than in placebo group (n = 6, 15%) (fig 1) and was mostly due to headache. Most of the discontinuations occurred early in the study (fig 3). Figure 3 Time from dose 1 to withdrawal from the treatment (intention to treat (ITT) population, Kaplan–Meier plot). Table 4 Most commonly reported adverse events (AE) (⩾5% of total subjects) in the intention to treat (ITT) population (no. of patients (%)) CRx-102 (n = 42) Placebo (n = 41) Total (n = 83) Subjects with at least one AE 27 (64) 13 (32) 40 (48) Headache 22 (52) 6 (15) 28 (34) Nausea 9 (21) 0 9 (11) DISCUSSION HOA is a frequent disease in people more than 60 years of age and imparts a considerable disease burden, on the individual6 and in society.2 However, few studies have formally addressed the efficacy of symptom-modifying drugs in HOA.10 15 This phase 2 study demonstrated that CRx-102, a combination of a low dose prednisolone and a titrated dose of dipyridamole, was superior to placebo across a variety of patient-reported measures.

ndividual6 and in society.2 However, few studies have formally addressed the efficacy of symptom-modifying drugs in HOA.10 15 This phase 2 study demonstrated that CRx-102, a combination of a low dose prednisolone and a titrated dose of dipyridamole, was superior to placebo across a variety of patient-reported measures. In HOA the DIP and PIP joints as well as the first carpometacarpal (CMC) joint are particularly affected,27 but patients with HOA tend to have involvement of multiple joints, a condition that is often referred to as generalised OA.28 The joints most frequently involved in this study population were the first metatarsophalangeal (MTP) joint, the knee and additionally the lumbar and cervical spine (table 1). We do not know how CRx-102 influenced these other joint areas, since we were not measuring low back pain, pain in the big toe or were not using the Western Ontario and McMaster Universities (WOMAC) index,29 a specific measure of knee and hip OA. However, the question on joint pain did not specifically address pain in the finger joints and the clear differentiation between CRx-102 and placebo for this measure may indirectly support an efficacy that goes beyond the finger joints.

o and McMaster Universities (WOMAC) index,29 a specific measure of knee and hip OA. However, the question on joint pain did not specifically address pain in the finger joints and the clear differentiation between CRx-102 and placebo for this measure may indirectly support an efficacy that goes beyond the finger joints. Inflammation has been recognised as a feature of the disease process in OA.30–33 A synergistic drug candidate such as CRx-102 comprises two components that are designed to act synergistically through multiple pathways, providing a novel therapeutic effect that neither component can achieve on their own. CRx-102 has been shown to have strong anti-inflammatory effects in preclinical assays with a greater average percentage inhibition of tumour necrosis factor alpha (TNFα) and interleukin 1 (IL1) release from human blood buffy coat cells than the inhibition seen with each of the single agents alone.20 Similarly, in the rat model of lipopolysaccharide (LPS)-induced TNFα release, the inhibition of release was greater with CRx-102 than the individual components.21 This finding was replicated in the rat models of adjuvant and collagen-induced arthritis.21

than the inhibition seen with each of the single agents alone.20 Similarly, in the rat model of lipopolysaccharide (LPS)-induced TNFα release, the inhibition of release was greater with CRx-102 than the individual components.21 This finding was replicated in the rat models of adjuvant and collagen-induced arthritis.21 The preclinical studies on the mechanism of action of CRx-102 indicate that CRx-102 acts through multiple molecular pathways to create a synergistic immunomodulatory effect without amplifying traditional glucocorticoid-associated side effects. Additionally, the components of CRx-102 affect the activity of key transcription factors, but not their nuclear localisation nor does it increase glucocorticoid receptor translocation or transcription from positive glucocorticoid response element promoters relative to low dose prednisolone on its own.22 Dipyridamole may contribute to the action of CRx-102 by increasing cAMP in part though inhibition of phosphodiesterases relevant to inflammation, as well as through modulating adenosine transport resulting in increased extra-cellular endogenous adenosine.34 Adenosine can suppress the release of TNFα from activated monocytes and macrophages, but a small placebo-controlled study was unable to demonstrate clinical improvement following treatment with dipyridamole.35

s well as through modulating adenosine transport resulting in increased extra-cellular endogenous adenosine.34 Adenosine can suppress the release of TNFα from activated monocytes and macrophages, but a small placebo-controlled study was unable to demonstrate clinical improvement following treatment with dipyridamole.35 This study used the pain dimension in AUSCAN as the primary endpoint, but also included an assessment of two other core measures, function and patient global assessment.36 37 AUSCAN has been developed as a disease-specific patient-reported measure in hand OA. Development of the AUSCAN HOA index was based on the questions and experience from the WOMAC38 as well as information from patient interviews. 24 AUSCAN was later validated in separate studies25 and has also been validated in the Norwegian language.26 Consistent with the present findings, Allen et al39 showed that the AUSCAN index can reliably measure changes in pain, stiffness and function, thereby providing a meaningful endpoint for clinical trials in HOA. For feasibility reasons we chose not to include other hand indexes that are alternatives to AUSCAN in the assessment of HOA.40 41 The patients had to stop treatment with NSAIDs and paracetamol at the screening visit, but a disease flare was not required for inclusion in the study. We unfortunately did not capture detailed information about ongoing medication at the screening visit, and we do not have detailed data on the changes in efficacy endpoints from screening to randomisation.

th NSAIDs and paracetamol at the screening visit, but a disease flare was not required for inclusion in the study. We unfortunately did not capture detailed information about ongoing medication at the screening visit, and we do not have detailed data on the changes in efficacy endpoints from screening to randomisation. In rheumatoid arthritis, several composite disease measures have been developed over the past years, and these are most often based on joint counts. Similar composite disease specific measures do not exist in HOA. We examined the number of finger joints with tenderness, soft tissue swelling, limited motion and bony swelling at baseline and during follow-up. It is reasonable to assume that tenderness and soft tissue swelling at least in part reflect the inflammatory component of the disease. Joint counts were numerically improved in the CRx-102 treatment group compared to the placebo group (tables 3 and 4). The number of joints with bony swelling did not change during the study; this endpoint was not expected to be influenced by anti-inflammatory therapy. More research is needed to address the development of joint count-indices for HOA that are valid, reliable and responsive.

ompared to the placebo group (tables 3 and 4). The number of joints with bony swelling did not change during the study; this endpoint was not expected to be influenced by anti-inflammatory therapy. More research is needed to address the development of joint count-indices for HOA that are valid, reliable and responsive. Headache was the most commonly reported adverse event in this trial and was most frequently observed during the first days of treatment (table 4 and fig 3). This type of headache has previously been associated with dipyridamole administration42 and also with other cardiovascular pharmaceutical products with vasodilatating properties. As in the current study, these headaches typically occur during the first days of treatment.42 This early and high withdrawal rate is a potential limitation of CRx-102. Thus, formulation development is necessary to optimise the synergistic benefits of CRx-102 demonstrated in this and other phase 2 studies and to minimise the observed incidence of headache. The objective of the formulation should be to deliver pulsed-doses of prednisolone with concurrently releasing dipyridamole at a rate that maintains the anti-inflammatory synergy and minimises the vasodilator effects that are known to cause headaches.

phase 2 studies and to minimise the observed incidence of headache. The objective of the formulation should be to deliver pulsed-doses of prednisolone with concurrently releasing dipyridamole at a rate that maintains the anti-inflammatory synergy and minimises the vasodilator effects that are known to cause headaches. In summary, preclinical studies have supported that the synergistic drug candidate CRx-102, comprised of dipyramidole and low dose prednisolone, has biological effects that exceed the effects of each individual component. This placebo-controlled phase 2 study suggests that the combination is effective in patients with HOA. Follow-up studies should be initiated to compare the clinical effects of CRx-102 versus the individual components in HOA and other rheumatic joint diseases. Competing interests: MN, YZ and JL are employees of CombinatoRx Inc, the producer of CRx-102. The statistical analyses were performed by the sponsor in collaboration with the principal investigator (TKK). Funding: This study was funded by CombinatoRx Inc, the producer of CRx-102. Ethics approval: The regional ethics committee evaluated the study, the storage and analyses of data was licensed from the data inspectorate, and approval for the collection of biologic material was obtained from the Department of Health.

Rheumatoid arthritis (RA) is a chronic inflammatory disorder affecting synovial tissue in multiple joints. Early treatment with disease-modifying antirheumatic drugs (DMARDs) has become the cornerstone of therapy. Recently, new biological therapies, including rituximab, have become available. Rituximab is a chimaeric monoclonal antibody directed against the CD20 antigen expressed by B cells, which significantly improves disease symptoms in patients with high levels of disease activity despite treatment with methotrexate (MTX) or tumour necrosis factor (TNF) blockers.1–3 This clinical effect strongly supports the notion that B cells play a critical role in the pathogenesis of RA, although the exact mechanism of rituximab treatment in RA remains to be elucidated. We have previously shown that rituximab treatment causes a rapid and specific decrease in numbers of B cells at the primary site of inflammation, the rheumatoid synovium,4 which was recently confirmed in another study.5 The early synovial tissue response varies between patients, which is in contrast with the marked B cell depletion observed in the peripheral blood of nearly all patients with RA. Interestingly, in the earlier, smaller studies there was no significant decrease in numbers of inflammatory cells other than synovial B cells 4–8 weeks after initiation of treatment.4 5

tween patients, which is in contrast with the marked B cell depletion observed in the peripheral blood of nearly all patients with RA. Interestingly, in the earlier, smaller studies there was no significant decrease in numbers of inflammatory cells other than synovial B cells 4–8 weeks after initiation of treatment.4 5 Currently, no data are available on the synovial tissue response to rituximab treatment after more prolonged follow-up and its predictive value related to clinical improvement. The current study was performed to investigate the kinetics of this response in detail and to identify possible predictors of clinical response in patients with RA. PATIENTS AND METHODS Patients and treatment protocol A total of 24 patients were included in this study analysing synovial biopsies at three timepoints: before treatment, at 4 weeks and 16 weeks after initiation of rituximab treatment; 17 of these patients participated in a previously reported study on the synovial tissue response to rituximab at 4 weeks only.4 The patients had active RA;6 active disease was defined as having ⩾4 tender joints and ⩾4 swollen joints of 28 joints assessed, and at least one of the following: erythrocyte sedimentation rate (ESR) ⩾28 mm/h, serum C-reactive protein (CRP) levels ⩾15 mg/litre, or morning stiffness ⩾45 min. In addition, patients needed to be positive for IgM-RF and/or anti-citrullinated peptide antibodies (ACPA) and have active arthritis (defined by the presence of pain and swelling) of a wrist, knee or ankle joint, amenable for arthroscopy.

serum C-reactive protein (CRP) levels ⩾15 mg/litre, or morning stiffness ⩾45 min. In addition, patients needed to be positive for IgM-RF and/or anti-citrullinated peptide antibodies (ACPA) and have active arthritis (defined by the presence of pain and swelling) of a wrist, knee or ankle joint, amenable for arthroscopy. All study patients were on stable doses of MTX (5–30 mg/week) for at least 28 days prior to enrolment. Stable prednisone therapy (⩽10 mg/day) and non-steroidal anti-inflammatory drug (NSAID) treatments were allowed. All other DMARDs and biological agents were withdrawn at least 4 weeks prior to study inclusion, with a washout period for leflunomide, infliximab, adalimumab and etanercept of >8 weeks prior to randomisation. The study protocol was approved by the Medical Ethics Committee of the Academic Medical Center/University of Amsterdam, and all patients gave written informed consent before participation in the study. Treatment consisted of two infusions of 1000 mg of rituximab (Roche, Woerden, The Netherlands) on days 1 and 15 after premedication with 2 mg clemastine fumarate intravenously and 1000 mg acetaminophen orally. Peri-infusional treatment with corticosteroids was not allowed, as this could have influenced the features of synovial inflammation.

two infusions of 1000 mg of rituximab (Roche, Woerden, The Netherlands) on days 1 and 15 after premedication with 2 mg clemastine fumarate intravenously and 1000 mg acetaminophen orally. Peri-infusional treatment with corticosteroids was not allowed, as this could have influenced the features of synovial inflammation. The 28-joint Disease Activity Score (DAS28)7 was measured every month after treatment. Serum levels of IgM-rheumatoid factor (RF) and ACPA (anti-CCP2 ELISA, Immunoscan RA, Mark 2, Euro Diagnostica, Arnhem, the Netherlands) were determined at baseline and weeks 4, 16, 24 and 36 after treatment. Synovial biopsies were obtained from the same clinically involved joint at three timepoints: before as well as 4 and 16 weeks after the first infusion of rituximab. Blood lymphocyte populations Flow cytometry on fresh peripheral blood samples was performed at baseline and at 4, 16 and 24 weeks after the first rituximab infusion. B cells and T cell subsets were detected by real-time fluorescence-activated cell sorting (FACS) using a FACSCalibur Flow Cytometer (Becton Dickinson, San Jose, California, USA) with antibodies against CD19, CD3, CD4 and CD8 (all from Becton Dickinson). CD19 was chosen because of its similar expression to CD20 on B cell subsets, without interference with the circulating rituximab antibody. The lower limit of detection of CD19+ B cells was set at 0.01×109 cells/litre.

San Jose, California, USA) with antibodies against CD19, CD3, CD4 and CD8 (all from Becton Dickinson). CD19 was chosen because of its similar expression to CD20 on B cell subsets, without interference with the circulating rituximab antibody. The lower limit of detection of CD19+ B cells was set at 0.01×109 cells/litre. Synovial biopsy Serial synovial biopsies were collected by needle arthroscopy from the same actively inflamed joint under local anaesthesia as previously described,8 before as well as 4 and 16 weeks after the first infusion of rituximab. To minimise sampling error at least six biopsy samples were obtained from different sites in the joint during each procedure.9 10 Specimens were directly embedded en bloc in TissueTek OCT (Miles Diagnostics, Elkhart, Indiana, USA) and subsequently snap-frozen in liquid nitrogen.

s after the first infusion of rituximab. To minimise sampling error at least six biopsy samples were obtained from different sites in the joint during each procedure.9 10 Specimens were directly embedded en bloc in TissueTek OCT (Miles Diagnostics, Elkhart, Indiana, USA) and subsequently snap-frozen in liquid nitrogen. Immunohistochemistry From each frozen tissue block serial sections (5 μm) were cut and stained with the following mouse monoclonal antibodies: anti-CD3 (SK7; Becton Dickinson), anti-CD4 (SK3; Becton Dickinson) and anti-CD8 (DK25; Dako, Glostrup, Denmark) to detect T cells, anti-CD22 (CLB-B-ly; Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, Amsterdam, the Netherlands) to detect B cells, anti-CD68 (EBM11; Dako) to detect macrophages, anti-CD55 (clone 67; Serotec, Oxford, UK) to detect fibroblast-like synoviocytes, anti-CD138 (clone B-B4; Immunotech, Marseille, France) to detect plasma cells and anti-CD23 (clone MHM6; Dako) as well as anti-CD21L (a kind gift from Dr Y Liu, Schering-Plough, Dardilly, France) to detect follicular dendritic cells (FDCs), as described previously.4 11 12 Sections with non-assessable tissue, defined by the absence of an intimal lining layer, were not analysed. For control sections, the primary antibodies were omitted or irrelevant antibodies were applied.

Liu, Schering-Plough, Dardilly, France) to detect follicular dendritic cells (FDCs), as described previously.4 11 12 Sections with non-assessable tissue, defined by the absence of an intimal lining layer, were not analysed. For control sections, the primary antibodies were omitted or irrelevant antibodies were applied. To determine the distribution of B cells and FDCs in lymphocyte aggregates, we evaluated two separate tissue sections (each section representing one level of the six biopsy samples that were embedded en bloc), located 50 μm apart, which were stained with the anti-CD22, anti-CD23 and anti-CD21L antibodies. CD22 rather than CD19 was used to detect B cells in tissue, as it results in more reliable staining on the tissue level. The presence of FDCs was assessed by CD23 staining and subsequently confirmed by CD21L staining.

loc), located 50 μm apart, which were stained with the anti-CD22, anti-CD23 and anti-CD21L antibodies. CD22 rather than CD19 was used to detect B cells in tissue, as it results in more reliable staining on the tissue level. The presence of FDCs was assessed by CD23 staining and subsequently confirmed by CD21L staining. Digital image analysis All sections were coded and randomly analysed by computer-assisted image analysis using a Leica DM-RXA light microscope and a program written in the programming language QUIPS with specialised Qwin software (both from Leica, Cambridge, UK).13 14 An independent observer (Marjolein Vinkenoog, Division of Clinical Immunology and Rheumatology of the Academic Medical Centre/University of Amsterdam, The Netherlands) who was unaware of the clinical data performed the acquisition and image analysis. For all markers, 18 high-power fields were analysed, as previously described and validated.15 CD68 expression was analysed separately in the intimal lining layer and the synovial sublining. CD22 expression was analysed on two levels as described above, and the mean number of B cells was used for statistical analysis. Data are expressed as the number of cells per mm2, representing cell density.

and validated.15 CD68 expression was analysed separately in the intimal lining layer and the synovial sublining. CD22 expression was analysed on two levels as described above, and the mean number of B cells was used for statistical analysis. Data are expressed as the number of cells per mm2, representing cell density. Lymphocyte aggregates were counted and classified as previously described.16 Aggregates with a radius of 2–5 cells were defined as grade 1, those with a radius of 5–10 cells were graded as 2, and those with a radius of >10 cells were graded as 3. The total number of grade 1, 2 and 3 was added up per section. The resulting figure was termed the total number of aggregates. Statistical analysis Changes in the DAS28 were compared using the Student paired t test. Changes in peripheral blood lymphocytes, synovial cell populations, as well as serum levels of total IgM, total IgG, IgM-RF and ACPA were compared using Wilcoxon signed rank test for paired data. A mixed linear model was used as a repeated measurements method to confirm the results of the separate paired t test for analysis of changes in synovial parameters. Correlations between changes in synovial cell populations, serum IgM-RF and ACPA levels and the DAS28 between baseline, 4 weeks and 16 weeks after treatment were determined by Spearman correlation coefficient. Differences in baseline synovial cell populations and changes in cell populations between responders and non-responders to treatment were determined by the Mann–Whitney U test for unpaired data.

els and the DAS28 between baseline, 4 weeks and 16 weeks after treatment were determined by Spearman correlation coefficient. Differences in baseline synovial cell populations and changes in cell populations between responders and non-responders to treatment were determined by the Mann–Whitney U test for unpaired data. For analysis of possible predictive biomarkers for clinical response we calculated the Spearman correlation coefficient between the change in DAS28 (between baseline and week 24) and changes in B cells, T cells, plasma cells and intimal and sublining macrophages (between baseline, 4 weeks and 16 weeks). For dichotomous analysis response was determined as a decrease in DAS28 of ⩾1.2 at 24 weeks17 and also as according to the European League Against Rheumatism (EULAR) response criteria.18 The Mann–Whitney U test for unpaired data was used to compare the clinical response in patients with FDCs or lymphocyte aggregates at baseline compare to those without. Parameters that were significantly related to the decrease in DAS28 were subsequently analysed by univariate linear regression analysis to assess the predictive value. Finally, these parameters were analysed together with multiple linear regression analysis in a backward model. RESULTS Clinical and demographic features Clinical and demographic details are shown in table 1.

The Mann–Whitney U test for unpaired data was used to compare the clinical response in patients with FDCs or lymphocyte aggregates at baseline compare to those without. Parameters that were significantly related to the decrease in DAS28 were subsequently analysed by univariate linear regression analysis to assess the predictive value. Finally, these parameters were analysed together with multiple linear regression analysis in a backward model. RESULTS Clinical and demographic features Clinical and demographic details are shown in table 1. Table 1 Baseline characteristics of the study patients Demographic Median age, years (range) 55 (22–75) Female, no. (%) 18 (75) Disease status: Median disease duration, years (range) 12 (0.9–50) Erosive disease, no. (%) 24 (100) Nodular disease, no. (%) 9 (38) Median IgM-RF, kU/litre (IQR) 75 (46–159) Median ACPA, kU/litre (IQR) 240 (95–1031) Median DAS28, (IQR) 6.5 (1.1) Median ESR, mm/h (IQR) 37 (19–55) Median CRP, mg/dl (IQR) 24 (10–76) Medications: Median no. of previous DMARDs, (IQR) 4 (2–5) Median no. of previous biological agents, (IQR) 2 (1–3) Median methotrexate dosage, mg/week (IQR) 15 (10–25) Corticosteroids, no. (%) 16 (67) Median prednisone dosage, mg/day (IQR) 5 (0–10) ACPA, anti-cyclic citrullinated peptide; CRP, C-reactive protein; DAS28, 28-joint Disease Activity Score; DMARDs, disease-modifying antirheumatic drugs; ESR, erythrocyte sedimentation rate; IQR, interquartile range; RF, rheumatoid factor.

Corticosteroids, no. (%) 16 (67) Median prednisone dosage, mg/day (IQR) 5 (0–10) ACPA, anti-cyclic citrullinated peptide; CRP, C-reactive protein; DAS28, 28-joint Disease Activity Score; DMARDs, disease-modifying antirheumatic drugs; ESR, erythrocyte sedimentation rate; IQR, interquartile range; RF, rheumatoid factor. Clinical response to treatment with rituximab One patient withdrew from the study after 20 weeks because of insufficient clinical response. For analysis related to response at week 24 the observations from her week 20 visit were carried forward. The DAS28 was unaltered at 4 weeks (peri-infusional corticosteroids were not allowed), but there was a significant decrease in DAS28 at 16 and 24 weeks (compared to baseline: mean (SD) decrease of 1.6 (1.1) and 1.6 (1.6), respectively; both p<0.001). At 24 weeks 16 of the 24 patients (67%) had a decrease in DAS28 of at least 1.2. In all, 3 patients (13%) had a good response according to the EULAR response criteria, 14 patients (58%) a moderate response and 7 patients (29%) did not fulfil the EULAR response criteria. Decreased IgM-RF and ACPA levels after rituximab treatment The IgM-RF and ACPA levels were not available for the week 24 visit and week 36 for the patient who withdrew from the study after 20 weeks because of insufficient clinical response. For another patient the visit at week 36 was not performed because of personal circumstances. Thus, serial IgM-RF and ACPA levels were available for 22 patients.

els were not available for the week 24 visit and week 36 for the patient who withdrew from the study after 20 weeks because of insufficient clinical response. For another patient the visit at week 36 was not performed because of personal circumstances. Thus, serial IgM-RF and ACPA levels were available for 22 patients. There was a highly significant decrease in serum levels of IgM-RF at 16 (p = 0.006) and 24 weeks (p<0.001) after treatment (fig 1). There was a trend towards lower ACPA levels at 24 weeks with a statistically significant decrease after 36 weeks (p = 0.015) (fig 1). At 24 weeks the serum levels of IgM-RF and ACPA decreased significantly more than those of their respective antibody classes (p = 0.001 for IgM-RF compared with total serum IgM; p = 0.026 for ACPA compared with total serum IgG levels).

weeks with a statistically significant decrease after 36 weeks (p = 0.015) (fig 1). At 24 weeks the serum levels of IgM-RF and ACPA decreased significantly more than those of their respective antibody classes (p = 0.001 for IgM-RF compared with total serum IgM; p = 0.026 for ACPA compared with total serum IgG levels). Figure 1 Change in levels of IgM-rheumatoid factor (IgM-RF; A) and anti-citrullinated peptide antibodies (ACPA; B) before and during the 36 weeks after initiation of rituximab treatment. The circles represent outliers (values of more than 1.5 box lengths from the upper or lower edge of the box; the box length is the interquartile range). *p<0.05, **p<0.01. Depletion of B cells in peripheral blood CD19+ B cells in PB were undetectable 4 and 16 weeks after rituximab treatment (table 2). Low numbers of B cells could be measured in 3/24 patients at week 4 and 4/24 patients at week 16 (0.01×109/litre). At 24 weeks B cells started to return in 13 of 23 analysed patients (median 0.01×109/litre (IQR 0.00–0.02). The total numbers of T cells and T cell subsets did not change significantly after treatment (table 2).

mbers of B cells could be measured in 3/24 patients at week 4 and 4/24 patients at week 16 (0.01×109/litre). At 24 weeks B cells started to return in 13 of 23 analysed patients (median 0.01×109/litre (IQR 0.00–0.02). The total numbers of T cells and T cell subsets did not change significantly after treatment (table 2). Table 2 Cells in peripheral blood and synovial tissue obtained from patients with rheumatoid arthritis before, 4 and 16 weeks after rituximab treatment Before rituximab 4 weeks after rituximab p Value* 16 weeks after rituximab p Value† Blood lymphocytes: CD19 0.13 (0.09–0.19) <0.01 (0.00–0.00) <0.001 <0.01 (0.00–0.00) <0.001 CD3 1.32 (0.81–1.68) 1.17 (0.90–1.90) 0.78 1.26 (1.00–1.87) 0.49 CD3CD4 0.93 (0.52–1.14) 0.78 (0.63–1.35) 0.64 0.85 (0.69–1.22) 0.66 CD3CD8 0.29 (0.19–0.48) 0.34 (0.20–0.52) 0.51 0.34 (0.22–0.49) 0.38 Cellular markers in synovial tissue: CD22 38 (3–158) 9 (0–39) 0.002 8 (0–40) 0.015 CD3 249 (119–845) 391 (62–1114) 0.65 130 (44–383) 0.010 CD4 403 (2–1702) 78 (15–1128) 0.92 152 (55–659) 0.068 CD8 8 (0–27) 3 (0–6) 0.67 1 (0–10) 0.048 CD138 137 (58–496) 174 (0–623) 0.71 76 (0–213) 0.48 CD68L 293 (93–502) 148 (66–346) 0.043 133 (18–215) 0.001 CD68SL 548 (134–1076) 434 (88–1291) 0.112 191 (52–563) 0.023 Aggregates 8 (0–32) 4 (0–24) 0.078 1 (0–2) 0.007 CD68+ macrophages were analysed separately in the intimal lining layer (L) and synovial sublining (SL). Values for peripheral blood lymphocytes are median (interquartile range (IQR)) 109/litre, for cellular markers in synovium median (IQR) cells/mm2. The total no. of lymphocyte aggregates (aggregates) was counted per section.

acrophages were analysed separately in the intimal lining layer (L) and synovial sublining (SL). Values for peripheral blood lymphocytes are median (interquartile range (IQR)) 109/litre, for cellular markers in synovium median (IQR) cells/mm2. The total no. of lymphocyte aggregates (aggregates) was counted per section. *Wilcoxon signed rank test for paired data between baseline and week 4. †Wilcoxon signed rank test for paired data between baseline and week 16. The effects of rituximab are not limited to synovial B cells: changes in plasma cells, T cells, FDCs and macrophages Out of the total of 24 patients analysed, baseline biopsies of 2 patients did not pass synovial tissue quality control; of these patients only samples taken after 4 and 16 weeks were included in the analysis. In two other patients the biopsy samples taken after 16 weeks and in one patient the biopsy taken after 4 weeks was of insufficient quality to be included in the analysis.

of 2 patients did not pass synovial tissue quality control; of these patients only samples taken after 4 and 16 weeks were included in the analysis. In two other patients the biopsy samples taken after 16 weeks and in one patient the biopsy taken after 4 weeks was of insufficient quality to be included in the analysis. Extension of the study population in the present study confirmed our previous observation,4 showing a highly significant reduction of synovial B cells at 4 weeks, but not in all patients (fig 2, table 2). Similar results were obtained when we used anti-CD19 antibodies to detect synovial B cells by immunohistochemistry (data not shown). There was no statistically significant additional reduction of synovial B cells at 16 weeks on the group level, but there was a clear trend towards more pronounced B cell depletion in 7 of the 15 patients who had persistent B cells at week 4 (median 39 cells/mm2 (IQR 12–89) and 4 (0–22), at respectively 4 and 16 weeks after rituximab infusion). In five patients there was clear persistence of B cells (median 52 (IQR 9–551) and 101 (63–313) at baseline and at 16 weeks, respectively).

cell depletion in 7 of the 15 patients who had persistent B cells at week 4 (median 39 cells/mm2 (IQR 12–89) and 4 (0–22), at respectively 4 and 16 weeks after rituximab infusion). In five patients there was clear persistence of B cells (median 52 (IQR 9–551) and 101 (63–313) at baseline and at 16 weeks, respectively). Figure 2 Change in number of CD22+ B cells (A), CD3+ T cells (B), lymphocyte aggregates (C), intimal (D) and sublining (E) CD68+ macrophages and CD138+ plasma cells (F) in synovial tissue. Biopsies were obtained before, 4 and 16 weeks after initiation of rituximab treatment. The circles represent outliers (values of more than 1.5 box lengths from the upper or lower edge of the box; the box length is the interquartile range). *p<0.05, **p<0.01. Since B cells are precursors of plasma cells, and B cell depletion could indirectly result in a decrease in short-lived plasma cells associated with autoantibody production,19 we examined the consequences of rituximab treatment for plasma cell numbers in the synovium. There was a marked reduction of plasma cells in a subset of the patients (fig 2, table 2). On the group level this change did not reach statistical significance due to the variability of the response.

autoantibody production,19 we examined the consequences of rituximab treatment for plasma cell numbers in the synovium. There was a marked reduction of plasma cells in a subset of the patients (fig 2, table 2). On the group level this change did not reach statistical significance due to the variability of the response. The number of T cells was unaltered at 4 weeks, but there was a significant decrease in T cell numbers at 16 weeks (fig 2, table 2). Consistent with the reduction of synovial B cells and T cells (the major cell populations in the lymphocyte aggregates), we observed a trend towards reduced numbers of lymphocyte aggregates at 4 weeks with a significant decrease of aggregates of all sizes at 16 weeks (fig 2, table 2). FDCs were found in four patients at baseline, but were undetectable in all patients 16 weeks after treatment. Hence, rituximab treatment had a clear effect on lymphocyte aggregates and germinal centres (defined by the presence of FDCs). Of interest, we also found a significant decrease in intimal macrophages at 4 weeks, which was even more pronounced at 16 weeks (fig 2, table 2). In addition, there was a trend towards lower numbers of sublining macrophages at 4 weeks with a statistically significant reduction at 16 weeks (fig 2, table 2). Using repeated measurement methods for analysis of changes in synovial cell populations between baseline and 4 and 16 weeks after treatment yielded the same results as separate paired t test analysis.

er numbers of sublining macrophages at 4 weeks with a statistically significant reduction at 16 weeks (fig 2, table 2). Using repeated measurement methods for analysis of changes in synovial cell populations between baseline and 4 and 16 weeks after treatment yielded the same results as separate paired t test analysis. Persistence of B cells after 16 weeks was correlated with persistence of plasma cells (r = 0.70; p<0.001), T cells (r = 0.69; p<0.001), and lymphocyte aggregates (r = 0.72; p<0.001) at the same timepoint. Changes in synovial plasma cells predict clinical response Since the clinical response to rituximab treatment may be variable, we studied whether changes in synovial cell populations were related to clinical improvement. Of importance, there were no baseline characteristics of the synovium that could significantly predict clinical response to treatment, although there was perhaps a minor trend towards more B cells at baseline in responders compared to non-responders (figs 3 and 4). The decrease in synovial B cells between baseline and 4 weeks or between 4 weeks and 16 weeks was not significantly different between responders versus non-responders to treatment.

eatment, although there was perhaps a minor trend towards more B cells at baseline in responders compared to non-responders (figs 3 and 4). The decrease in synovial B cells between baseline and 4 weeks or between 4 weeks and 16 weeks was not significantly different between responders versus non-responders to treatment. Figure 3 Differences in synovial B cells at baseline (A), respectively changes in B cells after rituximab treatment (B and C), in clinical responders versus non-responders. There was no statistically significant difference in B cell numbers at baseline or in the reduction in synovial B cells between responders and non-responders. The circles represent outliers (values of more than 1.5 box lengths from the upper or lower edge of the box; the box length is the interquartile range). *p<0.05, **p<0.01. Figure 4 Differences in synovial plasma cells (A), intimal macrophages (C), sublining macrophages (E) and T cells (G) at baseline respectively changes in these cells (B, D, F, H) after rituximab treatment in clinical responders versus non-responders. There was no statistically significant difference in these cells at baseline between responders versus non-responders. In light of the kinetics of the changes after treatment (fig 2), we compared the decrease in synovial cell populations other than B cells between 4 weeks and 16 weeks in relationship to clinical response. There was a highly significant difference in reduction of intimal macrophages (p = 0.008), respectively plasma cells (p = 0.002), between responders compared to non-responders with a similar trend for sublining macrophages. The circles represent outliers (values of more than 1.5 box lengths from the upper or lower edge of the box; the box length is the interquartile range). *p<0.05, **p<0.01. Additionally, no changes in other inflammatory cells between baseline and 16 weeks after treatment differed between responders and non-responders, although a trend was found for plasma cells (p = 0.115). As described above, the changes in synovial cells other than B cells were found between 4 and 16 weeks rather than between baseline and 4 weeks, except for intimal macrophages where there was already a reduction at 4 weeks (fig 2). In light of the kinetics of the changes shown in table 2 and fig 2, we compared the decrease in synovial cell populations other than B cells between 4 weeks and 16 weeks in relationship to clinical response (fig 4).

line and 4 weeks, except for intimal macrophages where there was already a reduction at 4 weeks (fig 2). In light of the kinetics of the changes shown in table 2 and fig 2, we compared the decrease in synovial cell populations other than B cells between 4 weeks and 16 weeks in relationship to clinical response (fig 4). Of interest, the change in plasma cells differed significantly between responders and non-responders (p = 0.002) (fig 4B and 5). Within the clinical responder group there was a significant decrease in plasma cells after treatment (p = 0.005), but not so in non-responders. Similarly, there was a significant difference in reduction of intimal macrophages between responders and non-responders (p = 0.008; fig 4D), with a similar trend for sublining macrophages (fig 4F).

esponder group there was a significant decrease in plasma cells after treatment (p = 0.005), but not so in non-responders. Similarly, there was a significant difference in reduction of intimal macrophages between responders and non-responders (p = 0.008; fig 4D), with a similar trend for sublining macrophages (fig 4F). Figure 5 Change in the number of CD138+ plasma cells in representative serial synovial tissue samples obtained at 4 (A and C) and 16 (B and D) weeks after initiation of rituximab treatment. Different patterns of response were identified. In patients who responded to treatment we observed a reduction in plasma cells between 4 and 16 weeks after treatment (compare A and B), while in patients who did not fulfil the response criteria, plasma cells persisted (compare C and D) (Original magnification × 20). Linear regression analysis revealed a significant relationship between the decrease in plasma cell numbers and the decrease in 28-joint Disease Activity Score (DAS28) at week 24. Linear regression analysis was performed to establish the predictive value of the changes in synovial cell populations for the size of the clinical response after 24 weeks. Consistent with the analyses described above, we found a positive correlation between the change in intimal macrophages (r = 0.51; p = 0.04) as well as plasma cells (r = 0.46, p = 0.003) between 4 and 16 weeks on the one hand and the decrease in DAS28 after 24 weeks on the other. Accordingly, linear regression analysis revealed that the decrease in plasma cells between 4 and 16 weeks could predict the decrease in DAS28 at 24 weeks after treatment (R2 = 0.26, p = 0.002). The reduction of macrophages also showed a trend for predicting clinical improvement at week 24 in the univariate analysis (p = 0.051), but when analysed together with the decrease in plasma cells in a multiple linear regression model, it was not an independent predictor of the decrease in DAS28 (p = 0.216, partial correlation 0.399).

macrophages also showed a trend for predicting clinical improvement at week 24 in the univariate analysis (p = 0.051), but when analysed together with the decrease in plasma cells in a multiple linear regression model, it was not an independent predictor of the decrease in DAS28 (p = 0.216, partial correlation 0.399). When linear regression analysis was performed on the combination of changes in plasma cells between 4 and 16 weeks and the baseline DAS28 using multiple linear regression analysis, the decrease in plasma cells continued to predict the decease in DAS28 at week 24 (p = 0.034), whereas the baseline DAS28 alone could not predict the clinical response (p = 0.623). Interestingly, the change in plasma cell numbers was also correlated with a decrease in the ACPA levels at week 16 (r = 0.52, p = 0.03). When we used the change in synovial cells between baseline and 16 weeks, similar trends were observed. As noted above and shown in table 2 and fig 2, the major changes in synovial cell populations other than B cells were observed between 4 and 16 weeks after treatment, secondary to the changes in synovial B cells. Collectively, the results suggest that the clinical response can be predicted by changes in cell types other than B cells, especially the number of synovial plasma cells that are derived from B cells. This change is also correlated to the reduction in serum ACPA levels.

When we used the change in synovial cells between baseline and 16 weeks, similar trends were observed. As noted above and shown in table 2 and fig 2, the major changes in synovial cell populations other than B cells were observed between 4 and 16 weeks after treatment, secondary to the changes in synovial B cells. Collectively, the results suggest that the clinical response can be predicted by changes in cell types other than B cells, especially the number of synovial plasma cells that are derived from B cells. This change is also correlated to the reduction in serum ACPA levels. DISCUSSION The results presented here confirm the previously reported variable tissue response of B cells after rituximab treatment, in contrast to the nearly complete depletion of B cells in the peripheral blood. Moreover, this study shows for the first time the secondary effects on cell populations other than synovial B cells, supporting the concept that B cells orchestrate synovial inflammation. In particular the change in short-lived synovial plasma cells (derived from B cells) between 4 and 16 weeks after initiation of treatment is related to the clinical response over time. These findings are consistent with the kinetics of the gradual clinical response and the slow but sure decrease in levels of circulating antibodies observed after rituximab treatment.

plasma cells (derived from B cells) between 4 and 16 weeks after initiation of treatment is related to the clinical response over time. These findings are consistent with the kinetics of the gradual clinical response and the slow but sure decrease in levels of circulating antibodies observed after rituximab treatment. B cells may drive the inflammatory processes involved in RA by different mechanisms. First, B cells may drive synovial inflammation by production of autoantibodies;20 they are the precursors of short-lived plasma cells associated with production of autoantibodies, such as IgM-RF and ACPA. Second, B cells are effective antigen-presenting cells and activators of T cells.21 Third, B cells may promote synovial inflammation by producing pro-inflammatory cytokines and chemokines.22–24 Thus, depletion of B cells could interfere with different mechanisms involved in the pathogenesis.

s, such as IgM-RF and ACPA. Second, B cells are effective antigen-presenting cells and activators of T cells.21 Third, B cells may promote synovial inflammation by producing pro-inflammatory cytokines and chemokines.22–24 Thus, depletion of B cells could interfere with different mechanisms involved in the pathogenesis. The results from the present study show that rituximab treatment may indeed deplete B cells at the primary site of inflammation, the synovium, although there is persistence of synovial B cells in a subset of patients. The discrepancy with the complete B cell depletion observed in peripheral blood in nearly all patients might be explained by the expression of protective factors in the tissue, such as BLyS25 and CD55 (decay-accelerating factor (DAF)),26 as well as the requirement for B cells to access the circulation for efficient depletion. This difference underscores the importance of analysis of different compartments to understand the effects of treatment, as has also been shown after, for example, Campath-1H treatment.27 The variable response in the synovial tissue with regard to B cell depletion suggests that the standard therapeutic regimen is perhaps not optimal in all patients. Of note, persistence of synovial B cells was related to persistence of plasma cells. Future studies need to address the question whether it is possible to induce clinical improvement in non-responders with persistent B cells and plasma cells in the synovium. Conceivably, a subset of patients would benefit from a more intense dosing schedule. It is also possible that persistence of plasma cells in non-responders is related to the presence of long-lived plasma cells in the synovium of a subset of patients. Plasma cells with different longevity could be induced by mechanisms such as epitope spreading. For these patients alternative approaches may be considered, interfering with, for example, APRIL (a proliferation-inducing ligand) and B lymphocyte stimulator (BLyS). The present study strengthens the rationale for evaluating changes in biomarkers after targeted therapies interfering with B cells and plasma cells to further optimise the clinical response in the context of personalised medicine.

for example, APRIL (a proliferation-inducing ligand) and B lymphocyte stimulator (BLyS). The present study strengthens the rationale for evaluating changes in biomarkers after targeted therapies interfering with B cells and plasma cells to further optimise the clinical response in the context of personalised medicine. The relationship between the change in plasma cells and clinical improvement suggests that rituximab exerts its effects at least in part by an indirect effect on short-lived autoreactive plasma cells that are associated with the production of autoantibodies. Consistent with these results, there was a reduction in RF and ACPA levels after treatment; the reduction in plasma cells was directly related to the decrease in ACPA levels at week 16. These data are also in agreement with a previous study showing a trend towards lower synovial immunoglobulin synthesis 2 months after rituximab treatment, although that study was not powered to detect statistically significant changes.5 A role for autoantibodies in RA has been suggested since the discovery of RF in RA and regained interest in the late nineties.20 The previously reported decrease in RF and ACPA levels after rituximab treatment relative to minor changes in total immunoglobulins suggests a role for short-lived plasma cells in their production.2 3 28 This notion is supported by observations in a mouse model; rheumatoid factor transgenic mice were crossed with mice of the autoimmune-prone MRL/lpr strain, after which a spontaneous rheumatoid factor response developed. This response was mediated by continuous generation of short-lived plasmablasts.19 In addition, it has been suggested that small immune complexes containing autoantibodies may drive synovial inflammation by triggering Fcγ receptor IIIA, which is expressed by intimal macrophages.29 The results presented here would support the concept that autoantibody production by B cells and plasma cells is critically involved in promoting synovial inflammation.

mmune complexes containing autoantibodies may drive synovial inflammation by triggering Fcγ receptor IIIA, which is expressed by intimal macrophages.29 The results presented here would support the concept that autoantibody production by B cells and plasma cells is critically involved in promoting synovial inflammation. Depletion of B cells did not only indirectly result in a decrease in synovial plasma cells, but there was also an effect on other major cell populations, such as T cells and macrophages. This indicates that B cells have an important role in sustaining the inflammatory cell infiltrate in the rheumatoid synovium. The decrease in synovial T cells and the disruption of lymphocyte aggregates and germinal centres, as shown by the disappearance of FDCs, supports the hypothesis that B cells influence T cell activity and organisation in the synovial tissue. Lymphocyte aggregates could be disrupted by the absence of B cell derived factors such as lymphotoxin beta.30 However, it is quite likely that the explanation is more complex, since previous work has shown that, in contrast to large follicles, relatively small lymphocyte aggregates usually contain very few B cells.31 Interestingly, B cell depletion also diminished macrophage infiltration, which is in agreement with the concept that there is a consistent relationship between clinical improvement and changes in synovial macrophages, independent of the primary mechanism of action of the treatment.32 33 Together, the change in T cells and macrophages could be explained by an indirect effect of B cell depletion on the expression of proinflammatory cytokines and chemokines involved in cell migration and retention, although this remains to be shown.

s, independent of the primary mechanism of action of the treatment.32 33 Together, the change in T cells and macrophages could be explained by an indirect effect of B cell depletion on the expression of proinflammatory cytokines and chemokines involved in cell migration and retention, although this remains to be shown. In conclusion, rituximab treatment results in a variable response on synovial B cells with secondary changes in numbers of other inflammatory cells, leading to diminished synovial inflammation. There is a direct relationship between the decrease in synovial plasma cells and clinical improvement over time. We would like to thank Dr Tom J M Smeets, Desiree Pots and Marjolein Vinkenoog for expert technical support, as well as the research nurses Margot Colombijn and Angelina Roelse. Competing interests: PPT has served as a consultant to Roche. Funding: DMG was supported by the Dutch Arthritis Association (Reumafonds). This research was supported by the European Community’s FP6 funding (Autocure). This publication reflects only the views of the authors; the European Community is not liable for any use that may be made of the information herein. Ethics approval: The study protocol was approved by the Medical Ethics Committee of the Academic Medical Center/University of Amsterdam and all patients gave written informed consent before participation in the study.

Gout is a disorder of purine metabolism characterised by acute, recurrent attacks of crystal arthritis. The single most important risk factor for developing gout is a raised level of serum uric acid (sUA), with supersaturation of uric acid in the extracellular fluid resulting in the precipitation of urate crystals.1 The 5-year cumulative risk of developing gout is 30.5% in men with an sUA level ⩾590 μmol/l (⩾10 mg/dl) and only 0.6% in those with an sUA level <420 μmol/l (<7.0 mg/dl).2 Deposition of urate crystals in the articular, periarticular and subcutaneous tissues3 results in episodes of acute arthritis (usually initially affecting the metatarsophalangeal joints) and the development of tophi.4 In addition, deposition of urate crystals in the renal tract may lead to impaired renal function. The definitive diagnosis of gout depends on identifying monosodium urate crystals in fluid from aspiration of an acutely affected joint. However, joint aspiration and crystal identification by polarising light microscopy is generally regarded as a specialist procedure which is seldom undertaken in general practice, where most patients with gout are diagnosed and treated.5 6 The diagnosis is usually made clinically, and often in retrospect, based on clinical symptoms described by patients, and on response to treatment.

light microscopy is generally regarded as a specialist procedure which is seldom undertaken in general practice, where most patients with gout are diagnosed and treated.5 6 The diagnosis is usually made clinically, and often in retrospect, based on clinical symptoms described by patients, and on response to treatment. The principal goal of treatment in chronic gout is to prevent crystal formation and promote crystal dissolution. Recent recommendations from the European League Against Rheumatism (EULAR)7 state that this is most effectively achieved by reducing and maintaining sUA levels below 360 μmol/l (6 mg/dl). Several studies provide evidence that reduced sUA is associated with reduced frequency of gout flares and reduction in tophi size.3 4 8 9 Effective management requires long-term administration of uric acid lowering treatment, initially guided by, and then monitored through, regular sUA testing. A recent study in the UK demonstrated the overall prevalence of gout to be 1.4%. Gout prevalence increased with age and was much higher among men.10 That study apart, however, contemporary data on the prevalence of gout in Europe are lacking, possibly because the episodic and chronic nature of the condition makes these data difficult to collect. We therefore undertook a study to investigate the prevalence of gout, its comorbidities, and its current clinical management in general practice in the UK and Germany.

A recent study in the UK demonstrated the overall prevalence of gout to be 1.4%. Gout prevalence increased with age and was much higher among men.10 That study apart, however, contemporary data on the prevalence of gout in Europe are lacking, possibly because the episodic and chronic nature of the condition makes these data difficult to collect. We therefore undertook a study to investigate the prevalence of gout, its comorbidities, and its current clinical management in general practice in the UK and Germany. METHODS A retrospective analysis was conducted using the IMS Disease Analyzer, a longitudinal database containing anonymised patient records maintained by 650 general practitioners (GPs) treating 2.5 million patients in the UK and 400 GPs and internists treating 2.4 million patients in Germany. Patients were included in the analysis if they had had a consultation with a diagnosis of gout (International Classification of Diseases (ICD-10) code M10 or “gout” written in the patient notes) between January 2000 and June 2005, and there was at least one additional recording of gout in their history (consultation with a diagnosis of gout or prescription for gout treatment). Patients were also required to have at least 24 months of recorded data before, and 18 months after, their index date (defined as their first consultation for gout between January 2000 and June 2005). Patients were excluded from the analysis if they were under 18 years of age or had been diagnosed with cancer. The data analysed included comorbidities, medication use, sUA level and frequency of flares in patients with gout.

Patients were included in the analysis if they had had a consultation with a diagnosis of gout (International Classification of Diseases (ICD-10) code M10 or “gout” written in the patient notes) between January 2000 and June 2005, and there was at least one additional recording of gout in their history (consultation with a diagnosis of gout or prescription for gout treatment). Patients were also required to have at least 24 months of recorded data before, and 18 months after, their index date (defined as their first consultation for gout between January 2000 and June 2005). Patients were excluded from the analysis if they were under 18 years of age or had been diagnosed with cancer. The data analysed included comorbidities, medication use, sUA level and frequency of flares in patients with gout. The following definitions were used for the purposes of identifying comorbidities: renal insufficiency (ICD-10 codes K767, N170, N171, N172, N178, N179, N180, N188, N189, N190 or I120; or a serum creatinine level of >150 μmol/l (>2 mg/dl)); alcoholism (ICD-10 codes F10, Y91, Y90 or K70); diabetes (a prescription for any antidiabetic drug; or ICD-10 codes E10, E11, E12, E13 or E14); heart failure (ICD-10 code I500, I501 or I509); hypertension (ICD-10 codes I100, I150, I152, I158, I159 or I270); myocardial infarction (ICD-10 codes I210, I211, I212, I213, I214, I219, I220, I221, I228, I229 or I252); and obesity (body mass index (BMI) >30 kg/m2; or ICD-10 codes E660, E661, E662, E668 or E669).

E14); heart failure (ICD-10 code I500, I501 or I509); hypertension (ICD-10 codes I100, I150, I152, I158, I159 or I270); myocardial infarction (ICD-10 codes I210, I211, I212, I213, I214, I219, I220, I221, I228, I229 or I252); and obesity (body mass index (BMI) >30 kg/m2; or ICD-10 codes E660, E661, E662, E668 or E669). Persistence was defined as the total time that any patient was prescribed a given drug, from initiation of treatment to the end of the last supplied prescription, without intervening discontinuation of that drug. Compliance was defined as the percentage of the prescribed doses that could actually have been taken by a patient during the period of persistence. Compliance was assessed by dividing the sum of the number of days of medication supplied by the duration of treatment while the patient was persistent (ie, the medication possession ratio).11 A gout flare was defined as either a visit resulting in a prescription for treating acute gout (short-term prescription for non-steroidal anti-inflammatory drugs (NSAIDs), colchicine or corticosteroids) or a gout-related emergency-room visit or admission to hospital.

Compliance was defined as the percentage of the prescribed doses that could actually have been taken by a patient during the period of persistence. Compliance was assessed by dividing the sum of the number of days of medication supplied by the duration of treatment while the patient was persistent (ie, the medication possession ratio).11 A gout flare was defined as either a visit resulting in a prescription for treating acute gout (short-term prescription for non-steroidal anti-inflammatory drugs (NSAIDs), colchicine or corticosteroids) or a gout-related emergency-room visit or admission to hospital. Mean and standard deviation were determined for all variables during the analysis. A logistic regression model was applied to investigate the relationship between sUA and frequency of gout flares. The outcome assessed was the total number of recorded flares during the observation period (±3–4 years), in relation to the maximum observed sUA level in that period. Persistence was analysed using Cox regression analysis, which models the risk of treatment discontinuation over time.

n sUA and frequency of gout flares. The outcome assessed was the total number of recorded flares during the observation period (±3–4 years), in relation to the maximum observed sUA level in that period. Persistence was analysed using Cox regression analysis, which models the risk of treatment discontinuation over time. RESULTS Gout prevalence During the observation period, the IMS Disease Analyzer contained 2 514 806 and 2 402 185 patient records in the UK and Germany, respectively. Overall, there were 34 071 patients in the UK and 34 797 in Germany who were reported to have gout, indicating a prevalence of 1.4% in both countries. On the basis of the inclusion and exclusion criteria, 7443 patients in the UK and 4006 in Germany were entered into the analysis (table 1). Over 80% of the study population was male in each of the countries. The mean age of the patients sampled was 66 years and 63 years in the UK and Germany, respectively, and the mean duration of gout history was 81 and 67 months, respectively. Table 1 Study population characteristics Characteristics UK(n = 7443) Germany(n = 4006) Male patients, No (%) 6074 (81.6) 3222 (80.4) Current age (years) 65.6 (13.8) 63.1 (13.1) Age at diagnosis (years) 61.6 (13.9) 58.6 (13.1) Duration of gout (months) 81.4 (66.7) 67.4 (28.6) Time of follow-up (months) 46.2 (14.4) 51.0 (14.5) Results are shown as mean (SD) unless stated otherwise.

7443) Germany(n = 4006) Male patients, No (%) 6074 (81.6) 3222 (80.4) Current age (years) 65.6 (13.8) 63.1 (13.1) Age at diagnosis (years) 61.6 (13.9) 58.6 (13.1) Duration of gout (months) 81.4 (66.7) 67.4 (28.6) Time of follow-up (months) 46.2 (14.4) 51.0 (14.5) Results are shown as mean (SD) unless stated otherwise. Pre-index comorbidities Renal insufficiency was reported in 9.5% of the study population in the UK and 4.8% in Germany. Figure 1 shows the prevalence of other comorbidities within the study population. In the UK study population, the most common comorbidity recorded was obesity (27.7%), while in Germany it was diabetes (25.9%). Hypertension was a common comorbidity in both study populations: 18.5% of the German population sampled and 17.5% of the UK patients had at least one record of hypertension during the study period. Heart failure and myocardial infarction were reported in 7.1% and 7.4% of the UK study population, and 10.8% and 5.8% of the study population in Germany, respectively. Alcoholism was recorded as a diagnosis in 3% of UK patients with gout and 0.4% of the German patients with gout. There was a trend for the prevalence of each comorbidity to increase with sUA level (fig 1). Figure 1 Pre-index comorbidities in (A) the UK study population and (B) the German study population. sUA, serum uric acid. Conversion: sUA (mg/dl) ×59.48  =  μmol/l. Drug use before and after the index consultation Figure 2 shows drug use before and after the index consultation in the UK and German study populations.

Pre-index comorbidities Renal insufficiency was reported in 9.5% of the study population in the UK and 4.8% in Germany. Figure 1 shows the prevalence of other comorbidities within the study population. In the UK study population, the most common comorbidity recorded was obesity (27.7%), while in Germany it was diabetes (25.9%). Hypertension was a common comorbidity in both study populations: 18.5% of the German population sampled and 17.5% of the UK patients had at least one record of hypertension during the study period. Heart failure and myocardial infarction were reported in 7.1% and 7.4% of the UK study population, and 10.8% and 5.8% of the study population in Germany, respectively. Alcoholism was recorded as a diagnosis in 3% of UK patients with gout and 0.4% of the German patients with gout. There was a trend for the prevalence of each comorbidity to increase with sUA level (fig 1). Figure 1 Pre-index comorbidities in (A) the UK study population and (B) the German study population. sUA, serum uric acid. Conversion: sUA (mg/dl) ×59.48  =  μmol/l. Drug use before and after the index consultation Figure 2 shows drug use before and after the index consultation in the UK and German study populations. Figure 2 Drug use before and after the index consultation in (A) the UK study population and (B) the German study population. NSAIDs, non-steroidal anti-inflammatory drugs. In the UK, 63% and in Germany 84.5% of patients with gout received treatment (fig 2). Among these patients, allopurinol was prescribed for most patients in both countries (89% UK; 93% Germany). Colchicine was used much less frequently (16% UK; 15% Germany). In the UK, the use of probenecid and sulfinpyrazone was limited to <1% of patients receiving gout pharmacological treatment. Benzbromarone was prescribed to <3% of patients in Germany. In addition to chronic gout treatment, 89.4% (UK) and 80.3% (Germany) of patients received prescriptions for oral NSAIDs as prophylaxis.

n the UK, the use of probenecid and sulfinpyrazone was limited to <1% of patients receiving gout pharmacological treatment. Benzbromarone was prescribed to <3% of patients in Germany. In addition to chronic gout treatment, 89.4% (UK) and 80.3% (Germany) of patients received prescriptions for oral NSAIDs as prophylaxis. Allopurinol was prescribed at an average daily dose of >200 mg or ⩽300 mg in 63.3% of patients in the UK and 65.7% of those in Germany (fig 3). Average daily doses in the range of 50–100 mg were prescribed in 21% of patients in the UK and 22.4% of those in Germany, and doses >300 mg/day were not often used (2.1% of patients in the UK and 3.4% of patients in Germany). Figure 3 Average dosages of allopurinol prescribed in (A) the UK and (B) Germany. In the UK, persistence with allopurinol (expressed as the percentage of patients receiving treatment for fixed time points) was 92.3% at 28 days, 76.3% at 100 days and 61.2% by 360 days (fig 4A). In Germany, persistence with allopurinol treatment was 98.0% at 28 days, 54.3% at 100 days and 31.0% by 360 days (fig 4B). Apparent rates of compliance while receiving treatment were 93% in the UK and 96% in Germany. Figure 4 Persistence with allopurinol treatment at 24 months in (A) the UK study population and (B) the German study population. sUA testing and gout flares sUA testing was conducted during the observation period in 14% of patients in the UK and 9% of those in Germany. The proportion of patients who had one or more sUA tests performed annually during the observation period was <1% in both countries.

opulation and (B) the German study population. sUA testing and gout flares sUA testing was conducted during the observation period in 14% of patients in the UK and 9% of those in Germany. The proportion of patients who had one or more sUA tests performed annually during the observation period was <1% in both countries. Overall, 72% of the study population in the UK and 41% in Germany experienced at least one gout flare during the observation period (table 2; 45.8%+26.2% and 17.2%+24.1%, respectively).The mean number of gout flares per patient was 1.7 in the UK and 2.9 in Germany (table 2). The frequency of gout flares was positively correlated with the level of sUA (table 2). Table 2 Frequency of gout flares during the observation period. The observation period was unique for each individual patient and depended on when the index date occurred within the period January 2000–June 2005 All patients sUA* >7 mg/dl sUA >8 mg/dl sUA >9 mg/dl sUA >10 mg/dl UK Germany UK Germany UK Germany UK Germany UK Germany Patients with one gout flare during their observation period (%) 45.8 17.2 – – 50.3 21.4 53.0 22.3 41.9 22.4 Patients with ⩾2 gout flares during their observation period (%) 26.2 24.1 – – 42.8 33.3 47.0 31.3 58.1 30.2 Mean number of gout flares/patient 1.7 2.9 – – 1.8 3.1 1.9 3.3 1.9 3.8 Mean number of gout flares/patient/year of observation period 0.38 0.68 0.42 0.65 0.43 0.67 0.42 0.91 0.58 1.97 Conversion: sUA (mg/dl) ×59.48  =  μmol/l. sUA, serum uric acid.

Table 2 Frequency of gout flares during the observation period. The observation period was unique for each individual patient and depended on when the index date occurred within the period January 2000–June 2005 All patients sUA* >7 mg/dl sUA >8 mg/dl sUA >9 mg/dl sUA >10 mg/dl UK Germany UK Germany UK Germany UK Germany UK Germany Patients with one gout flare during their observation period (%) 45.8 17.2 – – 50.3 21.4 53.0 22.3 41.9 22.4 Patients with ⩾2 gout flares during their observation period (%) 26.2 24.1 – – 42.8 33.3 47.0 31.3 58.1 30.2 Mean number of gout flares/patient 1.7 2.9 – – 1.8 3.1 1.9 3.3 1.9 3.8 Mean number of gout flares/patient/year of observation period 0.38 0.68 0.42 0.65 0.43 0.67 0.42 0.91 0.58 1.97 Conversion: sUA (mg/dl) ×59.48  =  μmol/l. sUA, serum uric acid. In the UK, patients with an sUA of 360–420 μmol/l (6–7 mg/dl) were 1.33 times more likely to experience a gout flare than those in whom sUA was <360 μmol/l (<6 mg/dl; odds ratio (OR) = 1.33; p = NS), while those with an sUA level ⩾530 μmol/l (⩾9 mg/dl) were more than twice as likely (OR = 2.15; p<0.01) (table 3). In Germany, patients with an sUA level of 360–420 μmol/l (6–7 mg/dl) were 1.37 times more likely to experience a gout flare than those in whom sUA was <360 μmol/l (<6 mg/dl; OR = 1.37; p = NS), and those with an sUA level ⩾530 μmol/l (⩾9 mg/dl) were more than twice as likely to experience a gout flare (OR = 2.48; p<0.01) (table 3).

atients with an sUA level of 360–420 μmol/l (6–7 mg/dl) were 1.37 times more likely to experience a gout flare than those in whom sUA was <360 μmol/l (<6 mg/dl; OR = 1.37; p = NS), and those with an sUA level ⩾530 μmol/l (⩾9 mg/dl) were more than twice as likely to experience a gout flare (OR = 2.48; p<0.01) (table 3). Table 3 Association between serum uric acid (sUA) level and the number of flares sUA level* (mg/dl) UK Germany Odds ratio vs sUA<6 mg/dl (CI) p Value Odds ratio vs sUA<6 mg/dl (CI) p Value 6–7 1.33 (0.92 to 1.94) NS 1.37 (0.91 to 2.05) NS >7–8 1.49 (1.21 to 2.42) <0.01 1.65 (1.17 to 2.33) <0.01 >8–9 1.71 (1.04 to 2.13) <0.01 2.37 (1.67 to 3.36) <0.01 ⩾9 2.149 (1.53 to 3.01) <0.01 2.48 (1.77 to 3.49) <0.01 *Conversion: sUA (mg/dl) ×59.48  =  μmol/l. DISCUSSION This study was undertaken to augment the limited data available on the epidemiology of gout and its management in practice in Europe. The study dealt exclusively with patients with a clinical diagnosis of gout but not patients with asymptomatic hyperuricaemia. The overall prevalence of gout of 1.4% was remarkably similar in two large IMS databases, each containing about 2.5 million patient records from the years 2000–5 in Germany and the UK. The overall prevalence was also remarkably similar to that found previously by Mikuls et al10 in the records of 1.8 million patients in the UK General Practice Research Database (GPRD) for the years 1990–9, using similar case definitions.10

about 2.5 million patient records from the years 2000–5 in Germany and the UK. The overall prevalence was also remarkably similar to that found previously by Mikuls et al10 in the records of 1.8 million patients in the UK General Practice Research Database (GPRD) for the years 1990–9, using similar case definitions.10 The consistency of these observations on the prevalence of gout consultations in Germany and the UK would seem to provide a measure of confidence in the robustness of estimates of the true prevalence of gout in primary care settings in Northern Europe. It is noteworthy, however, that the prevalence of gout was not greater in the current study, than in the UK GPRD for 1990–9,10 despite the fact that the mean ages of the patients with gout in the UK and German databases for 2000–5 were 6 years and 3 years older, respectively. The prevalence rates observed are, however, slightly higher than those found by Harris et al (0.95%) in 40 UK general practices in 199312 and fivefold higher than those observed (0.26%) in the survey by Currie in UK general practices 30 years ago.13 There are other data which suggest that the prevalence of gout might have increased over the past 30 years. The annual prevalence of self-reported gout in the National Health Interview Survey in the USA trebled between 1969 and 1996,14 and the prevalence of gout and hyperuricaemia requiring urate-lowering drug treatment was observed to have increased by 80% between 1990 and 1999 in a managed-care population in the USA.15 However, Mikuls et al showed that the overall incidence of consultations for gout remained relatively stable throughout the 1990s,10 suggesting that any increase in prevalence might be largely attributable to the changing age structure of the population. Nevertheless, gout is the most common cause of inflammatory arthritis in men over the age of 40 years,16 and the overall prevalence of gout is notably higher than that of rheumatoid arthritis, which was recently estimated to be 1.16% in women and 0.44% in men across the UK population.17

age structure of the population. Nevertheless, gout is the most common cause of inflammatory arthritis in men over the age of 40 years,16 and the overall prevalence of gout is notably higher than that of rheumatoid arthritis, which was recently estimated to be 1.16% in women and 0.44% in men across the UK population.17 As expected, the prevalence of comorbidities was high among patients with gout in both the UK and Germany. The comorbidities found to be most commonly associated with gout in this study were obesity in the UK and diabetes in Germany. Obesity was twice as common (27.7%) in UK patients as in German patients (13.3%), while diabetes was recorded as a comorbid diagnosis three times as often in the German patients (25.9%) as in those in the UK GPRD (8.3%). Similar prevalence rates of obesity have been estimated previously in the UK and Germany,18 and the somewhat surprising difference reported in this study may be attributable, in part, to the use of the IMS Disease Analyzer database. In the UK, GPs in the IMS Disease Analyzer panel can record items such as BMI in addition to using ICD-10 codes for obesity, but in Germany only the ICD-10 codes were considered in the study definition of obesity as there was no other electronically coded way for a GP to record obesity. Therefore, it is possible that British GPs were more likely to capture information electronically about obesity than those in Germany. The greater frequency with which diabetes was recorded in Germany is consistent with previous estimates of prevalence rates in Europe, which were more than twice as high in Germany as those in the UK.19

possible that British GPs were more likely to capture information electronically about obesity than those in Germany. The greater frequency with which diabetes was recorded in Germany is consistent with previous estimates of prevalence rates in Europe, which were more than twice as high in Germany as those in the UK.19 Over 60 years ago, Brochner-Mortensen observed that 78% of a group of 100 Scandinavian patients with gout were more than 10% overweight and 57% were more than 30% overweight.20 Grahame and Scott found that 48% of a UK cohort of 354 patients with gout attending a hospital clinic were more than 15% overweight.21 Obesity was not recorded as a specific comorbidity in the study by Mikuls et al.10 However, recently published data demonstrated that the prevalence of abdominal obesity and metabolic syndrome is as high as 62.9% in people with gout identified in the US Third National Health and Nutrition Survey (NHANES III),22 and diabetes was a comorbid diagnosis in 19.9% of 9482 patients with gout in a North American study of patients in a managed-care setting.23 It is therefore interesting and surprising that Mikuls et al also found the prevalence of diabetes to be remarkably low (6.4%) in patients with gout in their analysis from the UK GPRD for 1990–9.10 In that study, it was found that coronary artery disease and hypertension were the most common comorbidities in patients with gout (24.9% and 23.9%, respectively).10 The prevalence of hypertension among patients with gout was found to be as high as 59.7% in the North American managed-care population.23

RD for 1990–9.10 In that study, it was found that coronary artery disease and hypertension were the most common comorbidities in patients with gout (24.9% and 23.9%, respectively).10 The prevalence of hypertension among patients with gout was found to be as high as 59.7% in the North American managed-care population.23 There is some evidence to suggest that gout and hyperuricaemia may be independent risk factors for cardiovascular disease.24 25 The clinical data demonstrating the association of gout with obesity, hypertension, excessive alcohol consumption and metabolic syndrome is now overwhelming and incontrovertible.22 23 25 26 As these comorbidities are themselves major risk factors for cardiovascular disease, every diagnosis of gout should act as a “red flag” to alert doctors to assess patients for cardiovascular risk.7 22 27

ertension, excessive alcohol consumption and metabolic syndrome is now overwhelming and incontrovertible.22 23 25 26 As these comorbidities are themselves major risk factors for cardiovascular disease, every diagnosis of gout should act as a “red flag” to alert doctors to assess patients for cardiovascular risk.7 22 27 A similar proportion of patients were found to be receiving prescriptions for diuretic drugs in the UK and Germany in this study (∼ 30%), and by Mikuls et al.10 Diuretics, especially thiazide diuretics, can increase the sUA level as a consequence of volume depletion, increased renal tubular reabsorption of urate, stimulation of the urate-anion exchanger (URAT-1) and a decrease in renal urate excretion.28 Diuretic use has been found to be a risk factor for incident gout (relative risk (RR) = 1.77; 95% confidence interval 1.42 to 2.20) which is apparently independent of hypertension in the 12 years of follow-up of a cohort of men who had no previous history of gout in the health professionals follow-up study.26 However, a recent case–control study found no evidence that diuretics were a risk factor for the development of gout, and suggested that the cardiovascular conditions for which they were prescribed, which are themselves risk factors for gout, might have confounded previous inferences.29 This would imply that continuing flares of gout in these patients were not primarily attributable to continuing treatment with diuretics, but rather were a consequence of failure to treat hyperuricaemia adequately.

were prescribed, which are themselves risk factors for gout, might have confounded previous inferences.29 This would imply that continuing flares of gout in these patients were not primarily attributable to continuing treatment with diuretics, but rather were a consequence of failure to treat hyperuricaemia adequately. Allopurinol was by far the most commonly used drug for treating patients with chronic and interval gout in both the UK and Germany, accounting for about 90% of all such medication. However, despite the clear preference for allopurinol as the preferred uric acid lowering drug, the findings from the current study show that the management of patients with gout is suboptimal. A large proportion of patients with gout continued to have raised sUA levels and recurrent gout flares. Persistence of treatment with allopurinol was remarkably low in both the UK and Germany, but this study provides no information about why this should have been the case.

anagement of patients with gout is suboptimal. A large proportion of patients with gout continued to have raised sUA levels and recurrent gout flares. Persistence of treatment with allopurinol was remarkably low in both the UK and Germany, but this study provides no information about why this should have been the case. Until recently, the importance of follow-up sUA testing to guide treatment decisions in patients diagnosed with gout has not been emphasised or well defined by research evidence. However, there is now evidence from comparative randomised controlled trials,3 30 and from an observational study in a managed-care setting,1 that fewer than 50% of patients receiving allopurinol, in the most commonly prescribed dose of 300 mg daily, achieve optimum reductions in plasma urate concentrations. Recent evidence-based, consensus guidelines from EULAR7 emphasise the importance of ensuring that the sUA level is maintained below 360 μmol/l (6 mg/dl), but those from the British Society for Rheumatology27 recommend maintaining the sUA level below 300 μmol/l (5 mg/dl). Despite a lack of evidence for the optimum frequency for monitoring the sUA level, recent guidelines recommend 3-monthly measurements of sUA in the first year after the start of treatment with uric acid lowering drug treatment followed by annual measurements.27 The frequency of sUA testing was found to be extremely low in this study; sUA evaluation was performed in only 14% of the UK and 9% of the German study populations, and fewer than 5% of patients had two sUA tests performed in the observation period in either country. It might be argued that it is acceptable for the sUA level to be monitored less frequently in patients with good symptomatic control; however, the study clearly showed that even patients who continued to experience gout flares were not having regular sUA checks.

ts performed in the observation period in either country. It might be argued that it is acceptable for the sUA level to be monitored less frequently in patients with good symptomatic control; however, the study clearly showed that even patients who continued to experience gout flares were not having regular sUA checks. Seventy-two percent of the patients with gout in the UK, and over 40% of those in Germany, experienced at least one gout flare, and the frequency of gout flares was significantly correlated with sUA levels. Previous studies have demonstrated the presence of urate crystals in the joints of patients with asymptomatic gout,31 but reduction of the sUA below 360 μmol/l (6 mg/dl) leads to decreased intra-articular crystal deposition8 and more rapid reduction of tophi3 as well as a reduction in the frequency of gout flares.9 This suggests that maintaining the sUA level below 360 μmol/l (6 mg/dl) could prevent recurrent attacks of acute gout in the long term.1 9 Indeed, it would be of interest to undertake further research to investigate the possibility of a continuous relationship between sUA levels and the risk of crystal deposition, analogous to the relationship that has been demonstrated between cardiovascular risk and measures such as serum cholesterol levels and blood pressure.

it would be of interest to undertake further research to investigate the possibility of a continuous relationship between sUA levels and the risk of crystal deposition, analogous to the relationship that has been demonstrated between cardiovascular risk and measures such as serum cholesterol levels and blood pressure. This study has some limitations. Although the primary care consultation data provided by the IMS databases are an important source of information on the prevalence of gout, associated comorbidities and treatment, and can provide age and sex standardised estimates of the number of new consultations, the IMS Disease Analyzer does not contain community-based data that would allow for the measurement of population-wide incidence rates of gout in different countries. Although the IMS database populations in Germany and the UK closely mirror the populations at large in their male to female ratio, social class and geographical distribution, it is possible that not all practice types are represented.

or the measurement of population-wide incidence rates of gout in different countries. Although the IMS database populations in Germany and the UK closely mirror the populations at large in their male to female ratio, social class and geographical distribution, it is possible that not all practice types are represented. Possibly, the study might have underestimated the prevalence of gout in Germany, as several patients in the German database were excluded from the analysis because they only had a single recording of gout in their notes. Second attacks might have been under-recorded in the German database because of the free choice of healthcare that operates in Germany, where the doctor responsible for the patient’s records might not be the doctor consulted for a subsequent gout flare. There is also the possibility that the study underestimated the frequency of gout flares, because the option to self-medicate means not all gout flares would necessarily have been detailed in a patient’s records. Any such effect did not, however, prevent the study from demonstrating a correlation between sUA level and gout flares.

o the possibility that the study underestimated the frequency of gout flares, because the option to self-medicate means not all gout flares would necessarily have been detailed in a patient’s records. Any such effect did not, however, prevent the study from demonstrating a correlation between sUA level and gout flares. Another important question that was not dealt with was whether the doses of allopurinol were optimum in those patients with gout found to have a high sUA level and frequent gout flares. The study showed that 97.9% of patients prescribed allopurinol in the UK and 96.6% of those in Germany received doses of ⩽300 mg/day (fig 3), but how this was related to outcomes was not investigated. Neither did the study evaluate whether allopurinol doses were adjusted according to renal function to minimise the risk of toxicity in patients with renal impairment.27 32 The relationship between flares and the recommended7 27 prophylactic use of low-dose colchicine or NSAIDs to minimise the risk of gout flares following the initiation of urate-lowering treatment was not investigated.

Another important question that was not dealt with was whether the doses of allopurinol were optimum in those patients with gout found to have a high sUA level and frequent gout flares. The study showed that 97.9% of patients prescribed allopurinol in the UK and 96.6% of those in Germany received doses of ⩽300 mg/day (fig 3), but how this was related to outcomes was not investigated. Neither did the study evaluate whether allopurinol doses were adjusted according to renal function to minimise the risk of toxicity in patients with renal impairment.27 32 The relationship between flares and the recommended7 27 prophylactic use of low-dose colchicine or NSAIDs to minimise the risk of gout flares following the initiation of urate-lowering treatment was not investigated. In conclusion, this study has shown that the prevalence of gout in practice in the UK and Germany in the years 2000–5 was 1.4%, a figure that is consistent with previous UK data for 1990–9. Patients with gout had a high frequency of chronic comorbidities, several of which are associated with an increased risk of cardiovascular disease. Allopurinol was the preferred urate-lowering drug for the management of interval and chronic gout in both countries studied, but persistence with treatment was poor and gout flares were frequent despite urate-lowering treatment.

bidities, several of which are associated with an increased risk of cardiovascular disease. Allopurinol was the preferred urate-lowering drug for the management of interval and chronic gout in both countries studied, but persistence with treatment was poor and gout flares were frequent despite urate-lowering treatment. Evaluation of sUA levels to guide treatment was not undertaken sufficiently. The importance of regular monitoring of sUA levels and adjustment of treatment to optimise sUA was highlighted by data from this study, showing that the prevalence of comorbidities among patients with gout was directly related to sUA levels, and that patients with sUA levels >360 μmol/l (>6 mg/dl) had an increased risk of gout flares. We take full responsibility for the design, results and interpretation of the work and thank David Cooke of Medical Action Communications (supported by Ipsen) for preparation and editing of the manuscript. Competing interests: LA has received an unrestricted research grant from Ipsen. MB and VM are Ipsen employees. TG is a former Ipsen employee. ES, MG, GN: no competing interests. Funding: Financial support for this study was provided by Ipsen. Involvement by Ipsen in the study design, in the collection, analysis and interpretation of data, and in the writing of the report and the decision to submit for publication was limited to the role of the authors employed by Ipsen.

Within the last decade, an increasing number of rheumatologists worldwide have incorporated musculoskeletal ultrasound (MSUS) as a valuable imaging and research tool in their clinical practice. The European League Against Rheumatism (EULAR) has supported 14 courses on MSUS since 1998. There have been introductory and advanced MSUS courses in different European countries under the auspices of the EULAR Standing Committee on Education and Training. European rheumatologists highly experienced in MSUS have comprised the faculty of these courses. Many of them chair and organise MSUS training for rheumatologists in their own countries. In 2001, the first guidelines for performing MSUS in rheumatology were published by the EULAR Working Group for Musculoskeletal Ultrasound.1 These guidelines provided useful information on the technical basis for MSUS, equipment specifications, scanning methods and image acquisition along with the main pathological findings in each anatomical area. Within the last 3 years, a number of relevant papers on MSUS education, curriculum and competency for rheumatologists have been published.2–7 Brown et al produced an international interdisciplinary consensus on the specific indications, anatomic areas and knowledge and skills required by rheumatologists performing MSUS.4 Nevertheless, there is a wide variety of approaches to training in MSUS.8 Furthermore, there is a need for a standardised educational programme of efficient teaching and learning of MSUS for rheumatologists.

the specific indications, anatomic areas and knowledge and skills required by rheumatologists performing MSUS.4 Nevertheless, there is a wide variety of approaches to training in MSUS.8 Furthermore, there is a need for a standardised educational programme of efficient teaching and learning of MSUS for rheumatologists. The purpose of this project was to develop education guidelines for future EULAR ultrasound courses. These could also be recommended to national and local MSUS training programmes as a formal educational model and curriculum for rheumatologists. METHODS Study design The initial step involved a 2-month email-based forum on preliminary ideas for developing MSUS training guidelines, using a core group composed of the organisers of the recent past and future EULAR ultrasound courses. This preliminary approach included discussion on the following issues: (1) the desired model of MSUS education, (2) the theoretical and practical curriculum, (3) the proposed course schedule and duration and (4) the requirements and restrictions for attendance at the courses. We then undertook a consensus process through two consecutive written questionnaires sent to the 29 senior ultrasonographer rheumatologists who comprised the faculty of the 14th EULAR ultrasound course that was held in Sitges, Spain, 10–13 June 2007. These faculty members were from 12 European countries (Denmark, 2; Finland, 1; France, 2; Germany, 4; Hungary, 1; Ireland, 1; Italy, 3; The Netherlands, 2; Norway, 1; Spain, 10; Switzerland, 1; UK, 1).

logists who comprised the faculty of the 14th EULAR ultrasound course that was held in Sitges, Spain, 10–13 June 2007. These faculty members were from 12 European countries (Denmark, 2; Finland, 1; France, 2; Germany, 4; Hungary, 1; Ireland, 1; Italy, 3; The Netherlands, 2; Norway, 1; Spain, 10; Switzerland, 1; UK, 1). Questionnaire design and content We developed the first questionnaire based on the issues discussed among the core group. This questionnaire was sent to the 29 rheumatologists who participated as faculty in the 14th EULAR ultrasound course, and they were asked to respond within 2 months. An explanation of the purpose of the exercise accompanied the questionnaire. After 4 weeks, e-mail reminders were sent to the non-responders.

core group. This questionnaire was sent to the 29 rheumatologists who participated as faculty in the 14th EULAR ultrasound course, and they were asked to respond within 2 months. An explanation of the purpose of the exercise accompanied the questionnaire. After 4 weeks, e-mail reminders were sent to the non-responders. The first questionnaire included 30 questions divided into 8 sections to be answered by ticking “yes” or “no” with space for additional comments. Participants were allowed to vote on more than one option for each section. The eight sections that comprised the first questionnaire included the following topics: MSUS educational model, course timing, course curriculum, course duration, number of participants per teacher in practical sessions, time spent on hands-on sessions, requirements, and/or restrictions for attendance at the courses. The educational model section offered four possibilities: two-level education, three-level education (eg basic, intermediate and advanced), two-level education and additional courses on selected advanced subjects, and two-level education and additional modular courses on specific anatomic areas and/or diseases or group of diseases. The course timing section offered four possibilities: ultrasound courses with timing and location closely related to the annual EULAR Congresses, ultrasound courses scheduled apart from the annual EULAR Congresses, the basic or the basic and intermediate courses with timing and location closely related to the annual EULAR Congresses and the advanced course at different time. The course curriculum section offered four questions about the main educational objectives of the courses (eg examination technique versus examination technique plus basic pathology for the basic course; MSUS research in rheumatology versus pathological findings in other specialities for the advanced course). The course duration section provided three possibilities: 20 h over 3 days, 24 h over 3 days, and 20 h over 2.5 days. With regard to the number of participants per teacher in practical sessions, three options (four, five or six participants) were given. The offered possibilities of time spent on “hands-on” sessions were 40–50%, 50–60% or 60–70% of the total duration of each course. Nine questions about different possibilities of requirements and/or restrictions for attendance at the courses were included in the last section.

four, five or six participants) were given. The offered possibilities of time spent on “hands-on” sessions were 40–50%, 50–60% or 60–70% of the total duration of each course. Nine questions about different possibilities of requirements and/or restrictions for attendance at the courses were included in the last section. The second questionnaire consisted of two parts. Firstly, seven questions were related to areas of non-consensus and comments supplied in the first questionnaire, to be answered by ticking “yes” or “no” with space for any additional comments. Secondly, it included a list of 22 topics and 29 MSUS pathologies to be assigned to different levels of education, together with space for any suggestions of other topics or pathologies not included in the provided list. Participants were allowed to vote on more than one option for each section. The second questionnaire and the results from the first questionnaire were sent by e-mail to the responders to the first questionnaire. They were asked to respond within 2 months and after 4 weeks, e-mail reminders were sent to the non-responders. Analysis We calculated the percentage of respondents who answered yes and no to the questions included in each questionnaire. Group agreement with the issue under consideration was defined as agreement ⩾65%.

The second questionnaire consisted of two parts. Firstly, seven questions were related to areas of non-consensus and comments supplied in the first questionnaire, to be answered by ticking “yes” or “no” with space for any additional comments. Secondly, it included a list of 22 topics and 29 MSUS pathologies to be assigned to different levels of education, together with space for any suggestions of other topics or pathologies not included in the provided list. Participants were allowed to vote on more than one option for each section. The second questionnaire and the results from the first questionnaire were sent by e-mail to the responders to the first questionnaire. They were asked to respond within 2 months and after 4 weeks, e-mail reminders were sent to the non-responders. Analysis We calculated the percentage of respondents who answered yes and no to the questions included in each questionnaire. Group agreement with the issue under consideration was defined as agreement ⩾65%. RESULTS The response rate was 82.7% (24 out of 29) from the first questionnaire and 87.5% (21 out of 24) from the second questionnaire. The respondents were from 11 European countries (Denmark, 2; Finland, 1; France, 1; Germany, 2; Hungary, 1; Ireland, 1; Italy, 3; The Netherlands, 1; Norway, 1; Spain, 10; UK, 1). Because of the high percentage of respondents from one European country (Spain) with respect to the other countries, we calculated again the results for each question taking into account the answers from a maximum of four Spanish ultrasonographer rheumatologists. The four Spanish rheumatologists who had comprised the faculty of more EULAR ultrasound courses than the others were selected for this second analysis. The results of this separate analysis are only reported when they were different from the total group.

from a maximum of four Spanish ultrasonographer rheumatologists. The four Spanish rheumatologists who had comprised the faculty of more EULAR ultrasound courses than the others were selected for this second analysis. The results of this separate analysis are only reported when they were different from the total group. Results from the first questionnaire round The total agreement for a three-level education model was 65.2%. A total of 17.4% of the respondents voted for a two-level education model, 30.4% voted for two-level education and additional courses on selected advanced subjects and 26.1% voted for two-level education and additional modular courses on specific anatomic areas and/or diseases or group of diseases. Eight (33.3%) respondents voted for two options. After analysing again the answers including only four Spanish respondents, 50% of the respondents voted for a three-level education model, 22.2% of the respondents voted for a two-level education model, 33.3% voted for two-level education and additional courses on selected advanced subjects and 22.2% voted for two-level education and additional modular courses. Ultrasound courses with timing and location closely related to the annual EULAR Congresses was agreed by 87% of the respondents. However, there were comments on the possibility that future EULAR Congresses may take place in cities/countries where there are no colleagues willing or able to organise the MSUS courses.

Results from the first questionnaire round The total agreement for a three-level education model was 65.2%. A total of 17.4% of the respondents voted for a two-level education model, 30.4% voted for two-level education and additional courses on selected advanced subjects and 26.1% voted for two-level education and additional modular courses on specific anatomic areas and/or diseases or group of diseases. Eight (33.3%) respondents voted for two options. After analysing again the answers including only four Spanish respondents, 50% of the respondents voted for a three-level education model, 22.2% of the respondents voted for a two-level education model, 33.3% voted for two-level education and additional courses on selected advanced subjects and 22.2% voted for two-level education and additional modular courses. Ultrasound courses with timing and location closely related to the annual EULAR Congresses was agreed by 87% of the respondents. However, there were comments on the possibility that future EULAR Congresses may take place in cities/countries where there are no colleagues willing or able to organise the MSUS courses. In all, 48.3% of the respondents voted that the basic course should include only ultrasound examination technique and 48.3% that it should also include basic pathological findings. A total of 74% of the respondents voted that advanced course should include MSUS research and methodology, an update on MSUS in rheumatology and technological developments, and 65.2% of the respondents voted that the advanced course should include MSUS uncommon findings in rheumatology as well as MSUS findings in other specialities.

of the respondents voted that advanced course should include MSUS research and methodology, an update on MSUS in rheumatology and technological developments, and 65.2% of the respondents voted that the advanced course should include MSUS uncommon findings in rheumatology as well as MSUS findings in other specialities. The agreement on course duration was 48.3% for 20 h over 3 days, 43.5% on 24 h over 3 days and 13% on 20 h over 2.5 days. A maximum of six participants per teacher, and ideally four or five participants per teacher, was voted by 100% of respondents. With regard to time spent on “hands-on” scanning, more than 65% of the respondents voted for 50–60% of the total duration of each course. In all, 61% of respondents agreed that there should be some restrictions to participating in EULAR ultrasound courses. Only 8.7% voted for restrictions in attending the basic course, 69.6% for restricting attendance at the intermediate course and 82.6% voted for restrictions on attending the advanced course. After considering only four Spanish respondents, 44.4% voted for some restrictions to participating in EULAR ultrasound courses, 5.5% for restrictions in attending the basic course, 61.1% for restricting attendance at the intermediate course and 72.2% voted for restricting attendance at the advanced course.

ced course. After considering only four Spanish respondents, 44.4% voted for some restrictions to participating in EULAR ultrasound courses, 5.5% for restrictions in attending the basic course, 61.1% for restricting attendance at the intermediate course and 72.2% voted for restricting attendance at the advanced course. Results from the second questionnaire round A proposal was made for three-level courses with timing and location closely related to the annual EULAR Congress if local organisers are willing and able to conduct such courses; as well, if future EULAR Congresses take place in cities/countries where there are no colleagues willing or able to organise the MSUS courses, the courses would be organised by another colleague in their own country/city. This was agreed by 95.2% of the respondents. In all, 95.2% of the respondents voted that the basic course should focus on ultrasound examination techniques and basic pathology, the intermediate course should focus on a wide spectrum of rheumatologic pathology, and the advanced course should focus on current MSUS research in rheumatology, new technological developments, uncommon pathological findings in rheumatology, pathological findings in other specialities (such as nerve, ligament, muscle and sport-related lesions), and MSUS methodology. The total agreement for course duration of 20 h for 3 days was 70%.

In all, 95.2% of the respondents voted that the basic course should focus on ultrasound examination techniques and basic pathology, the intermediate course should focus on a wide spectrum of rheumatologic pathology, and the advanced course should focus on current MSUS research in rheumatology, new technological developments, uncommon pathological findings in rheumatology, pathological findings in other specialities (such as nerve, ligament, muscle and sport-related lesions), and MSUS methodology. The total agreement for course duration of 20 h for 3 days was 70%. The proposal was overwhelmingly agreed that there will be no restrictions on attending basic courses: however, the proposals that previous performance of ⩾100 MSUS scans before participating in an intermediate level course, and previous performance of ⩾300 MSUS scans before attending an advanced level course were agreed by 90.5% of the respondents. Eight topics and five pathologies were assigned to the basic course by more than 70% of the respondents, 7 topics and 16 pathologies to the intermediate course and 8 topics and 9 pathologies to the advanced course (table 1).

The proposal was overwhelmingly agreed that there will be no restrictions on attending basic courses: however, the proposals that previous performance of ⩾100 MSUS scans before participating in an intermediate level course, and previous performance of ⩾300 MSUS scans before attending an advanced level course were agreed by 90.5% of the respondents. Eight topics and five pathologies were assigned to the basic course by more than 70% of the respondents, 7 topics and 16 pathologies to the intermediate course and 8 topics and 9 pathologies to the advanced course (table 1). Table 1 Musculoskeletal ultrasound (MSUS) course curriculum and pathology content for basic, intermediate and advanced levels Level Content Basic Application, indications and limitations of MSUS in rheumatology. Ultrasound physics and technology Sonographic pattern of the different musculoskeletal tissues MSUS artefacts and pitfalls Standard sonographic scans of the shoulder, elbow, wrist and hand, hip, knee, ankle and foot Holding the probe and optimising the grey-scale settings of the sonographic system Image documentation Reporting ultrasound findings and diagnosis Basic course pathologies Joint synovitis Joint effusion Synovial hypertrophy Bursitis Tenosynovitis

c scans of the shoulder, elbow, wrist and hand, hip, knee, ankle and foot Holding the probe and optimising the grey-scale settings of the sonographic system Image documentation Reporting ultrasound findings and diagnosis Basic course pathologies Joint synovitis Joint effusion Synovial hypertrophy Bursitis Tenosynovitis Intermediate Colour and power Doppler physics and technology Application, indications and limitations of colour and power Doppler in rheumatology Use of the colour and power Doppler settings Colour and power Doppler artefacts Use of colour and power Doppler to detect synovial and entheseal inflammation Assessment and quantification of structural joint damage (bone, tendons, ligaments) Sonographic-guided periarticular and articular injections Intermediate course pathologies Joint synovitis, synovial hypertrophy, tenosynovitis Tendon calcification Enthesopathy Tendinosis Paratenonitis Tendon subluxation/luxation Intrasubstance tendon lesions Tendon impingement Complete tendon tear Partial tendon tear Bone erosions Osteophytes Ganglia and cysts Articular cartilage lesions Peri- and intra-articular microcrystal deposit Ligament, muscle, cartilage, fibrocartilage and synovial calcification

aratenonitis Tendon subluxation/luxation Intrasubstance tendon lesions Tendon impingement Complete tendon tear Partial tendon tear Bone erosions Osteophytes Ganglia and cysts Articular cartilage lesions Peri- and intra-articular microcrystal deposit Ligament, muscle, cartilage, fibrocartilage and synovial calcification Advanced Optimisation of colour and power Doppler settings Sonographic-guided musculoskeletal interventional procedures Assessment and quantification of synovial, tenosynovial and entheseal inflammatory activity Role of ultrasound in vasculitis Evaluation of vessels and detection of vasculitis by sonography Paediatric sonography: musculoskeletal sonoanatomy and pathological findings in rheumatic diseases Uncommon sonographic pathological findings in rheumatology MSUS technological development 3- and 4-dimensional MSUS Update on MSUS in rheumatology MSUS research and methodology Advanced course pathologies Peripheral nerve entrapment and lesions Ligament lesions Fibrocartilage lesions Myopathy Myositis Muscle injury Soft tissue masses Loose bodies Foreign bodies In summary, the group consensus on guidelines and curriculum for future EULAR MSUS courses are outlined below. 1. MSUS education model A three-level education model with basic, intermediate and advanced levels will be conducted.

Advanced Optimisation of colour and power Doppler settings Sonographic-guided musculoskeletal interventional procedures Assessment and quantification of synovial, tenosynovial and entheseal inflammatory activity Role of ultrasound in vasculitis Evaluation of vessels and detection of vasculitis by sonography Paediatric sonography: musculoskeletal sonoanatomy and pathological findings in rheumatic diseases Uncommon sonographic pathological findings in rheumatology MSUS technological development 3- and 4-dimensional MSUS Update on MSUS in rheumatology MSUS research and methodology Advanced course pathologies Peripheral nerve entrapment and lesions Ligament lesions Fibrocartilage lesions Myopathy Myositis Muscle injury Soft tissue masses Loose bodies Foreign bodies In summary, the group consensus on guidelines and curriculum for future EULAR MSUS courses are outlined below. 1. MSUS education model A three-level education model with basic, intermediate and advanced levels will be conducted. 2. Course timing There will be three-level courses with timing and location related to the annual EULAR Congresses if local organisers are willing and able to conduct the courses. If future EULAR Congresses take place in cities/countries where there are no colleagues willing or able to organise the MSUS courses, another colleague can organise three courses (basic, intermediate and advanced)/simultaneously in their country/city. With this course timing, 2 years is the minimum time in which all three courses can be attended by trainees interested in MSUS. They would have a minimum of 1 year for practising between consecutive levels.

her colleague can organise three courses (basic, intermediate and advanced)/simultaneously in their country/city. With this course timing, 2 years is the minimum time in which all three courses can be attended by trainees interested in MSUS. They would have a minimum of 1 year for practising between consecutive levels. 3. Course curriculum Basic courses will focus on examination technique and will include some basic pathology. Intermediate courses will focus on a wide spectrum of rheumatologic pathology. Advanced courses will focus on current MSUS research in rheumatology, new technological developments, uncommon pathological findings in rheumatology, pathological findings in other specialities such as nerve, ligament, muscle lesions, sport related lesions and MSUS methodology (table 1). 4. Course duration A total of 20 h for 3 days. 5. Number of participants per teacher in practical sessions Ideally four or five, maximum six participants per teacher. 6. Time spent on hands-on scanning In all, 50–60% of total time will be spent in practical training for all courses; 40–50% of total time will be theoretical teaching. 7. Attendance at the EULAR ultrasound courses There will be no prerequisites for attending basic courses. Previous performance of ⩾100 musculoskeletal ultrasound scans is strongly recommended before participating in an intermediate-level course and previous performance of ⩾300 musculoskeletal ultrasound scans is strongly recommended before attending an advanced-level course. 8. Certification The group voted for working towards certification of attendance and competency.

7. Attendance at the EULAR ultrasound courses There will be no prerequisites for attending basic courses. Previous performance of ⩾100 musculoskeletal ultrasound scans is strongly recommended before participating in an intermediate-level course and previous performance of ⩾300 musculoskeletal ultrasound scans is strongly recommended before attending an advanced-level course. 8. Certification The group voted for working towards certification of attendance and competency. DISCUSSION High-resolution MSUS has become an established imaging technique for evaluating periarticular and intra-articular structures involved in rheumatic diseases.9–11 In addition, ultrasound is a bedside tool for performing accurate and safe musculoskeletal injections.12 Within the last decade, several reports on the superiority of US over clinical evaluation and plain radiography for assessing early joint inflammatory and structural changes have directed MSUS applications in rheumatology towards early diagnosis, assessment of disease activity and monitoring of therapeutic response in patients with inflammatory arthritis.13–21 As a consequence of the recent technological development and increasing utility of MSUS, there is a great demand for appropriate education in this technique among rheumatologists worldwide. Since US is an operator-dependent imaging technique (mainly because of the intrinsic real time nature of US images acquisition) appropriate training is highly important to ensure skilled and safe use of MSUS by rheumatologists.22 23

eat demand for appropriate education in this technique among rheumatologists worldwide. Since US is an operator-dependent imaging technique (mainly because of the intrinsic real time nature of US images acquisition) appropriate training is highly important to ensure skilled and safe use of MSUS by rheumatologists.22 23 Sustained and extensive interest has occurred in attending EULAR ultrasound courses that have been organised over the last 9 years. Furthermore, several MSUS courses and workshops are offered by individual European national Societies of Radiology and Rheumatology and by Universities in many countries. In addition, MSUS is a compulsory part of rheumatology training in some European countries. However, until now there has been no agreed educational programme and curriculum on MSUS for European rheumatologists. We therefore aimed to develop guidelines on content and conducting MSUS courses for rheumatologists under the auspices of EULAR. These guidelines can also be followed by national and local societies and/or universities in order to standardise MSUS training across Europe. Standardisation of the MSUS education model is essential for validating the results of training in this technique.

MSUS courses for rheumatologists under the auspices of EULAR. These guidelines can also be followed by national and local societies and/or universities in order to standardise MSUS training across Europe. Standardisation of the MSUS education model is essential for validating the results of training in this technique. We used a consensus method among European rheumatologists highly experienced in performing and teaching MSUS. They were asked to vote on different options regarding content and conduct of MSUS courses. Controversial points were voted again in the next round after presenting the group replies from the previous round. The high response rate reflects their great motivation for MSUS education. Because there was a higher number of respondents from one European country (Spain) than from the other countries, we analysed again the questionnaires taking into account only the answers from the four Spanish ultrasonographer rheumatologists who had participated as faculty in more EULAR ultrasound courses than their peers in order to avoid unbalanced influence from the country in question. After excluding the answers from six Spanish rheumatologists, the results did not change considerably.

the answers from the four Spanish ultrasonographer rheumatologists who had participated as faculty in more EULAR ultrasound courses than their peers in order to avoid unbalanced influence from the country in question. After excluding the answers from six Spanish rheumatologists, the results did not change considerably. The most controversial issue concerned how many levels of MSUS training should be included. Arguments for the three-level education model (basic, intermediate and advanced) included that this would result in a higher competency achieved by trainees, more homogeneous groups in practical sessions in each course level, increased time for practical sessions in each course and a step-by-step training in MSUS. However, three-level courses may be more difficult to organise and conduct, as well as more time consuming for participants and teachers than two-level courses. Nevertheless, the 14th EULAR ultrasound course consisted of three simultaneous basic, intermediate and advanced courses (10–13 June 2007, Sitges, Spain). This course was successfully carried out with a careful, pre-planned organisation.

s well as more time consuming for participants and teachers than two-level courses. Nevertheless, the 14th EULAR ultrasound course consisted of three simultaneous basic, intermediate and advanced courses (10–13 June 2007, Sitges, Spain). This course was successfully carried out with a careful, pre-planned organisation. With regard to the other educational models proposed, they were more frequently voted for than the two-level education model. The two-level education model plus additional courses on selected advanced subjects is quite similar to a three-level education model. The two-level education model plus modular courses on anatomic areas and/or diseases would lead to a higher number of courses than in the three-level model, increasing time and money spent on ultrasound education by trainees and trainers. The total agreement for a three-level education model decreased from 65.2% to 50% after considering only four Spanish respondents. The six excluded Spanish rheumatologists voted for a three-level education model because most were involved as organisers of the 14th EULAR ultrasound course. Nevertheless, a three-level education model was agreed on by more respondents than a two-level education model (22%), two-level education and additional courses on selected advanced subjects (33%) and two-level education and additional modular courses (22%).

e involved as organisers of the 14th EULAR ultrasound course. Nevertheless, a three-level education model was agreed on by more respondents than a two-level education model (22%), two-level education and additional courses on selected advanced subjects (33%) and two-level education and additional modular courses (22%). With the three-level course timing, trainees can attend a course each year. They need a minimum of 2 years for attending basic, intermediate and advanced course and they have a minimum of 1 year for practising between consecutive levels. We consider that 2 years should be the minimal time necessary for achieving competency in MSUS. For the curriculum of the courses, we selected the principal topics and pathologies that have been considered relevant and appropriate for rheumatologic practice according to published data.6 Afterwards, they were assigned to the basic, intermediate and advanced level according to the results from the second questionnaire. Since practical training under expert supervision is essential for appropriate MSUS learning, the courses should have a 50% to 60% of total time spent in practical, hands-on sessions and no more than six participants per tutor in such sessions.

For the curriculum of the courses, we selected the principal topics and pathologies that have been considered relevant and appropriate for rheumatologic practice according to published data.6 Afterwards, they were assigned to the basic, intermediate and advanced level according to the results from the second questionnaire. Since practical training under expert supervision is essential for appropriate MSUS learning, the courses should have a 50% to 60% of total time spent in practical, hands-on sessions and no more than six participants per tutor in such sessions. Although the courses should be considered as a necessary starting point for developing and improving MSUS skills, training after courses by performing normal scans and diagnostic examinations is mandatory for consolidating the knowledge and skills provided during the courses. Then, a number of ultrasound examinations are recommended before attending intermediate and advanced courses. Certification is desirable. Guidelines for certification of competency will be developed in cooperation with the EULAR Standing Committee on Education and Training, the EULAR Committee on Musculoskeletal Imaging, and with other European institutions. In conclusion, we have developed European agreed guidelines for content and conducting EULAR MSUS courses, which will be also useful for standardising rheumatology MSUS training worldwide. Competing interests: None declared.

Primary Sjögren syndrome (pSS) is a systemic autoimmune disease, which shares a number of clinical, serological and genetic features with systemic lupus erythematosus (SLE).1 One of the features of this disease is the high prevalence of autoantibodies to the Ro and La components of a ribonuclear protein (RNP) complex. Ro/La autoantibodies occur in individual patients as targeting either Ro alone, or with specificity to both Ro and La, and autoantibody subgroups are stable over time. Major histocompatibility complex (MHC) disease associations have been known for many years,2 3 and more recently it has been recognised that these MHC haplotypes regulate the diversification of the autoantibody response.4 5 Ro/La autoantibodies occur in other rheumatological diseases such as SLE; however, they are highly specific for pSS and constitute one of the classification criteria for this disease.6 Serum Ro/La autoantibodies have been shown to pre-date disease onset, often for many years, and seropositivity predicts disease severity and extraglandular features in patients with pSS.7 More recently, a pathogenic role has been proposed as Ro/La autoantibodies form nucleic-acid-containing immune complexes that can trigger prolonged type I interferon production, leading to a self-perpetuating autoimmune reaction.8

ropositivity predicts disease severity and extraglandular features in patients with pSS.7 More recently, a pathogenic role has been proposed as Ro/La autoantibodies form nucleic-acid-containing immune complexes that can trigger prolonged type I interferon production, leading to a self-perpetuating autoimmune reaction.8 Deficiencies in the classical pathway of the complement system have been implicated in the aetiology and pathogenesis of autoimmune diseases such as SLE.9 Human complement receptor 1 (CR1, CD35) is an integral membrane complement control protein (CCP) whose primary role on erythrocytes is the non-inflammatory clearance of immune complexes opsonised with C3a and C4b. There is longstanding evidence, over 25 years in our own papers,10 for an acquired CR1 deficiency in SLE. A recent meta-analysis has demonstrated an association between a molecular weight variant of CR1 and SLE.11 Further, in the mouse, CR1/CR2 (encoded by alternatively spliced forms of the same gene) is a lupus susceptibility gene12 13 important in the modulation of the antinuclear autoantibody response.14 15 Human CR1 is located at 1q32 in the regulators of complement activation (RCA) complex, which also includes other complement regulatory encoding genes, namely factor H (CFH), complement 4 α and β binding protein’s (C4ABP and C4BBP), decay accelerating factor (DAF, CD56), complement receptor 2 (CR2, CD21) and membrane cofactor protein (MCP, CD46).

2 in the regulators of complement activation (RCA) complex, which also includes other complement regulatory encoding genes, namely factor H (CFH), complement 4 α and β binding protein’s (C4ABP and C4BBP), decay accelerating factor (DAF, CD56), complement receptor 2 (CR2, CD21) and membrane cofactor protein (MCP, CD46). Encompassing some 13 Mb of the long arm of chromosome 1, the RCA does not contain an even distribution of CCPs. In fact all CCPs can be found within one of two distinct blocks, located at the telomeric and centromeric ends of the cluster, we refer to these as the α and β blocks respectively. Both blocks are about 500 kb in length and are characterised by extensive imperfect genomic duplication and degeneracy generated by insertions and deletions (indels) such as those created through domain duplication/deletion and retroviral insertion. CR1 is encoded within the RCAα block, where it has been duplicated as part of a segment along with MCP-Like, to form CR1-L and MCP (or vice versa). Segment A (CR1 and MCP-L) and B (CR1-L and MCP) are located next to each other and encompass ∼350 kb of the α block.16

domain duplication/deletion and retroviral insertion. CR1 is encoded within the RCAα block, where it has been duplicated as part of a segment along with MCP-Like, to form CR1-L and MCP (or vice versa). Segment A (CR1 and MCP-L) and B (CR1-L and MCP) are located next to each other and encompass ∼350 kb of the α block.16 As demonstrated some time ago within the MHC,17–19 and more recently throughout the entire genome,20–22 genomic blocks are normally defined by gene and segmental duplications while also exhibiting other characteristics such as suppression of recombination and an increased frequency of copy number, indel and single nucleotide polymorphisms.23 A consequence of recombination suppression is that haplotypes of these regions are likely to have been inherited faithfully over many generations (ie, ancestral haplotypes (AH)) and will therefore be important in defining complex genetic interactions.24

uency of copy number, indel and single nucleotide polymorphisms.23 A consequence of recombination suppression is that haplotypes of these regions are likely to have been inherited faithfully over many generations (ie, ancestral haplotypes (AH)) and will therefore be important in defining complex genetic interactions.24 We recently reported a novel haplotyping approach, capable of interrogating the genetically complex RCAα block, by utilising the genomic duplication of CR1 and MCP (segment A and segment B).25 The assay uses the principles of the Genomic Matching Technique,23 used by us, and other groups, to interrogate the highly complex MHC region for the identification of donor/recipient matches prior to bone marrow transplantation.26–29 More recently the technique has been extended to forensic applications30 and used within the RCA complex on chromosome 125 and the class II region of the canine MHC.31 The technique involves the specific amplification, with a single primer pair, of multiple complex geometric elements, all linked and located within duplicated segments. The conceptual basis behind the technique is that duplicated elements make excellent haplospecific markers as they have evolved in line with the many other changes that have also occurred on each haplotype.

a single primer pair, of multiple complex geometric elements, all linked and located within duplicated segments. The conceptual basis behind the technique is that duplicated elements make excellent haplospecific markers as they have evolved in line with the many other changes that have also occurred on each haplotype. The aim of this study was to evaluate RCAα block haplotypes, as defined by the Genomic Matching Technique (GMT), in relation to both disease susceptibility and Ro/La autoantibody responses in patients with pSS. Specifically we hypothesised that genetic variation in the RCAα, in combination with HLA, would influence the diversification of the Ro/La autoantibody response.

lotypes, as defined by the Genomic Matching Technique (GMT), in relation to both disease susceptibility and Ro/La autoantibody responses in patients with pSS. Specifically we hypothesised that genetic variation in the RCAα, in combination with HLA, would influence the diversification of the Ro/La autoantibody response. MATERIALS AND METHODS Study participants Ninety-eight Caucasian controls and 115 Caucasian patients with pSS from the South Australian Sjögren syndrome research registry were included in the study. All patients met the revised 2002 American-European consensus research classification criteria for pSS,6 and controls were recruited from the same population base as the patients. Anti-Ro/La autoantibody specificity was determined by enzyme-linked immunosorbent assay (RELISA ANA Screening System, Immuno Concepts NA, Sacramento, California, USA) using recombinant Ro60 and La proteins, as part of standard diagnostic procedure. Sera from patients with anti-La were further tested by counterimmunoelectrophoresis (CIEP)32 to confirm whether or not anti-La antibodies detected by enzyme-linked immunosorbent assay were able to be detected by this method. HLA typing of patients with pSS (serological HLA-B and molecular DRB1) was performed by the Transplantation Laboratory, Australian Red Cross Blood Service, SA Division. Molecular DRB1 typing of the controls was performed by Conexio Genomics (Applecross, WA, Australia). The study was approved by the Human Ethics Committee of The Queen Elizabeth and Royal Adelaide Hospitals and all patients gave informed, written consent.

tion Laboratory, Australian Red Cross Blood Service, SA Division. Molecular DRB1 typing of the controls was performed by Conexio Genomics (Applecross, WA, Australia). The study was approved by the Human Ethics Committee of The Queen Elizabeth and Royal Adelaide Hospitals and all patients gave informed, written consent. RCAα block haplotyping Haplotypes of the RCAα block were obtained using the GMT assays as previously described.25 Briefly, two separate polymerase chain reaction reactions using primer sets CR1MCP5&6 and CR1MCP11&12 were performed on each genomic DNA sample. Each primer set was designed to amplify a duplicated element, located within the CR1 region of both segments (segment A containing CR1 and MCP-Like and segment B containing CR1-Like and MCP). The resulting mix of polymerase chain reaction products were used to define the haplotypic variation within the RCAα block. The polymerase chain reaction products were separated on the basis of size on a Corbett Research GS-3000 automated gel analysis system (Corbett Research (Australia), Mortlake, NSW, Australia). Haplotype assignment and nomenclature are as previously described.25 Statistical analysis Contingency table analysis of haplotype and genotype frequencies in patients with pSS versus controls was performed by multivariate logistic regression using both additive and dominant allele coding. Associations were further reported as odds ratios (OR) with 95% confidence intervals (CI) obtained by back transformation of the regression coefficients.

lotype and genotype frequencies in patients with pSS versus controls was performed by multivariate logistic regression using both additive and dominant allele coding. Associations were further reported as odds ratios (OR) with 95% confidence intervals (CI) obtained by back transformation of the regression coefficients. RESULTS RCAα block haplotype diversity We have previously demonstrated substantial polymorphism in GMT RCAα block haplotypes.25 More than 20 haplotypes have been defined, although the majority are rare. In the current study of 213 Caucasians (pSS and controls combined), there were three relatively common haplotypes (AH1, AH2 and AH3 as designated by McLure et al25 each with a frequency of >10%. These three haplotypes combined accounted for 56% of the total haplotypes in the sample. There were a further 14 haplotypes with a frequency between 1 and 3%. These frequencies were considered too low to be informative given the study sample sizes and were therefore combined for analysis purposes. RCAα block genotypes were in Hardy–Weinberg equilibrium in both patients with pSS and controls (p = 0.93, p = 0.21 respectively, exact test).

pes with a frequency between 1 and 3%. These frequencies were considered too low to be informative given the study sample sizes and were therefore combined for analysis purposes. RCAα block genotypes were in Hardy–Weinberg equilibrium in both patients with pSS and controls (p = 0.93, p = 0.21 respectively, exact test). RCAα block haplotypes are associated with Ro/La autoantibody-positive primary Sjögren syndrome CR1 haplotype frequencies were significantly different between patients with pSS and controls (global χ2 = 14.6, df = 3, p = 0.002, table 1). Both AH1 (OR 2.1, 95% CI 1.3, 3.3) and AH3 (OR 2.4, 95% CI 1.3, 4.4) were significantly increased in pSS relative to controls implying an association between both of these haplotypes and susceptibility to pSS. However, this association was primarily attributable to the Ro/La autoantibody-positive subgroup of patients with pSS (n = 96, p = 0.0003) and was not evident in the relative minority of patients with pSS who were autoantibody negative (n = 19, p = 0.68).

tion between both of these haplotypes and susceptibility to pSS. However, this association was primarily attributable to the Ro/La autoantibody-positive subgroup of patients with pSS (n = 96, p = 0.0003) and was not evident in the relative minority of patients with pSS who were autoantibody negative (n = 19, p = 0.68). Table 1 RCAα haplotype frequencies in patients with pSS compared with controls Haplotype pSS(2N = 230) Controls(2N = 196) OR (95% CI) p Value AH1 80 (34.8%) 46 (23.5%) 2.1 (1.3,3.3)* 0.002 AH2 25 (10.9%) 27 (13.8%) 1.3 (0.6,2.1) 0.75 AH3 38 (16.5%) 19 (9.7%) 2.4 (1.3,4.4) 0.006 Other 87 (37.8%) 104 (53.1%) 1 Global test: χ2 = 14.6, df = 3, p = 0.002 There were three common RCAα block ancestral haplotypes, AH1, AH2 and AH3. “Other” haplotypes, with a frequency between 1 and 3%, were combined for analysis purposes. Both AH1 and AH3 frequencies were increased in patients with pSS relative to controls. This association was attributable to the Ro/La autoantibody-positive subgroup of patients with pSS and was not evident in the minority of patients with pSS who were autoantibody negative (2N = 38).

ere combined for analysis purposes. Both AH1 and AH3 frequencies were increased in patients with pSS relative to controls. This association was attributable to the Ro/La autoantibody-positive subgroup of patients with pSS and was not evident in the minority of patients with pSS who were autoantibody negative (2N = 38). With the exception of the relatively rare AH3 homozygotes, the frequencies of all genotypes carrying either AH1 or AH3 were increased in Ro/La autoantibody-positive patients with pSS compared with controls (fig 1). This was accompanied by a striking concomitant decrease in the frequency of all other (X,X) genotypes. There was significant departure from an additive or allele dose model (p = 0.043) by logistic regression analysis, and the data were most consistent with both AH1 and AH3 exerting dominant genotypic effects. Under the dominant model, AH1 and AH3 appeared to be independently associated with autoantibody-positive pSS as there was no evidence of any interaction (p = 0.258). The odds ratios for the dominant genotypic association with autoantibody-positive pSS were 3.8 (95% CI 2.0, 7.3, p = 0.0001) for AH1 and 4.1 (95% CI 1.9, 8.7, p = 0.0003) for AH3. The association for the compound heterozygous AH1,AH3 genotype was predicted by multiplication of these odds ratios.

nteraction (p = 0.258). The odds ratios for the dominant genotypic association with autoantibody-positive pSS were 3.8 (95% CI 2.0, 7.3, p = 0.0001) for AH1 and 4.1 (95% CI 1.9, 8.7, p = 0.0003) for AH3. The association for the compound heterozygous AH1,AH3 genotype was predicted by multiplication of these odds ratios. Figure 1 RCAα ancestral haplotype (AH) 1 and AH3 genotype frequencies in patients with primary Sjögren syndrome who were Ro/La autoantibody positive compared with controls. There was a significant departure from an additive or allele dose model (p = 0.043) by logistic regression analysis, and the data are most consistent with AH1 and AH3 exerting dominant effects of similar size. Under the dominant model, AH1 and AH3 appear to be independently associated with autoantibody-positive primary Sjögren syndrome, ie, with multiplicative risks. Epistatic interaction between the RCAα block and the major histocompatibility complex in susceptibility to Ro/La autoantibody-positive pSS An association between both HLA-DRB1*0301 (DR3) and HLA-DRB1*1501 (DR15) and autoantibody-positive pSS is well established in Caucasians.4 5 Therefore it was of interest to examine relationships between MHC and RCAα block associations with autoantibody-positive pSS.

in susceptibility to Ro/La autoantibody-positive pSS An association between both HLA-DRB1*0301 (DR3) and HLA-DRB1*1501 (DR15) and autoantibody-positive pSS is well established in Caucasians.4 5 Therefore it was of interest to examine relationships between MHC and RCAα block associations with autoantibody-positive pSS. Analysis of the cross-classification of HLA-DR3, DR15 and RCAα AH1 and AH3 genotypic combinations in autoantibody-positive patients with pSS (n = 92) versus controls (n = 96) was performed by multivariate logistic regression using dominant coding and all two-factor interaction terms were evaluated. The odds ratios from this analysis are illustrated in fig 2A. HLA DR3, DR15 alleles and the RCAα AH3 haplotype were independent risk factors for autoantibody-positive pSS, with similar effect sizes. Therefore, the association for compound heterozygous genotypes was predicted by multiplication of these odds ratios. However, there was a disproportionately increased frequency of patients with pSS who carried both HLA DR3 and RCAα AH1 (interaction term, p = 0.021) demonstrating an epistatic relationship.33 In the absence of RCAα AH1, HLA DR3 was still associated with Ro/La-positive pSS (OR 2.7, p = 0.025), although the association was greatly enhanced by the presence of AH1 in the same individual (OR 15.7, p = 10−8). Conversely, the RCAα AH1 association was not associated with genetic susceptibility to pSS in the absence of DR3 (OR = 1.3, p = 0.6). Moreover, this epistatic genotypic combination of HLA DR3 and RCAα AH1 was the largest genetic risk factor for autoantibody-positive pSS. It was highly specific for autoantibody-positive pSS as it was present in 44 of 92 (48%) patients with pSS who were Ro/La autoantibody positive compared with only eight of 96 (8%) controls (fig 2B). The majority of patients with pSS who did not carry this epistatic genetic risk factor carried any combination of DR3/DR15/AH3, the other risk genotypes identified in the analysis. Therefore, only four of 92 (4%) autoantibody-positive patients with pSS did not carry any risk genotypes compared with 36 of 96 (38%) of controls.

ority of patients with pSS who did not carry this epistatic genetic risk factor carried any combination of DR3/DR15/AH3, the other risk genotypes identified in the analysis. Therefore, only four of 92 (4%) autoantibody-positive patients with pSS did not carry any risk genotypes compared with 36 of 96 (38%) of controls. Figure 2 Epistatic interaction between the MHC and RCAα block in Ro/La autoantibody-positive patients with primary Sjögren syndrome (pSS). (A) Odds ratios (y-axis, logarithmic scale) derived by logistic regression for the cross-classification of HLA DR3, DR15 and RCAα AH1, AH3 genotypic combinations (dominant coding) in Ro/La autoantibody-positive patients with pSS (n = 92) relative to controls (n = 98). The vertical bars represent 95% confidence intervals, and the horizontal line represents an odds ratio of 1 (no effect). HLA DR3, DR15 alleles and the RCAα AH3 haplotype were independent risk factors for autoantibody-positive pSS (ie, multiplicative risks), but there was an epistatic interaction between HLA DR3 and RCAα AH1 (interaction term p = 0.021). The genotypic combination of HLA DR3 and RCAα AH1 was the greatest genetic risk factor for autoantibody-positive pSS (OR 15.7, p = 10−8), but in the absence of DR3, there was no effect of RCAα AH1. (B) Pie chart depicting the relative proportions of risk genotypes. The HLA DR3–RCAα AH1 epistatic combination was present in 48% of autoantibody-positive patients with pSS compared with 8% of controls. The majority of other patients with pSS carried any combination of HLA DR3, DR15 and RCAα AH3. Genetic associations with Ro/La autoantibody subsets in primary Sjögren syndrome Of 115 patients with pSS, 19 (16%) were negative and 97 (84%) positive for anti-Ro/La autoantibodies. Seropositive Ro+La patients by enzyme-linked immunosorbent assay were further subdivided into non-precipitating La, ie, Ro+La (ppt–), or precipitating, ie, Ro+La (ppt+), on the basis of a precipitin line formed by anti-La antibodies on CIEP. Therefore, in addition to a seronegative subset, seropositive patients with pSS were classified into one of three serological subsets: anti-Ro alone (19 of 115 = 16%), anti-Ro+La(ppt−) (22 of 115 = 19%) and anti-Ro+La(ppt+) (55 of 115 = 46%). These subgroups are characterised by increasing titre and a more polyclonal autoantibody response in addition to higher rheumatoid factor and IgG levels.4 32

were classified into one of three serological subsets: anti-Ro alone (19 of 115 = 16%), anti-Ro+La(ppt−) (22 of 115 = 19%) and anti-Ro+La(ppt+) (55 of 115 = 46%). These subgroups are characterised by increasing titre and a more polyclonal autoantibody response in addition to higher rheumatoid factor and IgG levels.4 32 We, and others4 5 have previously demonstrated that HLA-DR3 and DR15 frequencies differ between autoantibody subsets, and this question is also of interest in relation to the RCAα haplotypes. The phenotypic prevalences of DR3, DR15 and RCAα AH1 and AH3 in controls, seronegative pSS and the three pSS autoantibody subsets are displayed in fig 3. While the data are difficult to interpret definitively because of the relatively small sample sizes in three of the serological subsets, some inferences may be drawn.

The phenotypic prevalences of DR3, DR15 and RCAα AH1 and AH3 in controls, seronegative pSS and the three pSS autoantibody subsets are displayed in fig 3. While the data are difficult to interpret definitively because of the relatively small sample sizes in three of the serological subsets, some inferences may be drawn. Figure 3 Phenotypic prevalence of HLA DR3 (A), HLA-DR15 (B), RCAα AH1 (C) and RCAα AH3 (D) haplotypes by diversification of the Ro/La autoantibody response within primary Sjögren syndrome (pSS). There were 115 patients with pSS in the study: 19 were seronegative, 19 with anti-Ro only, 22 with anti-Ro+La (ppt−) and 55 with anti-Ro+La (ppt+). The increased prevalence of HLA DR3 in autoantibody-positive pSS is largely B8-DR3 (the prevalence of B8-DR3 in the controls is not known) and there are genetic differences between the pSS autoantibody subgroups. The increased prevalence of HLA DR3 in autoantibody-positive pSS is largely B8-DR3, which is carried on the 8.1 ancestral haplotype. B8-DR3 is most strongly associated with the polyclonal Ro+La (ppt+) autoantibody response, whereas, as previously reported,4 DR15 is almost exclusively associated with the restricted Ro+La (ppt−) response. The RCAα AH1 prevalence, although elevated, is relatively constant in autoantibody-positive patients. The basis for the epistatic interaction between RCAα AH1 and DR3 is most likely restricted to the 8.1 ancestral haplotype, rather than other DR3 containing haplotypes, and may exert a primary influence on the autoantibody response in pSS. Interestingly, 8.1 ancestral haplotype only one, rather than two or more C4 genes and is therefore associated with a relative C4 deficiency.34 However, there are clearly other genes on the MHC 8.1 haplotype that contribute additionally to regulation of the autoantibody response. The RCAα AH3 haplotype has the highest prevalence in patients with anti-La and may therefore potentiate diversification of the autoantibody response.

with a relative C4 deficiency.34 However, there are clearly other genes on the MHC 8.1 haplotype that contribute additionally to regulation of the autoantibody response. The RCAα AH3 haplotype has the highest prevalence in patients with anti-La and may therefore potentiate diversification of the autoantibody response. The genes for C2 are also in the extended MHC region and type 1 C2 deficiency is encoded within the 18.1 haplotype, which carries B18-DR15. There was no evidence of any epistatic interaction between RCAα block haplotypes and DR15. Further, only four B18-DR15 (from a total of 52 DR15) haplotypes were observed in patients with pSS and, as expected, there were no associations. DISCUSSION The GMT haplotyping approach has been used in this study to identify AH of the RCAα block and their associations with pSS, a systemic autoimmune disease. Duplication and copy number variation reveal more genetic diversity than single nucleotide polymorphisms,19 35 yet for most haplotyping assays, interpreting the results and defining haplotypes in these regions is difficult. The GMT assay is useful in interrogating complex genomic regions, such as the MHC and RCA blocks, as it requires and utilises these features to efficiently define the genomic polymorphism and AH within the region. The GMT RCAα block haplotyping has revealed extensive haplotypic polymorphism in this region (which also includes CR1-L, MCP and MCP-L genes), with more than 20 AH defined,25 although the majority are rare.

it requires and utilises these features to efficiently define the genomic polymorphism and AH within the region. The GMT RCAα block haplotyping has revealed extensive haplotypic polymorphism in this region (which also includes CR1-L, MCP and MCP-L genes), with more than 20 AH defined,25 although the majority are rare. In this study we report that relatively frequent RCAα block haplotypes, AH1 and AH3, are associated with pSS, an autoimmune disease with a high prevalence of antinuclear Ro/La autoantibodies, and which shares both clinical and genetic susceptibility overlap with SLE.

it requires and utilises these features to efficiently define the genomic polymorphism and AH within the region. The GMT RCAα block haplotyping has revealed extensive haplotypic polymorphism in this region (which also includes CR1-L, MCP and MCP-L genes), with more than 20 AH defined,25 although the majority are rare. In this study we report that relatively frequent RCAα block haplotypes, AH1 and AH3, are associated with pSS, an autoimmune disease with a high prevalence of antinuclear Ro/La autoantibodies, and which shares both clinical and genetic susceptibility overlap with SLE. The RCAα block contains CR1, CR1-L, MCP and MCP-L genes, and possibly also CR2, although the boundaries of the block have not been precisely defined. Erythrocyte CR1 in humans provides an important mechanism for the non-inflammatory clearance of immune complexes opsonised with C3a and C4b, and genetic variation in CR1 expression and/or function may influence both the predisposition to autoantibody-mediated disease (the clearance hypotheses)36 and self-perpetuating autoimmune reactivity resulting from prolonged, immune complex-mediated, type I interferon production.8 Although an acquired loss of erythrocyte CR1 expression has been observed in diseases such as SLE and pSS, the phenotypic high/low CR1 expression polymorphism is not associated with SLE,11 although the genetic basis for this is not well defined. CR1 expression is downregulated by immune complexes and upregulated by interferon γ in SLE,37 and in the mouse, CR1/CR2 expression (encoded by the same gene) is upregulated by BAFF.38 Therefore, genetic differences in the compound regulation of CR1 expression are likely to be strongly influenced by the complex inflammatory disease milieu.

wnregulated by immune complexes and upregulated by interferon γ in SLE,37 and in the mouse, CR1/CR2 expression (encoded by the same gene) is upregulated by BAFF.38 Therefore, genetic differences in the compound regulation of CR1 expression are likely to be strongly influenced by the complex inflammatory disease milieu. In addition to clearance of immune complexes, CR1 and CR2 play a direct, instructive role in setting the threshold for B cell responses to antigen39 and potentially, the loss of B cell tolerance,40 a fundamental step in the pathogenesis of autoimmune disease. Further, cross-linking of MCP (CD46) with the TCR on naïve CD4+ T cells induces regulatory T cells,41 which are critical in the balance between autoimmunity and tolerance.42 Finally, complement control proteins may function as receptors for ligands other than complement components. The CR2 receptor is of particular interest in this regard, and a recent study has demonstrated an association between a common CR2 single nucleotide polymorphism based haplotype and SLE.43. CR2 is not only a functional receptor for interferon α,44 but is also the Epstein–Barr virus (EBV) receptor. There has long been a strongly suspected link between EBV infection and systemic autoimmunity. Intriguingly, molecular mimicry between the EBNA-1 epitope of the EBV virus and the Ro autoantigen has been demonstrated in initiation of the anti-Ro response in patients who ultimately developed SLE.45

virus (EBV) receptor. There has long been a strongly suspected link between EBV infection and systemic autoimmunity. Intriguingly, molecular mimicry between the EBNA-1 epitope of the EBV virus and the Ro autoantigen has been demonstrated in initiation of the anti-Ro response in patients who ultimately developed SLE.45 Similar to HLA haplotypes, RCAα block haplotypes exert an influence on Ro/La autoantibody responses in patients with pSS. Importantly, there was an epistatic interaction between RCAα AH1 and HLA B8-DR3 in patients with pSS with Ro/La autoantibodies, which is indicative of a biological relationship between gene products of these haplotypes and disease pathways. The most likely basis for this epistasis is an interaction between the CR1 receptor and C4, one of its ligands. The genes for C4 are in the extended MHC region. HLA B8-DR3 and a relative C4 insufficiency (C4A*Q0,C4B*1)34 are both part of the 8.1 ancestral haplotype. The genetic structure of the C4 region is itself complex and highly polymorphic with both allelic and copy number variation of C4A and C4B genes.46 Genetic variation in C4 is known to contribute to disease susceptibility associated with MHC haplotypes47–49, but other MHC genes are also likely to be involved.34 48 In addition to systemic diseases such as pSS and SLE, the MHC 8.1 haplotype is associated with a number of organ-specific autoimmune diseases such as type 1 diabetes mellitus, Hashimoto’s thyroiditis, Graves’ disease, myasthenia gravis and multiple sclerosis. Evaluation of RCAα/MHC interactions in these diseases, and their association with susceptibility, autoantibody production and clinical phenotype would clearly be of great interest.

c autoimmune diseases such as type 1 diabetes mellitus, Hashimoto’s thyroiditis, Graves’ disease, myasthenia gravis and multiple sclerosis. Evaluation of RCAα/MHC interactions in these diseases, and their association with susceptibility, autoantibody production and clinical phenotype would clearly be of great interest. In addition to demonstrating the utility of the GMT approach for interrogating polymorphism in complex genomic regions, this study has also demonstrated that normal population variation in the RCAα block contributes substantially to susceptibility to systemic autoimmune disease. This is therefore direct evidence of the importance of regulation of complement activation in disease pathogenesis. The differing functional effects of the RCAα block haplotypes have not yet been defined and will be the focus of future research. These are likely to be pleiotropic and may well involve further epistatic genetic interactions. Although the task of unravelling these effects will be highly complex, we predict they will provide important insights into underlying autoimmune disease mechanisms. The authors thank Professor Tom Gordon and Ms Dimi Beroukas from the Department of Allergy, Immunology and Arthritis, Flinders Medical Centre, for CIEP autoantibody screening of patient sera. We also gratefully acknowledge the support of the Arthritis Foundation of Australia and the CYO Immunogenetics Research Foundation. Genetic Technologies Ltd provided infrastructural support.

om the Department of Allergy, Immunology and Arthritis, Flinders Medical Centre, for CIEP autoantibody screening of patient sera. We also gratefully acknowledge the support of the Arthritis Foundation of Australia and the CYO Immunogenetics Research Foundation. Genetic Technologies Ltd provided infrastructural support. Competing interests: The CY O’Connor Village has a financial interest in Genetic Technologies Ltd, which has filed a provisional patent application, including RCAα block typing (PCT/AU2006001232, “Identification of ancestral haplotypes and uses thereof”).

In 2006, recommendations for the management of ankylosing spondylitis (AS) were published under the umbrella of the ASsessment in AS international working group (ASAS), and the European League Against Rheumatism (EULAR).1 These recommendations concern all aspects of management of AS (box 1). The Institutes of Medicine define clinical practice guidelines or recommendations as “systematically developed statements to assist practitioner and patient decision about appropriate health care for specific clinical circumstances”. Clinical guidelines may induce small improvements, both in processes and in the outcomes of care.2 However, if recommendations are to have effect, it is necessary, after having published them, to facilitate their dissemination.3 Simple top-down dissemination of monodisciplinary guidelines alone is not effective.4–8 A more powerful strategy to change behaviour is to involve doctors directly.9 Implementation experts indicate that multistage involvement in the development of a guideline can be a positive contributor to effective implementation of guidelines.10 11 Therefore, an ASAS-initiated project was performed in 2006, involving practising rheumatologists in 10 countries. The objectives of this study were (a) to disseminate and (b) to evaluate conceptual agreement with, and (c) self-reported application as well as (d) potential barriers to the application of, the ASAS/EULAR recommendations among rheumatologists from 10 different countries.

In 2006, recommendations for the management of ankylosing spondylitis (AS) were published under the umbrella of the ASsessment in AS international working group (ASAS), and the European League Against Rheumatism (EULAR).1 These recommendations concern all aspects of management of AS (box 1). The Institutes of Medicine define clinical practice guidelines or recommendations as “systematically developed statements to assist practitioner and patient decision about appropriate health care for specific clinical circumstances”. Clinical guidelines may induce small improvements, both in processes and in the outcomes of care.2 However, if recommendations are to have effect, it is necessary, after having published them, to facilitate their dissemination.3 Simple top-down dissemination of monodisciplinary guidelines alone is not effective.4–8 A more powerful strategy to change behaviour is to involve doctors directly.9 Implementation experts indicate that multistage involvement in the development of a guideline can be a positive contributor to effective implementation of guidelines.10 11 Therefore, an ASAS-initiated project was performed in 2006, involving practising rheumatologists in 10 countries. The objectives of this study were (a) to disseminate and (b) to evaluate conceptual agreement with, and (c) self-reported application as well as (d) potential barriers to the application of, the ASAS/EULAR recommendations among rheumatologists from 10 different countries. Box 1: ASAS/EULAR recommendations for the management of ankylosing spondylitis (AS)1 Treatment of AS should be tailored according to current manifestations of the disease (axial, peripheral, entheseal, extra-articular symptoms and signs), level of current symptoms, clinical findings and prognostic indicators: disease activity/inflammation; pain; function, disability, handicap; structural damage, hip involvement, spinal deformities; general clinical status (age, sex, comorbidity, concomitant drugs); wishes and expectations of the patient.

oms and signs), level of current symptoms, clinical findings and prognostic indicators: disease activity/inflammation; pain; function, disability, handicap; structural damage, hip involvement, spinal deformities; general clinical status (age, sex, comorbidity, concomitant drugs); wishes and expectations of the patient. Disease monitoring of patients with AS should include a patient history (eg, questionnaires), clinical measures, laboratory tests and imaging, all according to the clinical presentation, as well as the ASAS core set*. The frequency of monitoring should be decided on an individual basis depending on symptoms, severity and drug treatment. Optimal management of AS requires a combination of non-pharmacological and pharmacological treatments. Non-pharmacological treatment of AS should include patient education and regular exercise. Individual and group physical therapy should be considered and patient associations and self-help groups may be useful. Non-steroidal anti-inflammatory drugs (NSAIDs) are recommended as first-line drug treatment for patients with AS who have pain and stiffness. In those with increased gastrointestinal risk, non-selective NSAIDs plus a gastroprotective agent, or a selective COX-2 inhibitor could be used. Analgesics, such as paracetamol and opioids, might be considered for pain control in patients in whom NSAIDs are insufficient, contraindicated and/or poorly tolerated. Corticosteroid injections directed to the local site of musculoskeletal inflammation may be considered. The use of systemic corticosteroids for axial disease is not supported by evidence.

Analgesics, such as paracetamol and opioids, might be considered for pain control in patients in whom NSAIDs are insufficient, contraindicated and/or poorly tolerated. Corticosteroid injections directed to the local site of musculoskeletal inflammation may be considered. The use of systemic corticosteroids for axial disease is not supported by evidence. There is no evidence for the efficacy of disease-modifying antirheumatic drugs (DMARDs), including sulfasalazine and methotrexate, for the treatment of axial disease. Sulfasalazine may be considered in patients with peripheral arthritis. Anti-tumour necrosis factor (TNF) treatment should be given to patients with persistently high disease activity despite conventional treatments according to the ASAS recommendations. There is no evidence to support the obligatory use of DMARDs before, or concomitant with, anti-TNF treatment in patients with axial disease. Total hip arthroplasty should be considered in patients with refractory pain or disability and radiographic evidence of structural damage, independent of age. Spinal surgery—for example, corrective osteotomy and stabilisation procedures— may be of value in selected patients. *The ASAS core set includes domains on axial, peripheral and enthesopathological manifestations. One or more specific instruments are recommended for each domain. METHODS This project was initiated by ASAS, supervised by Maxime Dougados (France) and Tore K Kvien (Norway) and financially supported by Wyeth Europa pharmaceutical company.

*The ASAS core set includes domains on axial, peripheral and enthesopathological manifestations. One or more specific instruments are recommended for each domain. METHODS This project was initiated by ASAS, supervised by Maxime Dougados (France) and Tore K Kvien (Norway) and financially supported by Wyeth Europa pharmaceutical company. A questionnaire was prepared which included the text of the recommendations, demographic variables (age, sex, academic position or not, number of years of practice and mean number of rheumatic patients and patients with AS seen a month) and a numerical rating scale (NRS) from 1 to 10 for conceptual agreement and application of each recommendation (10 indicated maximal agreement and maximal application). The text was: “Do you conceptually agree with this recommendation?” and “Are you applying this recommendation in your daily practice?”. The questionnaire also included a list of potential barriers to the application of the recommendation (the rheumatologist could tick as many barriers as applicable). The barriers were different for each recommendation and were selected by the authors on the basis of clinical experience. Respondents could also volunteer additional barriers.

o included a list of potential barriers to the application of the recommendation (the rheumatologist could tick as many barriers as applicable). The barriers were different for each recommendation and were selected by the authors on the basis of clinical experience. Respondents could also volunteer additional barriers. Ten countries participated: Arabian Gulf, Belgium, Czech Republic, France, Germany, Italy, the Netherlands, Norway, Spain and the United Kingdom (UK). For each country, a national investigator (member of ASAS) translated the questionnaire and sent it out to rheumatologists between March and September 2006. The list of rheumatologists who would receive the questionnaire was freely chosen by the national investigator. The rheumatologists were asked to complete and return the questionnaire to the national investigator. Statistical analysis was performed by LG on anonymous data with knowledge of country; analysis was descriptive for conceptual agreement and self-reported application. Multivariate logistic regression analyses were also performed. In the first, the dependent variable was conceptual agreement per doctor, binarised at the mean conceptual agreement for that recommendation. In the second analysis the dependent variable was the individual difference between agreement and application (ie, agreement minus application for a given doctor), also binarised at the mean. The independent variables entered in both analyses were country and the characteristics of the rheumatologist. Statistical analyses were performed with SAS version 9.0.

iable was the individual difference between agreement and application (ie, agreement minus application for a given doctor), also binarised at the mean. The independent variables entered in both analyses were country and the characteristics of the rheumatologist. Statistical analyses were performed with SAS version 9.0. RESULTS Response rate and participants A total of 7206 questionnaires were sent out. The response rate varied across countries from 49% (Italy) to 11% (Spain), but the number of mailed questionnaires also varied widely between countries (table 1). Thus, the total number of questionnaires returned and analysed was 1507 (21% of all questionnaires), of which 413 (27%) of all analysed questionnaires were from France and 301 (20%) were from Germany (table 1).

taly) to 11% (Spain), but the number of mailed questionnaires also varied widely between countries (table 1). Thus, the total number of questionnaires returned and analysed was 1507 (21% of all questionnaires), of which 413 (27%) of all analysed questionnaires were from France and 301 (20%) were from Germany (table 1). Table 1 Response rate and characteristics of the 1507 European rheumatologists who were respondents to the questionnaire evaluating the ASAS/EULAR recommendations for the management of ankylosing spondylitis (AS) Characteristics Arabian Gulf (n = 89) Belgium (n = 84) Czech Republic (n = 73) France (n = 413) Germany (n = 301) Italy (n = 172) Netherlands (n = 73) Norway (n = 72) Spain (n = 119) UK (n = 111) All countries (n = 1507) Respondents (%) 28.5 28.8 20.9 16.0 27.0 49.1 30.2 36.0 10.9 17.3 20.9 Distribution across countries (% of all respondents) 5.9 5.6 4.8 27.4 20.0 11.4 4.8 4.8 7.9 7.4 100 Men (%) 79.8 62.2 27.4 60.1 63.1 68.6 65.8 47.2 60.5 75.7 62.1 Age (years), mean (SD) 46.4 (6.5) 49.1 (10.8) 47.8 (8.8) 48.3 (8.9) 50.9 (8.5) 48.7 (10.2) 51.1 (7.7) 50.0 (9.5) 46.8 (9.3) 45.8 (8.4) 48.6 (9.1) Number of years of practice, mean (SD) 14.1 (7.6) 18.3 (10.7) 15.8 (7.9) 18.5 (8.7) 15.6 (9.6) 17.6 (9.9) 16.0 (8.6) 14.1 (9.2) 18.0 (9.9) 14.7 (8.3) 16.8 (9.3) Academic position (%) 65.2 27.4 19.2 13.1 77.7 22.1 19.2 25.0 28.6 19.8 33.8 Number of rheumatic patients seen a month, mean (SD) 188.9 (133.1) 220.3 (126.4) 268.6 (192.9) 275.1 (159.9) 224.7 (202.7) 160.8 (120.5) 277.7 (162.7) 123.7 (222.2) 302.1 (143.7) 217.6 (153.7) 234.4 (171.7)

.2) 18.0 (9.9) 14.7 (8.3) 16.8 (9.3) Academic position (%) 65.2 27.4 19.2 13.1 77.7 22.1 19.2 25.0 28.6 19.8 33.8 Number of rheumatic patients seen a month, mean (SD) 188.9 (133.1) 220.3 (126.4) 268.6 (192.9) 275.1 (159.9) 224.7 (202.7) 160.8 (120.5) 277.7 (162.7) 123.7 (222.2) 302.1 (143.7) 217.6 (153.7) 234.4 (171.7) Range 8–800 3–500 20–750 20–1200 5–1000 5–800 40–850 0–1800 40–850 10–1250 0–1800 Number of patients with AS seen per month, mean (SD) 10.1 (32.7) 15.4 (13.5) 16.8 (16.7) 12.0 (25.9) 25.0 (43.3) 9.9 (12.5) 23.9 (61.1) 12.7 (10.0) 18.4 (16.6) 12.9 (9.8) 15.6 (29.8) Range 1–300 1–50 2–70 1–415 1–600 1–100 1–510 0–40 1–100 2–60 0–600 Table 1 shows the characteristics of the rheumatologists. Of respondents, 62% were men, mean (SD) age 49 (9) years, mean (SD) years of practice 17 (9), and 34% reported having an academic position. The mean (SD) number of patients seen a month was 234 (172) and the mean (SD) number of patients with AS seen a month was 15.6 (29.8).

s the characteristics of the rheumatologists. Of respondents, 62% were men, mean (SD) age 49 (9) years, mean (SD) years of practice 17 (9), and 34% reported having an academic position. The mean (SD) number of patients seen a month was 234 (172) and the mean (SD) number of patients with AS seen a month was 15.6 (29.8). Conceptual agreement Conceptual agreement with the recommendations was evaluated separately for each recommendation, and was generally high. Mean (SD) for all recommendations and all countries was 8.9 (0.9) (table 2), and more than 80% of all rheumatologists had an agreement ⩾7 for all recommendations combined (table 2). Agreement was highest for recommendation 3 (“optimal management of AS requires a combination of non-pharmacological and pharmacological treatments”) and recommendation 1 (“tailoring of treatment”) (mean (SD) 9.5 (1.1) and 9.3 (1.2), respectively). Agreement was lowest for recommendation 6 (“analgesics, such as paracetamol and opioids, might be considered”) with a mean (SD) value of 8.3 (2.0).

res a combination of non-pharmacological and pharmacological treatments”) and recommendation 1 (“tailoring of treatment”) (mean (SD) 9.5 (1.1) and 9.3 (1.2), respectively). Agreement was lowest for recommendation 6 (“analgesics, such as paracetamol and opioids, might be considered”) with a mean (SD) value of 8.3 (2.0). Table 2 Scores for self-reported conceptual agreement with each ASAS/EULAR recommendation, scores for self-reported application of each recommendation and difference between conceptual agreement and self-reported application (agreement – application) of each recommendation across countries Recommendation Agreement All countries Application All countries Difference between agreement and application Arabian Gulf Belgium Czech R. France Germany Italy Netherlands Norway Spain UK All countries 1 9.3 (1.2) 8.7 (1.4) 0.9 (1.2) 0.3 (1.6) 1.1 (1.4) 0.3 (1.2) 0.5 (1.0) 0.6 (1.4) 0.5 (1.6) 0.8 (1.5) 0.7 (1.1) 1.0 (2.0) 0.6 (1.4) [97.2] [93.4] [36.2] 10 (1–10) 9 (1–10) 0 (–9 to 9) 2 8.9 (1.5) 7.7 (2.0) 1.2 (1.5) 1.3 (2.4) 1.2 (1.4) 0.8 (1.4) 0.6 (1.2) 1.4 (1.6) 1.6 (1.9) 1.2 (1.9) 1.

s Norway Spain UK All countries 1 9.3 (1.2) 8.7 (1.4) 0.9 (1.2) 0.3 (1.6) 1.1 (1.4) 0.3 (1.2) 0.5 (1.0) 0.6 (1.4) 0.5 (1.6) 0.8 (1.5) 0.7 (1.1) 1.0 (2.0) 0.6 (1.4) [97.2] [93.4] [36.2] 10 (1–10) 9 (1–10) 0 (–9 to 9) 2 8.9 (1.5) 7.7 (2.0) 1.2 (1.5) 1.3 (2.4) 1.2 (1.4) 0.8 (1.4) 0.6 (1.2) 1.4 (1.6) 1.6 (1.9) 1.2 (1.9) 1. (1.6) 2.0 (1.9) 1.1 (1.6) [92.5] [75.6] [51.1] 9 (1–10) 8 (1–10) 1 (−8 to 9) 3 9.5 (1.1) 8.8 (1.5) 1.3 (1.7) 0.3 (1.7) 0.3 (0.9) 0.5 (1.3) 0.6 (1.2) 1.1 (1.4) 0.2 (0.5) 0.7 (1.2) 1.1 (1.6) 0.9 (1.6) 0.7 (1.4) [97.7] [91.1] [35.0] 10 (1–10) 9 (2–10) 0 (–9 to 8) 4 8.9 (1.6) 7.5 (2.1) 1.9 (2.1) 1.0 (1.6) 1.0 (1.4) 1.4 (1.7) 1.1 (1.7) 2.2 (2.1) 0.5 (0.9) 0.9 (1.0) 1.9 (2.0) 1.5 (1.8) 1.4 (1.8) [90.8] [69.7] [56.5] 10 (1–10) 8 (1–10) 0 (−9 to 9) 5 9.2 (1.4) 8.9 (1.5) 0.3 (0.8) 0.7 (2.0) 0.4 (1.0) 0.1 (0.7) 0.4 (0.9) 0.2 (0.8) 0.1 (0.6) 0.4 (0.8) 0.2 (0.7) 0.2 (1.1) 0.3 (0.9) [94.5] [92.4] [19.7] 10 (1–10) 9 (1–10) 0 (−9 to 7) 6 8.3 (2.0) 7.8 (2.3) 0.7 (1.4) 0.5 (1.6) 1.0 (1.5) 0.3 (1.1) 0.6 (1.2) 0.7 (1.1) 0.6 (1.1) 0.9 (1.2) 0.6 (1.1) 0.4 (0.9) 0.5 (1.2) [81.8] [73.9] [32.5] 9 (1–10) 8 (1–10) 0 (−9 to 7) 7 8.6 (1.8) 8.1 (2.1) 0.6 (1.1) 0.4 (1.7) 0.5 (1.3) 0.4 (1.2) 0.8 (1.5) 0.9 (1.6) 0.5 (1.2) 0.7 (1.1) 0.7 (1.4) 0.6 (1.1) 0.6 (1.4) [87.5] [79.6] [32.1] 9 (1–10) 9 (1–10) 0 (−8 to 9) 8 8.6 (1.9) 8.3 (2.1) 0.3 (1.2) 0.2 (2.1) 0.3 (0.7) 0.3 (1.1) 0.5 (1.1) 0.2 (1.1) 0.0 (1.0) 0.5 (0.9) 0.4 (1.3) 0.2 (1.0) 0.3 (1.2) [79.7] [82.7] [22.8] 9 (1–10) 9 (1–10) 0 (−9 to 9) 9 9.0 (1.6) 8.0 (2.3) 1.6 (2.3) 0.4 (1.8) 2.7 (2.9) 0.6 (1.3) 0.9 (1.8) 0.9 (1.9) 0.4 (0.9) 0.8 (1.2) 0.8 (1.5) 2.8 (3.0) 1.0 (2.0) [96.6] [79.1] [37.8] 10 (1–10) 9 (1–10) 0 (−9 to 9) 10 8.6 (1.7) 7.8 (2.2) 1.2 (1.6) 0.7 (1.5) 1.1 (1.7) 0.6 (1.4) 0.7 (1.4) 1.4 (1.9) 0.6 (1.0) 0.6 (1.2) 1.0 (1.5) 0.5 (1.0) 0.8 (1.5) [87.1] [74.5] [36.7] 9 (1–10) 8 (1–10) 0 (−9 to 9) Mean 8.9 (0.9) 8.2 (1.0) 0.9 (0.8) 0.5 (1.3) 0.9 (0.8) 0.5 (0.7) 0.7 (0.7) 1.0 (0.8) 0.5 (0.5) 0.8 (0.7) 0.9 (0.7) 1.0 (0.9) 0.7 (0.8) [90.8] [81.7] [30.2] 9 (1–10) 8 (1–10) 0.6 (−6.6 to 4.4) See box 1 for the text of the recommendations.

1.0 (1.5) 0.5 (1.0) 0.8 (1.5) [87.1] [74.5] [36.7] 9 (1–10) 8 (1–10) 0 (−9 to 9) Mean 8.9 (0.9) 8.2 (1.0) 0.9 (0.8) 0.5 (1.3) 0.9 (0.8) 0.5 (0.7) 0.7 (0.7) 1.0 (0.8) 0.5 (0.5) 0.8 (0.7) 0.9 (0.7) 1.0 (0.9) 0.7 (0.8) [90.8] [81.7] [30.2] 9 (1–10) 8 (1–10) 0.6 (−6.6 to 4.4) See box 1 for the text of the recommendations. Agreement (first column): the question was: “Do you conceptually agree with this recommendation?” (1–10 scale where 10 indicates high agreement). Results are presented for all countries (pooled results) as mean (SD) agreement with recommendation, [% rheumatologists with an agreement ⩾7], median (range). Application (second column): The question was: “Are you applying this recommendation in your daily practice?” (1–10 scale where 10 indicates high application). Results are presented for all countries (pooled results) as mean (SD) self-declared application of recommendation, [% rheumatologists with an application ⩾7], median (range). Individual difference (agreement-application): results are presented separately for each country and (last column) as pooled results, as mean (SD). Last column also shows [% rheumatologists with a difference ⩾1] and median (range). Agreement varied across countries; mean agreement with all recommendations was highest in the Czech Republic (mean (SD) 9.4 (0.5)) and in Norway (mean (SD) 9.2 (0.6)) and was lowest in Belgium (mean (SD) 8.5 (1.4)).

Individual difference (agreement-application): results are presented separately for each country and (last column) as pooled results, as mean (SD). Last column also shows [% rheumatologists with a difference ⩾1] and median (range). Agreement varied across countries; mean agreement with all recommendations was highest in the Czech Republic (mean (SD) 9.4 (0.5)) and in Norway (mean (SD) 9.2 (0.6)) and was lowest in Belgium (mean (SD) 8.5 (1.4)). Multivariate analyses were performed to explain agreement with the recommendations. For all recommendations pooled, the country of the investigator was not statistically associated with agreement. Only female gender (odds ratio (OR) = 1.33, 95% confidence interval (CI) 1.03 to 1.72, p = 0.028) and academic position (OR = 1.31, 95% CI 1.02 to 1.70, p = 0.032) were predictive of higher agreement. However, this statistical significance did not reflect clinical relevance. Mean (SD) agreement scores were 8.93 (0.930) versus 8.84 (0.90) for female versus male rheumatologists, and 8.95 (0.81) versus 8.80 (0.96) for academics versus non-academics, thus high for all doctors.

0.032) were predictive of higher agreement. However, this statistical significance did not reflect clinical relevance. Mean (SD) agreement scores were 8.93 (0.930) versus 8.84 (0.90) for female versus male rheumatologists, and 8.95 (0.81) versus 8.80 (0.96) for academics versus non-academics, thus high for all doctors. Application of recommendations Self-declared application of recommendations was also high, but lower than conceptual agreement (mean (SD) 8.2 (1.0)) (table 2). Application scores higher than 7 were reported by 81.7% of the doctors. Self-reported application was highest for recommendation 5 (“non-steroidal anti-inflammatory drugs (NSAIDs) are recommended as first-line drug treatment”) with mean (SD) score 8.9 (1.5). Self-reported application was lowest for recommendation 4 (“non-pharmacological treatment of AS should include patient education and regular exercise. Individual and group physical therapy should be considered and patient associations and self-help groups may be useful”) and recommendation 2 (“disease monitoring of patients with AS”) with mean (SD) scores of 7.5 (2.1) and 7.7 (2.0), respectively. Application varied across countries; mean (SD) self-reported application of all recommendations was highest in the Czech Republic (8.5 (0.9)) and was lowest in the UK (7.8 (0.9)).

Application of recommendations Self-declared application of recommendations was also high, but lower than conceptual agreement (mean (SD) 8.2 (1.0)) (table 2). Application scores higher than 7 were reported by 81.7% of the doctors. Self-reported application was highest for recommendation 5 (“non-steroidal anti-inflammatory drugs (NSAIDs) are recommended as first-line drug treatment”) with mean (SD) score 8.9 (1.5). Self-reported application was lowest for recommendation 4 (“non-pharmacological treatment of AS should include patient education and regular exercise. Individual and group physical therapy should be considered and patient associations and self-help groups may be useful”) and recommendation 2 (“disease monitoring of patients with AS”) with mean (SD) scores of 7.5 (2.1) and 7.7 (2.0), respectively. Application varied across countries; mean (SD) self-reported application of all recommendations was highest in the Czech Republic (8.5 (0.9)) and was lowest in the UK (7.8 (0.9)). Difference between agreement and self-reported application We calculated, for each recommendation, the difference for each doctor between self-declared agreement and application—that is, agreement minus application. This score could range from 0 (no difference between agreement and application) to 10 (total agreement but no application). Theoretically, negative scores could also be obtained (if application was higher than agreement).

each doctor between self-declared agreement and application—that is, agreement minus application. This score could range from 0 (no difference between agreement and application) to 10 (total agreement but no application). Theoretically, negative scores could also be obtained (if application was higher than agreement). The differences between agreement and application varied across recommendations and across countries (table 2). The difference between agreement and application had mean values above 1.0 for three recommendations. The highest values were 1.4 (1.8), for recommendation 4 (“non-pharmacological treatment of AS”), 1.1 (1.6), for recommendation 2 (“disease monitoring of patients with AS”) and 1.0 (2.0), for recommendation 9 (“anti-tumour necrosis factor (TNF)”). The lowest values were observed for recommendations 5 and 8 (mean 0.3 (0.9) and 0.3 (1.2), respectively). There were no negative scores. The largest overall country differences between agreement and application were recorded in Italy (1.0 (0.8)) and the UK (1.0 (0.9)). For recommendation 4, mean difference scores were >1 for all countries except the Netherlands and Norway, but were >2 only for Italy. For recommendation 2, mean difference scores were >1 for all countries except France and Germany, but were >2 only for the UK. For recommendation 9 (“anti-TNF”), mean difference scores were <1 for all countries except UK (2.8 (3.0)), Czech Republic (2.7 (2.9)) and the Arabian Gulf (1.6 (2.3)).

t were >2 only for Italy. For recommendation 2, mean difference scores were >1 for all countries except France and Germany, but were >2 only for the UK. For recommendation 9 (“anti-TNF”), mean difference scores were <1 for all countries except UK (2.8 (3.0)), Czech Republic (2.7 (2.9)) and the Arabian Gulf (1.6 (2.3)). Multivariate analyses were performed to explain this difference between agreement and application. Country was the only independent variable which was significantly associated with the difference between agreement and application (p<0.001) (data not shown). Barriers to application of recommendations Table 3 shows items presented as potential barriers and ticked by more than 25% of rheumatologists. A high proportion of doctors felt there was no specific barrier to the application of recommendations 1, 5, 6, 7, 8 and 10 (more than 40% ticked the item “there is no specific barrier”). The potential presence of barriers for application was most frequently reported for recommendation 4 (no barrier, 19.7%), 2, 3 and 9 (no barrier, 30.3%, 32.0% and 34.5%, respectively). Table 3 Barriers to application of recommendations Recomm. Item Arabian Gulf Belgium Czech R. France Germany Italy Netherlands Norway Spain UK All countries 1 There is no barrier 29 (32.6) 37 (44.1) 23 (31.5) 234 (56.7) 151 (50.2) 66 (38.4) 31 (42.5) 31 (43.1) 53 (44.5) 22 (19.8) 677 (44.9) Consultation time to conduct assessment fully 21 (23.6) 22 (26.2) 19 (26.0) 55 (13.3) 72 (23.9) 47 (27.3) 33 (45.2) 20 (27.8) 59 (49.6) 45 (40.5) 393 (26.1)

orway Spain UK All countries 1 There is no barrier 29 (32.6) 37 (44.1) 23 (31.5) 234 (56.7) 151 (50.2) 66 (38.4) 31 (42.5) 31 (43.1) 53 (44.5) 22 (19.8) 677 (44.9) Consultation time to conduct assessment fully 21 (23.6) 22 (26.2) 19 (26.0) 55 (13.3) 72 (23.9) 47 (27.3) 33 (45.2) 20 (27.8) 59 (49.6) 45 (40.5) 393 (26.1) 2 There is no barrier 16 (18.0) 20 (23.8) 17 (23.3) 145 (35.1) 140 (46.5) 46 (26.7) 16 (21.9) 25 (34.7) 18 (15.1) 13 (11.7) 456 (30.3) Lack of time to conduct frequent monitoring 33 (37.1) 42 (50.0) 33 (45.2) NA 58 (19.3) 68 (39.5) 38 (52.1) 29 (40.3) 84 (70.6) 79 (71.2) 464 (42.4)* 3 There is no barrier 28 (31.5) 33 (39.3) 13 (17.8) 138 (33.4) 107 (35.6) 32 (18.6) 50 (68.5) 30 (41.7) 26 (21.9) 25 (22.5) 482 (32.0) Insufficient number of qualified health professionals—eg, physiotherapists 23 (25.8) 11 (13.1) 18 (24.7) 70 (16.9) 44 (14.6) 97 (56.4) 4 (5.5) 12 (16.7) 67 (56.3) 45 (40.5) 391 (25.9) Lack of facilities for education 19 (21.4) 10 (11.9) 19 (26.0) 122 (29.5) 59 (19.6) 62 (36.1) 2 (2.7) 11 (15.3) 60 (50.4) 14 (12.6) 378 (25.1)

A and PsA where it occurred in plasma cells. This HSPG has been used as a marker of epithelial cells and plasma cells.40–44 In the present study it was a selective marker for the latter cell type, which is in agreement with other studies that used syndecan-1 to identify these cells in chronically inflamed synovia.43 44 Syndecan-2 was present mainly in blood vessels where it occurred in endothelial cells, pericytes and smooth muscle cells. Expression of syndecan-2 has been shown in mesenchymal, neuronal and cancer cells.45 46 In line with our results this HSPG has also been found in endothelial cells, for example in the brain where it is highly expressed.47 There is evidence that syndecan-2 is involved in angiogenesis that may be related to the interaction of the HSPG with vascular endothelial growth factor.48 This is of particular interest since in the RA synovium angiogenesis is a significant pathological event responsible for its enlargement and invasive properties, and suggests an involvement of syndecan-2 in these mechanisms.

icient number of qualified health professionals—eg, physiotherapists 23 (25.8) 11 (13.1) 18 (24.7) 70 (16.9) 44 (14.6) 97 (56.4) 4 (5.5) 12 (16.7) 67 (56.3) 45 (40.5) 391 (25.9) Lack of facilities for education 19 (21.4) 10 (11.9) 19 (26.0) 122 (29.5) 59 (19.6) 62 (36.1) 2 (2.7) 11 (15.3) 60 (50.4) 14 (12.6) 378 (25.1) 4 There is no barrier 10 (11.2) 16 (19.1) 16 (21.9) 72 (17.4) 79 (26.25) 14 (8.14) 34 (46.6) 23 (31.9) 9 (7.6) 24 (21.6) 297 (19.7) Lack of patient compliance with recommendations 49 (55.1) 51 (60.7) 47 (64.4) 214 (51.8) 121 (40.2) 74 (43.0) 24 (32.9) 28 (38.9) 63 (52.9) 39 (35.1) 710 (47.1) Lack of facilities for education 36 (40.5) 18 (21.4) 21 (28.8) 158 (38.3) 99 (32.9) 59 (34.3) 4 (5.5) 15 (20.8) 60 (50.4) 26 (23.4) 496 (32.9) Insufficient number of qualified health professionals—eg, physiotherapists 35 (39.3) 16 (19.1) 28 (38.4) 73 (17.7) 55 (18.3) 104 (60.5) 4 (5.5) 17 (23.6) 67 (56.3) 56 (50.5) 455 (30.2) 5 There is no barrier 42 (47.2) 10 (11.9) 30 (41.1) 223 (54.0) 141 (46.8) 44 (25.6) 42 (57.5) 12 (16.7) 46 (38.7) 38 (34.2) 628 (41.7) Concerns about the safety of long-term use of NSAIDs/COX-2 inhibitor 34 (38.2) 50 (59.5) 37 (50.7) 117 (28.3) 95 (31.6) 73 (42.4) 17 (23.3) 42 (58.3) 48 (40.3) 57 (51.4) 570 (37.8) 6 There is no barrier 29 (32.6) 17 (20.2) 25 (34.3) 223 (54.0) 160 (53.2) 79 (45.9) 41 (56.2) 11 (15.3) 64 (53.8) 65 (58.6) 714 (47.4) Fear of addiction/tolerance to opioids with long-term use 38 (42.7) 50 (59.5) 35 (47.9) 99 (24.0) 61 (20.3) 46 (26.7) 14 (19.2) 51 (70.8) 25 (21.0) 31 (27.9) 450 (29.9)

5 There is no barrier 42 (47.2) 10 (11.9) 30 (41.1) 223 (54.0) 141 (46.8) 44 (25.6) 42 (57.5) 12 (16.7) 46 (38.7) 38 (34.2) 628 (41.7) Concerns about the safety of long-term use of NSAIDs/COX-2 inhibitor 34 (38.2) 50 (59.5) 37 (50.7) 117 (28.3) 95 (31.6) 73 (42.4) 17 (23.3) 42 (58.3) 48 (40.3) 57 (51.4) 570 (37.8) 6 There is no barrier 29 (32.6) 17 (20.2) 25 (34.3) 223 (54.0) 160 (53.2) 79 (45.9) 41 (56.2) 11 (15.3) 64 (53.8) 65 (58.6) 714 (47.4) Fear of addiction/tolerance to opioids with long-term use 38 (42.7) 50 (59.5) 35 (47.9) 99 (24.0) 61 (20.3) 46 (26.7) 14 (19.2) 51 (70.8) 25 (21.0) 31 (27.9) 450 (29.9) 7 There is no barrier 37 (41.6) 37 (44.1) 40 (54.8) 182 (44.1) 149 (49.5) 64 (37.2) 45 (61.6) 29 (40.3) 59 (49.6) 66 (59.5) 708 (47.0) Patient concerns about use of corticosteroid injections 43 (48.3) 31 (36.9) 27 (37.0) 157 (38.0) 78 (25.9) 55 (32.0) 11 (15.1) 25 (34.7) 36 (30.3) 21 (18.9) 484 (32.1) 8 There is no barrier 46 (51.7) 42 (50.0) 47 (64.4) 255 (61.7) 175 (58.1) 97 (56.4) 54 (74.0) 35 (48.6) 70 (58.8) 66 (59.5) 887 (58.9) 9 There is no barrier 21 (23.6) 24 (28.6) 5 (6.9) 175 (42.4) 119 (39.5) 55 (32.0) 43 (58.9) 22 (30.6) 49 (41.2) 13 (11.7) 526 (34.9) Insufficient funding 41 (46.1) 6 (7.1) 57 (78.1) 32 (7.8) 102 (33.9) 38 (22.1) 3 (4.1) 19 (26.4) 20 (16.8) 73 (65.8) 391 (25.9) Administrative burden associated with anti-TNF treatment 26 (29.2) 24 (28.6) 24 (32.9) 101 (24.5) 48 (16.0) 74 (43.0) 13 (17.8) 9 (12.5) 39 (32.8) 28 (25.2) 386 (25.6)

.2) 13 (11.7) 526 (34.9) Insufficient funding 41 (46.1) 6 (7.1) 57 (78.1) 32 (7.8) 102 (33.9) 38 (22.1) 3 (4.1) 19 (26.4) 20 (16.8) 73 (65.8) 391 (25.9) Administrative burden associated with anti-TNF treatment 26 (29.2) 24 (28.6) 24 (32.9) 101 (24.5) 48 (16.0) 74 (43.0) 13 (17.8) 9 (12.5) 39 (32.8) 28 (25.2) 386 (25.6) 10 There is no barrier 27 (30.3) 26 (30.9) 23 (31.5) 193 (46.7) 184 (61.1) 61 (35.5) 48 (65.8) 34 (47.2) 44 (36.9) 68 (61.3) 708 (47.0) Items ticked by more than 25% of rheumatologists are presented. For each item: number (percentage) where the item was ticked. *Percentage of available data. NA, not available, NSAIDs, non-steroidal anti-inflammatory drugs; Recomm., recommendation; TNF, tumour necrosis factor. The potential barrier “I am not familiar with this recommendation” was included in the questionnaire for all recommendations but was infrequently ticked (0.2% to 6% for all recommendations).

*Percentage of available data. NA, not available, NSAIDs, non-steroidal anti-inflammatory drugs; Recomm., recommendation; TNF, tumour necrosis factor. The potential barrier “I am not familiar with this recommendation” was included in the questionnaire for all recommendations but was infrequently ticked (0.2% to 6% for all recommendations). For recommendations 1 and 2, the most frequent barrier (26.1% and 42.4% of doctors, respectively) was “lack of time”. More than 70% of doctors in Spain and the UK reported this barrier for recommendation 2. For recommendation 4, the most frequent barrier (47.1% of doctors) was “lack of patient compliance with recommendations”. The second most frequent barrier (32.9%) was “lack of facilities for education”, particularly in Spain (50.4%). The most frequent barrier (37.8%) for recommendation 5 was “concerns about the safety of long-term use of NSAIDs/COX-2 inhibitors”, but 41.7% reported no specific barrier. For recommendation 6, 47.4% of doctors reported no specific barrier; but 29.9% reported that “fear of addiction/tolerance to opioids with long-term use” was a barrier. The most frequent barrier (32.1%) for recommendation 7 was “patient concerns about use of corticosteroid injections” but 47.0% of doctors reported no specific barrier for the application of this recommendation.

c barrier; but 29.9% reported that “fear of addiction/tolerance to opioids with long-term use” was a barrier. The most frequent barrier (32.1%) for recommendation 7 was “patient concerns about use of corticosteroid injections” but 47.0% of doctors reported no specific barrier for the application of this recommendation. For recommendation 9 regarding use of anti-TNF drugs in patients with AS, the most frequent barriers were “insufficient funding” (25.9%), and “administrative burden associated with anti-TNF treatment” (25.6%). Major differences across countries were observed for this item. Barriers related to funding were frequently reported in the Czech Republic (78.1%) and the UK (65.8%), whereas Italian doctors frequently reported barriers related to administrative burden (43.0%). DISCUSSION This large project has helped the dissemination of the ASAS/EULAR recommendations for the management of AS: 7206 rheumatologists received the questionnaire and thus also the text of the recommendations, and 1507 answered and are assumed to have at least read the recommendations. The results from the survey showed that conceptual agreement with the recommendations was very high (mean 8.9), and self-declared application was also high, though somewhat lower (mean 8.2).

naire and thus also the text of the recommendations, and 1507 answered and are assumed to have at least read the recommendations. The results from the survey showed that conceptual agreement with the recommendations was very high (mean 8.9), and self-declared application was also high, though somewhat lower (mean 8.2). Raising implementation of recommendations is not an easy task. Implementation experts indicate that the effect of issued recommendations on clinical practice is improved if practising clinicians actively contribute to the development of the recommendations.12 If this involvement is not feasible, creation of a sense of “ownership” through a process of dissemination is essential for successful implementation of recommendations in clinical practice.10 12 We have paid specific attention to these aspects, first, by involving worldwide representation of rheumatologists in the elaboration 1, and second, by focusing specifically on the doctors’ acceptance of the recommendations by including them in the validation procedure, and giving them the opportunity to express their agreement and practical applicability of each recommendation.

involving worldwide representation of rheumatologists in the elaboration 1, and second, by focusing specifically on the doctors’ acceptance of the recommendations by including them in the validation procedure, and giving them the opportunity to express their agreement and practical applicability of each recommendation. Agreements with the 10 recommendations were generally high. The ASAS group is possibly considered as an opinion leader to such extent that the rheumatologists did not feel justified, or diplomatic, to disagree with recommendations that were prepared by experts they trusted. Another factor that might have had an impact on the validity of the results is the method of data collection. Some authors have recommended feedback meetings rather than questionnaires as an effective method to create involvement among doctors during a validation procedure.9 However, meetings are time consuming and costly and their impact may be only moderate.9 Furthermore, some doctors might have had comments that were not expressed since they considered that their input would not have any impact in the final recommendations. Another potential explanation is related to selection of respondents by their agreement. This is a possible bias which needs to be discussed because the response rate (21%) is very low, and the percentage of academic doctors is high (34%). However, it should be noted that these response rates in the context of an implementation project were reasonable. Response rates varied across countries, and were higher in countries where fewer rheumatologists were solicited, presumably because in these countries a selection was operated before sending the questionnaires—that is, these doctors were all treating patients with AS and wanted to answer. The reason for this low response rate, in particular, in certain countries is probably a suboptimal selection of the rheumatologists receiving the questionnaires. The list of recipients was freely determined by the principal investigator for each country, and in some countries a national society of rheumatology members’ list was used. This selection method may be insufficiently directed towards interested rheumatologists, involved with patients with AS, which may explain the low response rate. It also may explain why some of the responding rheumatologists reported having a small number of patients with AS.

ety of rheumatology members’ list was used. This selection method may be insufficiently directed towards interested rheumatologists, involved with patients with AS, which may explain the low response rate. It also may explain why some of the responding rheumatologists reported having a small number of patients with AS. The application of recommendations for management is generally important for the quality of patient care. Self-reported application of these recommendations was generally high across Europe. The implementation of guidelines and recommendations remains a challenge world wide, as barriers exist at several levels. These barriers may be generic, such as national limitations to prescriptions of expensive drugs or a general lack of adequate resources and/or poor infrastructure. Other potential barriers include the organisational level, the healthcare provider and patient factors. Some barriers are, however, potentially correctable, and the goal of the implementation of the recommendations for AS management is to translate evidence-based AS management recommendations into “real-life” practice which ultimately will lead to improved health status for patients with AS. It should be noted, however, that this study did not allow us to truly assess application of the ASAS-EULAR recommendations, but only the self-reported application. The possibility of a gap between self-declared application of recommendations, and their true application cannot be excluded.

status for patients with AS. It should be noted, however, that this study did not allow us to truly assess application of the ASAS-EULAR recommendations, but only the self-reported application. The possibility of a gap between self-declared application of recommendations, and their true application cannot be excluded. The differences between agreement and application are of particular interest, since these differences reflect areas in which the doctors agreed with the recommendations, but reported they did not implement them. This section of the results also disclosed differences between countries, which is important since equality in access to healthcare should be a general goal in Europe The difference between agreement and application was high for recommendations 2, 4 and 9 and the overall country differences were highest in Italy and the UK. Recommendation 2 was deemed difficult to apply because of lack of time and this concern was especially reported by doctors in the UK. The country difference between agreement and application was pronounced for recommendation 9 (“anti-TNF”), and the largest discrepancy between agreement and the possibility of applying the recommendations was reported by doctors from the UK and the Czech Republic. The barriers cited by doctors refer to funding and administrative burden of anti-TNF. Thus, according to our results, there are inequalities in access to treatment and healthcare in Europe. The unfortunate situation for patients with AS in UK is not surprising since previous studies also have indicated that access to anti-TNF drugs for patients with rheumatoid arthritis is more limited in the UK than in, for example, the Netherlands and the Scandinavian countries.13

to treatment and healthcare in Europe. The unfortunate situation for patients with AS in UK is not surprising since previous studies also have indicated that access to anti-TNF drugs for patients with rheumatoid arthritis is more limited in the UK than in, for example, the Netherlands and the Scandinavian countries.13 In conclusion, this project supported the dissemination of evidence-based recommendations for AS. However, the project also disclosed intercountry differences for conceptual agreement with the recommendations and implementation of the recommendations in clinical practice. The results show that inequalities exist in the provision of healthcare for patients with AS in Europe, even between countries who are members of the European Union. We acknowledge all the ASAS members, and all rheumatologists for their participation, and Wyeth Europa for providing financial support for meetings to plan and conduct the project and for logistical support when the questionnaires were mailed to rheumatologists. Thanks are due to Andrew Bayode, commercial director, Wyeth, USA for his support for the project. Wyeth played no role in designing the study, nor in its analysis or decision to publish results. Funding: This work was supported by an unrestricted educational grant from Wyeth Europa.

We acknowledge all the ASAS members, and all rheumatologists for their participation, and Wyeth Europa for providing financial support for meetings to plan and conduct the project and for logistical support when the questionnaires were mailed to rheumatologists. Thanks are due to Andrew Bayode, commercial director, Wyeth, USA for his support for the project. Wyeth played no role in designing the study, nor in its analysis or decision to publish results. Funding: This work was supported by an unrestricted educational grant from Wyeth Europa. Competing interests: MD has received fees for his participation at Wyeth Advisory Committees, his centre has received grants from Wyeth and several other pharmaceutical companies in order to conduct clinical trials. CP, Wyeth Medical Director. TKK has received research grants and honorarium from Wyeth and several other pharmaceutical companies.

Proteoglycans are composed of glycosaminoglycan (GAG) chains, such as heparan sulphate, chondroitin sulphate, keratan sulphate or dermatan sulphate, covalently attached to a core protein. Two major classes of proteoglycans contain heparan sulphate chains: syndecans, which have a transmembrane domain in their core proteins, and glypicans, which are attached to the cell membrane by glycosylphosphatidylinositol (GPI)-anchors.1 2 In the basement membrane perlecan is the major component that bears heparan sulphate.

or classes of proteoglycans contain heparan sulphate chains: syndecans, which have a transmembrane domain in their core proteins, and glypicans, which are attached to the cell membrane by glycosylphosphatidylinositol (GPI)-anchors.1 2 In the basement membrane perlecan is the major component that bears heparan sulphate. To date, four syndecans and six glypicans have been identified. Syndecans are the major source of cell surface heparan sulphate. They are expressed in a cell-, tissue- and development-specific manner.1 Syndecan-1 and -4 have been shown in endothelial cells,3 however, syndecan-1 is mainly expressed on epithelial cells with syndecan-4 expression on many cell types. Changes in their expression occur during embryogenesis, wound healing and carcinogenesis.4–6 Although syndecan-2 has been identified as an endothelial heparan sulphate proteoglycan (HSPG),7 expression within tissues has also been shown on fibroblasts, for example, in skin and periodontium,8 with expression also occurring on carcinoma cells.9 Syndecan-3 was first identified on neuronal cells and has been associated with the generation of cerebellar fibrillar plaques in Alzheimer disease.10 It is also an HSPG of the musculoskeletal system.11–14 Glypicans are widely expressed in embryonic and adult tissues such as ovary, intestine and central nervous system, and are involved in growth factor signalling.15 They play a role in tissue growth, regeneration and cancer.

fibrillar plaques in Alzheimer disease.10 It is also an HSPG of the musculoskeletal system.11–14 Glypicans are widely expressed in embryonic and adult tissues such as ovary, intestine and central nervous system, and are involved in growth factor signalling.15 They play a role in tissue growth, regeneration and cancer. Rheumatoid arthritis (RA) is characterised by chronic inflammation of the synovium of the joints, resulting in stiffness, pain and—as the disease progresses—erosion of the joint tissues and deformities.16 Psoriatic arthritis (PsA) resembles RA in being an inflammatory disease leading to joint destruction, but differs from RA in several ways including the distribution of affected joints, the presence of skin lesions and enthesopathy, and the absence of rheumatoid factor, characteristic rheumatoid erosions and periarticular osteopoenia on radiographs. The exact pathogenesis of RA and PsA is largely unknown, but it is clear that a number of factors may be involved either individually or in combination. During RA, characteristic histopathological changes occur; the synovial lining layer undergoes thickening and hypertrophy, and in the sublining leukocytes such as monocytes, T cells and B cells migrate into the tissue where they accumulate.17

at a number of factors may be involved either individually or in combination. During RA, characteristic histopathological changes occur; the synovial lining layer undergoes thickening and hypertrophy, and in the sublining leukocytes such as monocytes, T cells and B cells migrate into the tissue where they accumulate.17 There is increasing evidence that HSPGs are involved in inflammation.18 Using animal knockout models and isolated cells, syndecan-1 and syndecan-4 have been shown to be involved in regulating inflammatory responses,19 binding chemokines20 21 and forming chemokine gradients.22 23 The chemokine CXCL8 has been shown to bind to syndecan-2 in cultured human umbilical vein endothelial cells24 and we recently showed the induction of a CXCL8 binding site on syndecan-3 in the endothelial cells of the RA synovium.14 However, little is known about the expression of syndecans and glypicans by the various cell types of the chronically inflamed synovium, although other proteoglycans bearing GAGs such as dermatan sulphate have been identified in this tissue.25 HSPGs are of interest as they are co-receptors for cytokines (eg, fibroblast growth factor), presenters of chemokines and are involved in cell–matrix and cell–cell adhesion.26–30 Hence they are likely candidates involved in several pathomechanisms in chronically inflamed synovia, such as angiogenesis and the migration and retention of leukocytes. Therefore, this study aimed to compare the expression patterns of syndecan-1, -2, -3, and -4, and glypican-1, -3 and -4 in the RA, PsA and normal synovium.

they are likely candidates involved in several pathomechanisms in chronically inflamed synovia, such as angiogenesis and the migration and retention of leukocytes. Therefore, this study aimed to compare the expression patterns of syndecan-1, -2, -3, and -4, and glypican-1, -3 and -4 in the RA, PsA and normal synovium. MATERIALS AND METHODS Tissue samples All samples of synovia were obtained, with informed consent, from the suprapatellar pouch and medial gutter from knee joints. Clinical and demographic details are given in table 1. For early RA, synovia were taken by arthroscopic needle biopsy from patients diagnosed with RA according to the American College of Rheumatology (ACR) criteria and had a mean disease duration of <12 months. RA synovia were also taken at total knee replacement or synovectomies from patients who fulfilled the ACR criteria and had a disease duration >5 years. Synovia were obtained by arthroscopic biopsy of knees of patients with PsA. Osteoarthritic (OA) synovia were obtained at total knee replacement operations. Non-RA synovia, which were histologically normal, were taken arthroscopically from patients who had knee joint symptoms for suspected joint damage. Tissue was frozen in liquid nitrogen-cooled isopentane and stored in the vapour phase of liquid nitrogen. Normal human skin samples were kindly provided by Dr E Johnston and Dr C Mangham, Department of Pathology, University of Birmingham, UK.

pically from patients who had knee joint symptoms for suspected joint damage. Tissue was frozen in liquid nitrogen-cooled isopentane and stored in the vapour phase of liquid nitrogen. Normal human skin samples were kindly provided by Dr E Johnston and Dr C Mangham, Department of Pathology, University of Birmingham, UK. Table 1 Clinical and demographic details of patients Patient Gender Age Diagnosis Duration (years) ESR RF (titre) Histology score Medication 1 M 52 Normal (meniscal tear) 1.2 ND ND 0 None 2 F 53 Normal (articular cartilage degeneration) 0.25 ND ND 1 None 3 M 57 Normal (articular cartilage degeneration) 1 ND ND 0 None 4 F 57 Normal (meniscal tear) 1.5 ND ND 0 None 5 M 58 Normal (meniscal tear and cartilage degeneration) 0.5 ND ND 0 None 6 F 41 Normal (articular cartilage degeneration) 0.5 ND ND 0 None 7 M 38 Normal (articular cartilage damage) 1 ND ND 0 None 8 F 55 Normal (meniscal and articular cartilage damage) 2 ND Negative 0 None 9 F 29 Normal (articular cartilage degeneration) 11 <10 Negative 0 None 10 M 28 Normal (meniscal tear) 1 ND ND 0 None 11 M 44 Normal (meniscal tear) 15 ND ND 0 NSAID 12 M 55 Normal (articular cartilage damage) >1 ND ND 0 None 1 ND ND ERA 1.5 14 Positive (640) 3 NSAID 2 ND ND ERA 0.25 23 Positive (80) 1 NSAID 3 ND ND ERA 0.6 ND ND 1 ND 4 ND ND ERA 0.25 51 Positive (640) 2 NSAID 5 ND ND ERA 0.5 44 Positive (640) 2 None 6 ND ND ERA 0.3 34 Positive (170) 1 None 7 ND ND ERA 1 48 Positive (160) 2 None 8 ND ND ERA 0.1 12 Positive (640) 3 NSAID 1 F 48 RA 18 24 Positive (400) 3 MTX, NSAID 2 F 47 RA 15 18 Negative 1 St, NSAID, analgesic 3 F 62 RA 26 64 Positive (>1280) 2 Sulphasalazine, NSAID 4 F 45 RA 20 71 Positive (640) 3 Sulphasalazine, azathioprine, NSAID 5 M 73 RA 15 59 Positive(1280) 2 d-Penicillamine, NSAID, analgesic 6 F 43 RA 5 23 Positive (640) 2 MTX 7 F 52 RA 20 30 Positive (1280) 3 MTX, Hydroxychloroquine, NSAID, st, gold 8 F 67 RA 7 105 Positive (160) 3 Steroid, NSAID 9 F 63 RA 8 100 Positive (320) 3 MTX, st, analgesic 10 F 62 RA 38 100 Positive (640) 3 MTX, st 11 F 55 RA 21 68 Positive (1280) 2 NSAID 13 F 68 RA 28 70 Positive (800) 2 Analgesic 1 M 42 OA 2 ND Negative 1 None 2 F 61 OA >4 ND ND 2 Analgesic 3 F 72 OA >5 ND ND 1 NSAID, analgesic 4 F 33 OA 20 ND Negative 1 NSAID 5 M 78 OA 15 ND ND 1 NSAID 6 F 83 OA 13 ND ND 2 NSAID 1 M 22 PsA 8 11 Negative 1 Aulin 2 F 40 PsA <1 28 Negative 2 NSAID, st, tar pomeda, reductil 3 F 60 PsA 13 4 Negative 2 Betnovate, Lipitor 4 F 50 PsA <1 5 Negative 0 DMARD 5 M 35 PsA 2 11 Negative 2 NSAID,

>5 ND ND 1 NSAID, analgesic 4 F 33 OA 20 ND Negative 1 NSAID 5 M 78 OA 15 ND ND 1 NSAID 6 F 83 OA 13 ND ND 2 NSAID 1 M 22 PsA 8 11 Negative 1 Aulin 2 F 40 PsA <1 28 Negative 2 NSAID, st, tar pomeda, reductil 3 F 60 PsA 13 4 Negative 2 Betnovate, Lipitor 4 F 50 PsA <1 5 Negative 0 DMARD 5 M 35 PsA 2 11 Negative 2 NSAID, Losec, analgesic 6 F 55 PsA 28 10 Negative 3 St, MTX 7 F 27 PsA 20 24 Negative 3 NSAID, st DMARD, disease-modifying antirheumatic drug; ERA, early rheumatoid arthritis; MTX, methotrexate; ND, not determined; NSAID, non-steroidal anti-inflammatory drug; OA, osteoarthritis; PsA, psoriatic arthritis; RA longstanding rheumatoid arthritis; RF, rheumatoid factor; St, steroid. Antibodies The antibodies used were: anti-syndecan-1 (clone B-B4; Serotec, Oxford, UK); anti-syndecan-2 (mouse IgG1, clone 10H4), anti-syndecan-3 (mouse IgG1 clone 1C7), anti-syndecan-4 (mouse IgG1 clone 8G3), anti-glypican-1 (mouse IgG1 clone S1) (all prepared and supplied by Dr G David, University of Leuven, Belgium), anti-glypican-3 (Santa Cruz Biotechnologies, Santa Cruz, California, USA), anti-glypican-4 (rabbit polyclonal antibody kindly donated by Professor HD Haubeck, Germany), anti-human IgG (rabbit polyclonal, Dako code A 0423; Dako, Glostrup, Denmark), anti-CD20 (Dako clone L26), anti-CD68 (rabbit polyclonal; Santa Cruz Biotech, code sc-9139). The antibodies to the syndecans and glypicans have already been shown to be specific using biochemical approaches (eg, western blotting) and specific staining has been obtained for other human tissues.31–37

Glostrup, Denmark), anti-CD20 (Dako clone L26), anti-CD68 (rabbit polyclonal; Santa Cruz Biotech, code sc-9139). The antibodies to the syndecans and glypicans have already been shown to be specific using biochemical approaches (eg, western blotting) and specific staining has been obtained for other human tissues.31–37 Immunohistochemistry The method of Roskams et al38 was followed, using the anti-syndecan and anti-glypican-1 antibodies, with minor modifications; for glypican-3 the method of Patterson et al14 for polyclonal antibodies was used with minor modifications, for glypican-4 the polyclonal antibody was diluted in 10% human serum and detected with swine anti-rabbit horse radish peroxidase conjugate diluted in 10% human serum. Briefly, 10 μm thick serial cryostat sections of synovia and skin were dried for 1 h and stored at −80°C until required. Prior to immunohistochemical analysis, slides were left to equilibrate to room temperature for 30 min, fixed in acetone (4°C) for 10 min, air dried, and then sections were rehydrated in phosphate buffered saline (PBS) for 5 min. All primary antibodies were used at 5 μg/ml since initial experiments showed that this concentration gave the optimum specific staining over the range 2.5–10 μg/ml. Antibody binding was detected using DAB staining kit (Vector Labs, Burlingame, California, USA). Sections were counterstained with Mayer Haematoxylin and mounted. In control experiments isotype matched control mouse Ig or rabbit Ig were added instead of primary antibodies, and stained with the DAB kit as above.

μg/ml. Antibody binding was detected using DAB staining kit (Vector Labs, Burlingame, California, USA). Sections were counterstained with Mayer Haematoxylin and mounted. In control experiments isotype matched control mouse Ig or rabbit Ig were added instead of primary antibodies, and stained with the DAB kit as above. Double label immunofluorescence For co-localisation of syndecan-1 and IgG, sections were treated with anti-syndecan-1 and anti-IgG (both at 5 μg/ml) added together, washed in PBS and incubated with goat anti-mouse IgG1–fluorescein isothiocyanate (FITC) (1:500 Southern Biotech code 1070-02; Southern Biotech, Birmingham, Alabama, USA) and swine anti-rabbit Ig-biotinylated (1:500 Dako code E0353) containing 10% human serum. The detection of the biotinylated secondary antibody was by using streptavidin-Alexa 594 (1:1000 Molecular Probes code S-11227; Molecular Probes, Eugene, Oregon, USA). For co-localisation of syndecan-1 and CD20, sections were treated with anti-syndecan-1 (5 μg/ml) and anti-CD20 (2 μg/ml), washed in PBS and incubated with goat anti-mouse IgG1–FITC (as above) and goat anti-mouse IgG2b-Texas red (1:100 Southern Biotech code 1090-07) containing 10% human serum. For co-localisation of syndecan-3 and CD68, sections were incubated with anti-syndecan-3 (5 μg/ml) and anti-CD68 (4 μg/ml), washed in PBS and treated with goat anti-mouse IgG-Alexa 488 (1:500 Molecular Probes code A11001) and goat anti-rabbit-Texas red (1:100 Santa Cruz Biotechnology code sc-2780) containing 10% human serum. All sections were washed in PBS, rinsed with water, air dried and mounted. In control experiments, isotype-matched control mouse, goat or rabbit Ig were added instead of primary antibodies, and treated with relevant secondary antibodies as above.

(1:100 Santa Cruz Biotechnology code sc-2780) containing 10% human serum. All sections were washed in PBS, rinsed with water, air dried and mounted. In control experiments, isotype-matched control mouse, goat or rabbit Ig were added instead of primary antibodies, and treated with relevant secondary antibodies as above. Analysis The immunoreactivity of HSPG expression was scored blind by two individuals using the following scoring system that consisted of a combination of staining intensity and number of positive cells: no staining (–), weak (+), strong (++) and very strong (+++) (table 2). A histological scoring system was used to assess the degree of synovial inflammation:14 0 for normal intimal layer and no leukocyte infiltration; 1 for small foci/areas of leukocyte infiltration with two or more synovial lining cells; 2 for synovitis with moderate infiltrations and three or more synovial lining cells; 3 for severe synovitis with widespread infiltration and four or more synovial lining cells (table 1). RESULTS Cryosections of RA, PsA, OA and non-RA synovia, and normal skin were incubated with antibodies to each of syndecans-1 to -4, glypicans-1, –3, and -4 and perlecan.

Analysis The immunoreactivity of HSPG expression was scored blind by two individuals using the following scoring system that consisted of a combination of staining intensity and number of positive cells: no staining (–), weak (+), strong (++) and very strong (+++) (table 2). A histological scoring system was used to assess the degree of synovial inflammation:14 0 for normal intimal layer and no leukocyte infiltration; 1 for small foci/areas of leukocyte infiltration with two or more synovial lining cells; 2 for synovitis with moderate infiltrations and three or more synovial lining cells; 3 for severe synovitis with widespread infiltration and four or more synovial lining cells (table 1). RESULTS Cryosections of RA, PsA, OA and non-RA synovia, and normal skin were incubated with antibodies to each of syndecans-1 to -4, glypicans-1, –3, and -4 and perlecan. Using immunohistochemistry, intense staining for syndecan-1 was seen in the lymphocytic infiltrates of patients with longstanding RA and PsA (fig 1, table 2). Staining was also present in these infiltrates in early RA (fig 1A), but was weak in OA and absent in normal synovia. Using double label immunofluorescence staining, these syndecan-1 positive cells co-localised with IgG suggesting that they were plasma cells (fig 2A,B) and their histological appearance was that of plasma cells. Syndecan-1 and IgG immunoreactivity were negative in normal synovia. The syndecan-1 positive cells did not co-localise with CD20 indicating that they were not immature or mature B cells (fig 2C,D).

sed with IgG suggesting that they were plasma cells (fig 2A,B) and their histological appearance was that of plasma cells. Syndecan-1 and IgG immunoreactivity were negative in normal synovia. The syndecan-1 positive cells did not co-localise with CD20 indicating that they were not immature or mature B cells (fig 2C,D). Figure 1 Expression of syndecan-1 in synovia. Sections of human knee synovium were incubated with an antibody to syndecan-1 and processed using immunoperoxidase and diaminobenzidine (brown). Staining is present in the mononuclear infiltrates of patients with early rheumatoid arthritis (A), longstanding rheumatoid arthritis (B) and psoriatic arthritis (C). (D) is a negative control section from the same individual as (B) except the primary antibody was replaced by isotype-matched control IgG1. Original magnification A ×250, B and D ×60, C ×120. Figure 2 Co-localisation of syndecan-1 with IgG. Sections of synovia from longstanding rheumatoid arthritis were treated with antibodies to syndecan-1 and human IgG and observed by double label immunofluorescence microscopy. Syndecan-1 positive cells (A) co-localise with IgG (B). Lack of co-distribution of syndecan-1 (C) and CD20 (D). Original magnification ×100 (A-D).

ctions of synovia from longstanding rheumatoid arthritis were treated with antibodies to syndecan-1 and human IgG and observed by double label immunofluorescence microscopy. Syndecan-1 positive cells (A) co-localise with IgG (B). Lack of co-distribution of syndecan-1 (C) and CD20 (D). Original magnification ×100 (A-D). Table 2 Cellular distribution of heparan sulphate proteoglycans in normal (n = 12), early rheumatoid (n = 8), longstanding rheumatoid (n = 13),osteoarthritis (n = 6) and psoriatic (n = 7) synovium Tissue Syndecan-1 Syndecan-2 Syndecan-3 Syndecan-4 Glypican-1 Glypican-3 Glypican-4 Normal Lining layer − − + − − − + Sublining: Fibroblasts − − + − − − − Mononuclear infiltrates − − − − − − − Macrophages − − + − − − − Endothelium − + + − − − + Smooth muscle cells − + − − − − + Pericytes − + − − − − + Early rheumatoid ND Lining layer − + ++ + − ++ Sublining: Fibroblasts − − + − − − Mononuclear infiltrates ++ + − − − − Macrophages − − ++ − − − Endothelium − ++ +++ − − + Smooth muscle cells − ++ − − − + Pericytes − + − − − + Longstanding rheumatoid Lining layer − − ++ − − − ++ Sublining: Fibroblasts − − + − − − − Mononuclear infiltrates +++ + − − − − − Macrophages − − ++ − − − − Endothelium − ++ +++ − − − + Smooth muscle cells − ++ − − − − + Pericytes − ++ − − − − + Osteoarthritis Lining layer − − + − − − + Sublining: Fibroblasts − − + − − − − Mononuclear infiltrates + − − − − − − Macrophages − − ++ − − − − Endothelium − ++ ++ − − − + Smooth muscle cells − ++ − − − − + Pericytes − + − − − − + Psoriatic Lining layer − + ++ + − − ++ Sublining: Fibroblasts − − + − − − − Mononuclear infiltrates +++ + − − − − − Macrophages − − ++ − − − − Endothelium − ++ +++ − − − + Smooth muscle cells − ++ − − - − + Pericytes − ++ − − − − + Immunoreactivity of cells: − no staining; + weak staining; ++ strong staining; +++ very strong staining.

layer − + ++ + − − ++ Sublining: Fibroblasts − − + − − − − Mononuclear infiltrates +++ + − − − − − Macrophages − − ++ − − − − Endothelium − ++ +++ − − − + Smooth muscle cells − ++ − − - − + Pericytes − ++ − − − − + Immunoreactivity of cells: − no staining; + weak staining; ++ strong staining; +++ very strong staining. ND, not determined. Syndecan-2 showed a different distribution compared to syndecan-1. It occurred mainly within the walls of blood vessels (fig 3A) in all samples with strong staining in RA, OA and PsA and weak staining in normal synovia (table 2). The endothelial cells, pericytes and smooth muscle cells were positive. Some additional staining was present in the lining layer of early RA and PsA samples.

It occurred mainly within the walls of blood vessels (fig 3A) in all samples with strong staining in RA, OA and PsA and weak staining in normal synovia (table 2). The endothelial cells, pericytes and smooth muscle cells were positive. Some additional staining was present in the lining layer of early RA and PsA samples. Figure 3 Expression of syndecan-2 and syndecan-3 in synovia. A. Immunohistochemistry showing staining of syndecan-2 in the walls of blood vessels in the sublining and in the lining layer of rheumatoid arthrtitis (RA) synovium. B. Syndecan-3 staining within endothelial cells and the lining layer in early RA. C. Detail of syndecan-3 staining of the lining layer in longstanding RA. D,E. Double label immunofluorescence microscopy with syndecan-3 (D) and CD68 (E) antibodies showing syndecan-3 localising to sublining macrophages in RA. Original magnification ×100 (A), ×200 (B) ×300 (C-E). Syndecan-3 also localised to blood vessels where it showed a more selective distribution than syndecan-2; intense labelling occurred on endothelial cells but pericytes and smooth muscle cells were negative (fig 3B, table 2). Staining was stronger in the endothelial cells of chronically inflamed synovia from RA and PsA, compared to OA and normal. This is in agreement with our earlier study showing more intense staining of this HSPG in synovial endothelial cells of longstanding RA compared to normal.14 Cells within the sublining connective tissue were positive for syndecan-3. These cells were macrophages, as judged by their morphology, and this was confirmed by double label experiments that showed syndecan-3 co-localising with CD68 positive cells in RA and normal synovia (fig 3D,E). CD68 negative cells were also positive for syndecan-3 in the sublining. These cells were identified as being fibroblasts, based on their morphology, whereas lymphocytic infiltrates were negative. The lining layer stained strongly for syndecan-3 in RA and PsA and weakly in OA and normal (fig 3C, table 2).

ia (fig 3D,E). CD68 negative cells were also positive for syndecan-3 in the sublining. These cells were identified as being fibroblasts, based on their morphology, whereas lymphocytic infiltrates were negative. The lining layer stained strongly for syndecan-3 in RA and PsA and weakly in OA and normal (fig 3C, table 2). In synovia from all patients groups, glypican-4 staining occurred in the blood vessel wall, localising to the endothelium, pericytes and smooth muscle cells (fig 4, table 2). In addition, glypican-4 was demonstrated in the lining layer where strong staining was found in early RA, longstanding RA and PsA. The staining pattern of perlecan was localised to the lining layer, and in the sublining positivity was found in the walls of blood vessels (data not shown). This staining pattern was present in the synovia from all patient groups and is in agreement with a previous study of OA synovium.39

RA, longstanding RA and PsA. The staining pattern of perlecan was localised to the lining layer, and in the sublining positivity was found in the walls of blood vessels (data not shown). This staining pattern was present in the synovia from all patient groups and is in agreement with a previous study of OA synovium.39 Figure 4 Glypican-4 expression in synovia. A. Staining occurs in the lining layer and the blood vessels in early rheumatoid arthritis (RA) synovium. B. The presence of glypican-4 in the lining layer in longstanding RA. In normal synovium (C), staining occurs in the walls of blood vessels. Original magnifications A and C ×100, B ×50. Syndecan-4 immunoreactivity was negative in all synovia (fig 5A) apart from some weak staining in the lining layer of early RA and PsA (table 2). Synovial staining for glypican-1 and glypican-3 was negative (fig 5C,E). However, expression of syndecan-4, glypican-1 and glypican-3 was detected in the epidermis (fig 5B,D,F) and blood vessels of human skin, similar to that previously reported,8 providing a positive control for these antibodies.

RA and PsA (table 2). Synovial staining for glypican-1 and glypican-3 was negative (fig 5C,E). However, expression of syndecan-4, glypican-1 and glypican-3 was detected in the epidermis (fig 5B,D,F) and blood vessels of human skin, similar to that previously reported,8 providing a positive control for these antibodies. Figure 5 Immunohistochemistry for syndecan-4, glypican-1 and glypican-3 in synovia and skin. Lack of syndecan-4 (A), glypican-1 (C) and glypican-3 (E) staining in rheumatoid arthritis (RA) synovia. However, staining for syndecan-4 (B) and glypican-1 (D) does occur in keratinocytes throughout the epidermis and glypican-3 is present in the basal cell layer of the epidermis. Original magnifications: A and B ×150, C and F ×350, D and E ×60. Immunostaining was negative in all experiments in which the primary antibodies to HSPGs were replaced with the same concentrations of control mouse (fig 1D), goat or rabbit Ig. DISCUSSION The results of the present study show that the HSPGs expressed in chronically inflamed synovium show a differential expression pattern. Syndecan-1 was present in the mononuclear infiltrates of synovia from patients with RA and PsA where it occurred in plasma cells. This HSPG has been used as a marker of epithelial cells and plasma cells.40–44 In the present study it was a selective marker for the latter cell type, which is in agreement with other studies that used syndecan-1 to identify these cells in chronically inflamed synovia.43 44

s that may be related to the interaction of the HSPG with vascular endothelial growth factor.48 This is of particular interest since in the RA synovium angiogenesis is a significant pathological event responsible for its enlargement and invasive properties, and suggests an involvement of syndecan-2 in these mechanisms. Syndecan-3 was abundantly expressed in the endothelial cells of chronically inflamed synovia. This pattern resembled that of syndecan-2 but differed in being undetectable in the pericytes or smooth muscle cells of blood vessels. There has been a previous report of the expression of syndecan-3 by human endothelial cells. This has been shown in normal liver and intense reactivity occurs on endothelial cells in heptacellular carcinomas.38 49 In addition, syndecan-3 has recently been shown in the blood vessels of non-malignant ovarian tissues with intense reactivity occurring on blood vessels in ovarian cancer.50 In the RA synovium, out of the syndecans and glypicans tested, syndecan-3 was the most abundantly detected HSPG in endothelial cells. Endothelial syndecan-3 selectively binds the chemokine CXCL8, suggesting a role in leukocyte trafficking into the synovium, whereas syndecan-2 and glypican-4 do not appear to bind this chemokine despite the presence of these HSPGs in the synovial endothelium.14 A role for syndecan-3 in leukocyte extravasation is further suggested since heparan sulphate can act as an adhesion molecule involved in blood leukocyte-endothelial interactions.51 In the current study syndecan-3 was not purely a marker of endothelial cells since it also localised to CD68+ macrophages in the sublining. In addition, the lining layer was positive for syndecan-3 suggesting that macrophages may also be positive in this layer, since macrophages are known to be a major component of the lining layer.17 In this connection macrophages have been shown to express syndecan-3 in rat liver52 and human liver with chronic cholestatic disease.49 Syndecan-3 is also expressed by chondrocytes in normal and OA articular cartilage where it is a regulator of chondrocyte proliferation.12 53 Therefore, taken with our findings in the synovium, syndecan-3 appears to be an HSPG particularly associated with joint tissues.

nd human liver with chronic cholestatic disease.49 Syndecan-3 is also expressed by chondrocytes in normal and OA articular cartilage where it is a regulator of chondrocyte proliferation.12 53 Therefore, taken with our findings in the synovium, syndecan-3 appears to be an HSPG particularly associated with joint tissues. Glypican-4 showed a similar distribution to that of syndecan-2, localising to endothelial cells, pericytes and smooth muscle cells in the walls of blood vessels. In addition, this HSPG occurred in the lining layer. By contrast, glypican-1 and glypican-3 were not detectable indicating that there is specificity in glypican expression in the synovium. Using the same antibody as ours, glypican-4 has been shown to be expressed in human kidney and bone marrow stromal and haematopoietic cells;37 54 it is also expressed in development and binds fibroblast growth factor-2.55 56

ere not detectable indicating that there is specificity in glypican expression in the synovium. Using the same antibody as ours, glypican-4 has been shown to be expressed in human kidney and bone marrow stromal and haematopoietic cells;37 54 it is also expressed in development and binds fibroblast growth factor-2.55 56 Several differences in HSPG expression were noted between inflamed and normal synovia. Syndecan-1 was abundantly expressed in RA and PsA samples whereas it was absent in normal and weakly expressed in OA. The lack of syndecan-1 in normal synovia relates to the absence of plasma cells in these samples whereas in RA and PsA these cells were abundantly present. Normal synovia were taken from patients with suspected meniscal damage of the knee and were essentially non-inflamed. The lack of lymphocytes of the B lineage is in agreement with a recent study by Singh et al57 who could not demonstrate L26+ B lymphocytes in normal knee synovia whereas CD3+, CD4+ and CD8+ T lymphocytes could be detected. Other changes in our study included an increased staining of syndecan-3 in endothelial cells, sublining macrophages and lining layer cells of RA and PsA in comparison to normal. Syndecan-2 staining was more intense in the blood vessels of RA and PsA compared with normal, and glypican-4 labelling was more intense in the lining layer of RA and PsA compared with normal. These results suggest that there is an upregulation of selected HSPGs in the chronically inflamed synovium.

omparison to normal. Syndecan-2 staining was more intense in the blood vessels of RA and PsA compared with normal, and glypican-4 labelling was more intense in the lining layer of RA and PsA compared with normal. These results suggest that there is an upregulation of selected HSPGs in the chronically inflamed synovium. The major difference in HSPG expression between the synovial samples appeared related to the degree of histological inflammation (tables 1 and 2). Inflammation, as characterised by the thickening of the synovial lining layer and infiltration of the sublining by leukocytes, was absent in normal synovia and increased in OA, RA and PsA. Similarly the level of HSPG expression was lowest in normal synovia and increased in OA, followed by RA and PsA. There was no obvious difference between the different types of chronic inflammation in terms of their HSPG expression, since RA and PsA showed an identical pattern. Therefore the difference in the pathologies of these two diseases, such as the respective presence or absence of rheumatoid factor (table 1), did not relate the presence or absence of particular syndecans or glypicans.

chronic inflammation in terms of their HSPG expression, since RA and PsA showed an identical pattern. Therefore the difference in the pathologies of these two diseases, such as the respective presence or absence of rheumatoid factor (table 1), did not relate the presence or absence of particular syndecans or glypicans. HSPGs act as co-receptors for cytokines and are involved in presenting chemokines to chemokine receptors.1 30 58 In addition, HSPGs act as adhesion molecules, playing an important role in cell–matrix and cell–cell interactions.4 29 These functions are due to interactions with the heparan sulphate chains but may also include the core protein.29 42 The role of the core protein of HSPGs is becoming increasingly apparent, not only determining when and where the heparan sulphate chains are expressed but also playing a direct role in signalling.59 Cell–matrix interactions have been shown for syndecan-3 and syndecan-1 that interact with extracellular molecules such as collagen and fibronectin.4 The presence of syndecan-3 on macrophages and syndecan-1 on plasma cells suggests that these HSPGs may be involved in leukocyte–matrix interactions leading to the accumulation of these cells in the inflamed synovial tissue. Furthermore, the presence of syndecan-2 and syndecan-3 on synovial endothelial cells could be involved in angiogenesis, chemokine presentation and leukocyte extravasation. With respect to angiogenesis, it is interesting that growth factors that are thought to drive this process in the RA synovium require binding to HSPGs for their activity, underlining an important potential role for HSPGs in this mechanism.18 60 61

ved in angiogenesis, chemokine presentation and leukocyte extravasation. With respect to angiogenesis, it is interesting that growth factors that are thought to drive this process in the RA synovium require binding to HSPGs for their activity, underlining an important potential role for HSPGs in this mechanism.18 60 61 We wish to acknowledge Martine Gogarty for sections of early RA samples and Professor HD Haubeck, (Aachen Germany) for the donation of anti-glypican-4 antibody. The following surgeons and rheumatologists are thanked for their assistance: Dr R Butler, Dr J Dixey, Mr C McGeoch, Mr D Rees, Mr R Spencer-Jones, Mr R Wade, Mr S White and the Daycase unit (Robert Jones and Agnes Hunt Orthopaedic Hospital). AP designed the study, performed the immunohistochemistry and wrote the manuscript. AC carried out the double label immunofluorescence experiments. GD developed and supplied the antibodies. OF was involved in obtaining PsA synovia and data interpretation. BB was involved in obtaining early RA synovia. BA was involved in ideas and designing experiments. JM designed the study and wrote the manuscript. All authors have read and approved the final manuscript. Funding: Funding was received from the Arthritis Research Campaign (UK) and the Wellcome Trust (UK). Competing interests: None.

Septic arthritis (SA) due to bacterial infection is a serious and potentially life threatening disease that can lead to rapid destruction of the vulnerable articular hyaline cartilage and irreversible loss of joint function.1 2 The reported incidence of septic arthritis is 2–6 cases/100 000 inhabitants.3–7 Advancing age, rheumatoid arthritis (RA), osteoarthritis, immunosuppressive therapies and diabetes mellitus have been identified as the main risk factors for idiopathic septic arthritis. Iatrogenic infection resulting from joint surgeries, arthroscopies and needle insertions into joints is another source of septic arthritis. In recent years, the number of arthroscopies has greatly increased as well as the number of joint injections due to the emergence of intra-articular viscosupplementation therapies. The risk of procedure-related SA has been estimated at 0.5–2.0% for arthroscopies8–12 and 0.005–0.0002% for joint injections.13–15 In light of the growing number of individuals with one or more risk factors for SA, the incidence would be expected to be rising. However, very few of these studies have addressed this issue and most published studies are derived from selected populations. Therefore a retrospective, nationwide study was carried out to examine the incidence, cause and characteristics of septic arthritis in Iceland over a 13-year period.

d be expected to be rising. However, very few of these studies have addressed this issue and most published studies are derived from selected populations. Therefore a retrospective, nationwide study was carried out to examine the incidence, cause and characteristics of septic arthritis in Iceland over a 13-year period. MATERIALS AND METHODS Clinical setting and definition of cases Over a 13-year period from 1 January 1990 to 31 December 2002 a nationwide computerised and manual survey for culture positive joint fluids was performed in all microbiology laboratories in Iceland. In addition, medical records and archives at all hospitals that admit patients with SA were searched for the discharge diagnosis of bacterial arthritis (M009 according to the International Statistical Classification of Diseases and Related Health Problems (ICD)-9 and ICD-10 codes).

microbiology laboratories in Iceland. In addition, medical records and archives at all hospitals that admit patients with SA were searched for the discharge diagnosis of bacterial arthritis (M009 according to the International Statistical Classification of Diseases and Related Health Problems (ICD)-9 and ICD-10 codes). Data collection Hospital medical records were reviewed and clinical and laboratory data collected in a systematic manner. Iatrogenic SA was defined as occurring within 2 weeks after arthrocentesis or arthroscopy, or within 6 months after open joint surgery. The total number of arthroscopic and arthrocentesis procedures 1990–2002 were available from computerised records at the State Social Security Institute (SSI), which is the only agency in Iceland that reimburses for ambulatory medical procedures. The total import and sales of corticosteroids for intra-articular injections (methyl prednisolone (DepoMedrol; Pfizer NewYork, NewYork, USA), betamethasone (Diprospan; Schering-Plough, Kenilworth, New Jersey, USA) and triamcinolone (Lederspan; Meda AB, Solna, Sweden)) and for hyaluronic acid preparations (Hyalgan, Artzal Astra-Zeneca, Albertslund, Denmark) were similarily obtained from the SSI. Defined daily doses (DDD) for all drugs were obtained from the World Health Organization (WHO) Collaborating Centre for Drug Statistics Methodology (WHOCC). The study was approved by the Data Protection Authorities in Iceland and the National Bioethics Committee.

rtslund, Denmark) were similarily obtained from the SSI. Defined daily doses (DDD) for all drugs were obtained from the World Health Organization (WHO) Collaborating Centre for Drug Statistics Methodology (WHOCC). The study was approved by the Data Protection Authorities in Iceland and the National Bioethics Committee. Statistical methods The age-specific incidence of infectious arthritis during the period 1990–2002 (yearly and total) was obtained by dividing the number of cases by the mean population number for each gender in each 10-year age interval (the first 2 years and 2–9 years of life were computed separately). The age-standardised incidence was computed using the global population figure as reference. The same approach was used for obtaining the standardised values for ages 0–16 and >16 years using a truncated global population figure as reference. Regression analysis was performed for the age-standardised incidence for each gender. Linear regression was first used for finding the linear time trend and then quadratic regression was used to find whether the time trend was changing with time. The association of yearly use of intra-articular drugs (DDD/1000 inhabitants) and number of septic arthritis cases was analysed with linear regression with and without adjusting for common time trend.

ing the linear time trend and then quadratic regression was used to find whether the time trend was changing with time. The association of yearly use of intra-articular drugs (DDD/1000 inhabitants) and number of septic arthritis cases was analysed with linear regression with and without adjusting for common time trend. The same approach was used for analysing the association of number of procedures involving joints and the number of septic arthritis cases during the period 1990–2002. The level of significance used was p<0.05. All tests were two-tailed. The χ2 test with Yates correction for continuity was used for comparison between two groups using SigmaStat for Windows software, V.3.11 (SigmaStat, Systat Software GmbH, Erkath, Germany).

number of septic arthritis cases during the period 1990–2002. The level of significance used was p<0.05. All tests were two-tailed. The χ2 test with Yates correction for continuity was used for comparison between two groups using SigmaStat for Windows software, V.3.11 (SigmaStat, Systat Software GmbH, Erkath, Germany). RESULTS Incidence of septic arthritis From 1990 to 2002 a total of 253 cases of bacterial arthritis were diagnosed in Iceland, of which 159 were males and 94 females, giving an average incidence of 7.1 cases/100 000 inhabitants. Additionally, five cases of atypical mycobacterial joint infections were identified. These cases were not included in the overall analysis of SA. There were 69 children (0–16 years old) diagnosed with SA, 37 of those less than 2 years of age. Figure 1 shows age- and gender-specific incidence rates of SA for children 0–24 months old, ⩾2–10 years old and subsequent 10-year age groups. As shown in fig 1, women had lower incidence rates than men in all age groups (male/female ratio 1.7). For both sexes, the rate of SA rose sharply after the age of 50. Among children less than 2 years of age the age adjusted incidence was quite high at 27/100 000.

old, ⩾2–10 years old and subsequent 10-year age groups. As shown in fig 1, women had lower incidence rates than men in all age groups (male/female ratio 1.7). For both sexes, the rate of SA rose sharply after the age of 50. Among children less than 2 years of age the age adjusted incidence was quite high at 27/100 000. Figure 1 Age and gender specific incidence rates of septic arthritis for children 0–24 months, 2–9 years old and subsequent 10-year age groups. Change in the incidence of septic arthritis 1990–2002 The annual total number of cases of SA for children 0–16 years and for adults are shown in fig 2. The annual incidence of SA increased progressively from 4.2 per 100 000 in 1990 to 11.0 per 100 000 inhabitants in 2002. This change was primarily due to increased number of SA cases in adults and was statistically significant by linear trend, which showed a yearly increase of 0.61 adult cases per 100 000 (p<0.001). The mean annual incidence during the first 5 years (1990–94) increased from 4.2 cases per 100 000 adult inhabitants to 9.4 per 100 000 adults in 1998–2002 (95% CI 3.2–5.5 and 7.9–11.1, respectively; p<0.01).

istically significant by linear trend, which showed a yearly increase of 0.61 adult cases per 100 000 (p<0.001). The mean annual incidence during the first 5 years (1990–94) increased from 4.2 cases per 100 000 adult inhabitants to 9.4 per 100 000 adults in 1998–2002 (95% CI 3.2–5.5 and 7.9–11.1, respectively; p<0.01). Figure 2 The annual total number of cases of septic arthritis for children 0–16 years and adults. Iatrogenic septic arthritis Joint infections were iatrogenic in 1.4% (1/69 cases) of children and 41.8% (77/184 cases) of adults. Iatrogenic infections in adults were due to open joint surgery in 26 cases (14.1% of total), arthroscopy in 18 cases (9.8%) and arthrocentesis in 33 cases (17.9%) (table 1). The mean number of iatrogenic infections increased from 2.8 infections/year in 1990–1994 to 9.0 infections/year in 1998–2002 (p<0.01) and this was the main reason for the observed increase in SA during the study period. The overall percentage of iatrogenic infections was similar in young adults (age 17–65) and older adults (age ⩾65). Septic arthritis due to arthrocentesis was common in both age groups but post arthroscopy SA was primarily observed in the younger adults (15/101 vs 3/83; p = 0,02) whereas infections due to open surgery were more common in the older adults (18/83 vs 8/101; p = 0.01).

oung adults (age 17–65) and older adults (age ⩾65). Septic arthritis due to arthrocentesis was common in both age groups but post arthroscopy SA was primarily observed in the younger adults (15/101 vs 3/83; p = 0,02) whereas infections due to open surgery were more common in the older adults (18/83 vs 8/101; p = 0.01). Table 1 Iatrogenic septic arthritis in Iceland 1990–2002 Arthrocentesis Arthroscopy Open joint surgery Total No. of cases 33 18 26 77 Age (median) 63 47 70 60.1 Male/female ratio 2 17 1 2.1 Infected joint: Knee 64% (21) 89% (16) 62% (16) 69% Shoulder 12% (4) 11% (2) 16% (4) 13% Hip 3% (1) 0 16% (4) 6% Other 21% (7) 0 8% (2) 12% Synovial culture: Coagulase-positive Staphylococci 49% (16) 22% (4) 27% (7) 35% Coagulase-negative Staphylococci 21% (7) 28% (5) 20% (5) 22% Streptococci 9% (3) 11% (2) 27% (7) 16% Other 9% (3) 17% (3) 15% (4) 13% Negative 12% (4) 22% (4) 11% (3) 14% Figures in parentheses are absolute no. of cases.

1% (7) 0 8% (2) 12% Synovial culture: Coagulase-positive Staphylococci 49% (16) 22% (4) 27% (7) 35% Coagulase-negative Staphylococci 21% (7) 28% (5) 20% (5) 22% Streptococci 9% (3) 11% (2) 27% (7) 16% Other 9% (3) 17% (3) 15% (4) 13% Negative 12% (4) 22% (4) 11% (3) 14% Figures in parentheses are absolute no. of cases. To determine whether the rising incidence of SA was related to increased number of diagnostic or therapeutic joint procedures, information about the total number of arthroscopies and arthrocentesis were obtained from the SSI in Iceland (table 2). The annual number of documented arthrocentesis was only available for the practice of rheumatology and orthopaedics, but not for general practitioners. We were, however, able to retrieve accurate information about the annual sales of intra-articular glucocorticoid and hyaluronan preparations for 1990–2002. There was a correlation between the usage of intra-articular betamethasone (Diprospan®; p = 0.001), hyaluronic acid (p = 0.02), and the total usage of intra-articular drugs (p = 0.008) and the incidence of SA (table 2). The number of reimbursed arthrocentesis procedures performed by rheumatologists and orthopaedists did not change markedly over the study period, but the number of arthroscopies increased from an annual mean of 430 arthroscopies in 1990–1994 to 2303 arthroscopies in 1998–2002 (p<0.001).

nd the incidence of SA (table 2). The number of reimbursed arthrocentesis procedures performed by rheumatologists and orthopaedists did not change markedly over the study period, but the number of arthroscopies increased from an annual mean of 430 arthroscopies in 1990–1994 to 2303 arthroscopies in 1998–2002 (p<0.001). Table 2 The annual sales of intra-articular glucocorticoids and hyaluronans defined by DDD (defined daily dose) and the annual registered number of arthroscopies and arthrocentesis in Iceland Year Population No. cases Lederspan DepoMedrol Diprospan Hyaloronan Total Arthroscopy Arthrocentesis 1990 254 788 11 0.33 0.27 0.26 0.00 0.86 55 5829 1991 257 965 15 0.25 0.29 0.30 0.00 0.84 230 6253 1992 264 103 10 0.31 0.30 0.23 0.00 0.84 258 6808 1993 263 783 17 0.36 0.28 0.19 0.00 0.83 642 6612 1994 266 006 8 0.45 0.25 0.17 0.01 0.83 965 6404 1995 267 380 18 0.44 0.25 0.22 0.02 0.93 2245 6607 1996 268 927 16 0.46 0.24 0.20 0.07 0.97 3123 7558 1997 270 935 16 0.13 0.42 0.41 0.12 1.08 2847 8124 1998 273 764 26 0.39 0.25 0.40 0.17 1.21 1479 8297 1999 277 184 20 0.42 0.22 0.35 0.20 1.19 2157 8412 2000 281 154 27 0.40 0.22 0.28 0.22 1.12 2351 7551 2001 285 054 34 0.20 0.34 0.48 0.16 1.18 2795 7063 2002 287 559 36 0.00 0.33 0.61 0.12 1.06 2735 6891 Clinical and microbiological characteristics of septic arthritis The clinical characteristics of SA in children (0–2 years vs >2–16 years old) and adults (16–65 years vs >65 years old) are shown in table 3. No significant difference in joint distribution was observed between younger and older adults, although shoulder infections appeared to be more common in the elderly (13.8% vs 4.8%, p = 0.05). In adults, SA was polyarticular in 3.3% of cases.

16 years old) and adults (16–65 years vs >65 years old) are shown in table 3. No significant difference in joint distribution was observed between younger and older adults, although shoulder infections appeared to be more common in the elderly (13.8% vs 4.8%, p = 0.05). In adults, SA was polyarticular in 3.3% of cases. Table 3 Clinical and laboratory characteristics of children and adults with septic arthritis in Iceland 1990–2002 Children Adults <2 years 2–16 years Total 16–65 years >65 years Total No. of cases 31 38 69 101 83 184 Infected joint: Knee 35.5% 35% 35% 45% 53% 49% Hip 38.7% 32.4% 35.5% 12.6% 15% 13.7% Shoulder 6.5% 0% 2.9% 5.8% 13.8% 9.5% Ankle 9.7% 16.2% 13.2% 10.6% 5.7% 8.4% Hands and feet 0 7.9% 4.3% 11.9% 8.4% 10.3% Other 9.6% 8.5% 9.0% 14.1% 4.1% 9.0% Iatrogenic infections 0% 2.7% 1.5% 42% 40% 41.8% Admission data: Temperature (°C) 38.1 38.1 38.1 38.0 38.0 38.0 Peripheral blood: WBC (×109/litre) 13.6 10.5 11.9 11.6 9.5 10.7 ESR (mm/h) 52 28 35 54 80 68 CRP (mg/litre) 48 38 44 125 189 125 Culture positive 23% 24% 24% 28% 41% 35% Synovial fluid: WBC (×106/litre) 107 100 65 900 72 600 50 000 58 400 52 700 Gram stain positive 35% 35% 35% 46% 52% 49% Culture positive 45% 62% 54% 87% 81% 84% Normal admission values: Temperature <37.8°C 19% 26% 23% 41% 44% 42% WBC <10×109/litre 6% 43% 26% 36% 57% 46% ESR <20 mm/h 8% 26% 18% 18% 7% 12% CRP <10 mg/litre 9% 11% 7% 12% 18% 15% Median values for WBC, ESR, CRP and temperature are shown. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; WBC, white blood cell count.

Table 3 Clinical and laboratory characteristics of children and adults with septic arthritis in Iceland 1990–2002 Children Adults <2 years 2–16 years Total 16–65 years >65 years Total No. of cases 31 38 69 101 83 184 Infected joint: Knee 35.5% 35% 35% 45% 53% 49% Hip 38.7% 32.4% 35.5% 12.6% 15% 13.7% Shoulder 6.5% 0% 2.9% 5.8% 13.8% 9.5% Ankle 9.7% 16.2% 13.2% 10.6% 5.7% 8.4% Hands and feet 0 7.9% 4.3% 11.9% 8.4% 10.3% Other 9.6% 8.5% 9.0% 14.1% 4.1% 9.0% Iatrogenic infections 0% 2.7% 1.5% 42% 40% 41.8% Admission data: Temperature (°C) 38.1 38.1 38.1 38.0 38.0 38.0 Peripheral blood: WBC (×109/litre) 13.6 10.5 11.9 11.6 9.5 10.7 ESR (mm/h) 52 28 35 54 80 68 CRP (mg/litre) 48 38 44 125 189 125 Culture positive 23% 24% 24% 28% 41% 35% Synovial fluid: WBC (×106/litre) 107 100 65 900 72 600 50 000 58 400 52 700 Gram stain positive 35% 35% 35% 46% 52% 49% Culture positive 45% 62% 54% 87% 81% 84% Normal admission values: Temperature <37.8°C 19% 26% 23% 41% 44% 42% WBC <10×109/litre 6% 43% 26% 36% 57% 46% ESR <20 mm/h 8% 26% 18% 18% 7% 12% CRP <10 mg/litre 9% 11% 7% 12% 18% 15% Median values for WBC, ESR, CRP and temperature are shown. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; WBC, white blood cell count. Compared to adults the relative occurrence of SA of the hip was significantly higher in children (35.5% vs 13.7% respectively, p<0.001).

Table 3 Clinical and laboratory characteristics of children and adults with septic arthritis in Iceland 1990–2002 Children Adults <2 years 2–16 years Total 16–65 years >65 years Total No. of cases 31 38 69 101 83 184 Infected joint: Knee 35.5% 35% 35% 45% 53% 49% Hip 38.7% 32.4% 35.5% 12.6% 15% 13.7% Shoulder 6.5% 0% 2.9% 5.8% 13.8% 9.5% Ankle 9.7% 16.2% 13.2% 10.6% 5.7% 8.4% Hands and feet 0 7.9% 4.3% 11.9% 8.4% 10.3% Other 9.6% 8.5% 9.0% 14.1% 4.1% 9.0% Iatrogenic infections 0% 2.7% 1.5% 42% 40% 41.8% Admission data: Temperature (°C) 38.1 38.1 38.1 38.0 38.0 38.0 Peripheral blood: WBC (×109/litre) 13.6 10.5 11.9 11.6 9.5 10.7 ESR (mm/h) 52 28 35 54 80 68 CRP (mg/litre) 48 38 44 125 189 125 Culture positive 23% 24% 24% 28% 41% 35% Synovial fluid: WBC (×106/litre) 107 100 65 900 72 600 50 000 58 400 52 700 Gram stain positive 35% 35% 35% 46% 52% 49% Culture positive 45% 62% 54% 87% 81% 84% Normal admission values: Temperature <37.8°C 19% 26% 23% 41% 44% 42% WBC <10×109/litre 6% 43% 26% 36% 57% 46% ESR <20 mm/h 8% 26% 18% 18% 7% 12% CRP <10 mg/litre 9% 11% 7% 12% 18% 15% Median values for WBC, ESR, CRP and temperature are shown. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; WBC, white blood cell count. Compared to adults the relative occurrence of SA of the hip was significantly higher in children (35.5% vs 13.7% respectively, p<0.001). The median temperature on admission was 38–38.1°C in all age groups. Normal temperatures (<37.8°C) were quite frequent, seen in about one fourth of children and almost half of the adults (table 3). Similarily, peripheral white blood cell (WBC) count was normal (<10 000/μl) on admission in 26% of children and 46% of adults. Normal values for temperature and WBC count on admission were more common in adults (p = 0.066 and p = 0.05 respectively). Normal erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) values were infrequent and only 2.3% of children and 4.4% of adults had normal values for ESR and CRP on admission.

adults. Normal values for temperature and WBC count on admission were more common in adults (p = 0.066 and p = 0.05 respectively). Normal erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) values were infrequent and only 2.3% of children and 4.4% of adults had normal values for ESR and CRP on admission. Synovial fluid was obtained in 97.4% of cases (98.6% of children, 97.3% of adults). Results from synovial fluid analysis are shown in table 3. Although the leukocyte count was markedly elevated (median = 58 100/mm3) a relatively low synovial fluid leukocyte count (<20 000/mm3) was observed in 7% of children and 15% of adults. Synovial fluid culture was positive in 84% of adults but only in 54% of children. Blood cultures were obtained in 96% (66/69) of children and in 74% (137/184) of adults. They were positive in one quarter to one third of individuals tested, arguably more often in the elderly (41%, p = 0.056). Positive Gram stain and/or blood culture increased the positively identified septic arthritis cases considerably, from 54% to 61% in children and from 84% to 88% in adults. The yield of synovial fluid culture was significantly lower in adult patients who had received antibiotics prior to diagnostic joint aspiration (10/27 negative, 37%) compared to patients without antibiotic exposure (15/119 negative, 12.6%; p = 0.002).

iderably, from 54% to 61% in children and from 84% to 88% in adults. The yield of synovial fluid culture was significantly lower in adult patients who had received antibiotics prior to diagnostic joint aspiration (10/27 negative, 37%) compared to patients without antibiotic exposure (15/119 negative, 12.6%; p = 0.002). The contributing risk factors for SA among adults were history of a recent trauma to the involved joint in 24% of cases, osteoarthritis in 19%, diabetes mellitus (DM) in 7% all having type II DM and being treated with per oral hypoglycaemic drugs, rheumatoid arthritis in 4% all being treated with methotrexate, neoplasia in 2% and one patient was HIV positive. Additionally, three patients were receiving glucocorticoids. In 16 (8.7%) of the adult cases SA was due to infected prosthetic joints, 12 knees, 3 hip joints and 1 shoulder joint. The type and frequency of bacterial organisms is shown in table 4.

The contributing risk factors for SA among adults were history of a recent trauma to the involved joint in 24% of cases, osteoarthritis in 19%, diabetes mellitus (DM) in 7% all having type II DM and being treated with per oral hypoglycaemic drugs, rheumatoid arthritis in 4% all being treated with methotrexate, neoplasia in 2% and one patient was HIV positive. Additionally, three patients were receiving glucocorticoids. In 16 (8.7%) of the adult cases SA was due to infected prosthetic joints, 12 knees, 3 hip joints and 1 shoulder joint. The type and frequency of bacterial organisms is shown in table 4. Table 4 Results of synovial fluid culture in septic arthritis in Iceland Bacteria Children Adults <2 years, n = 31 2–16 years, n = 38 Total, n = 69 16–65 years, n = 101 >65 years, n = 83 Total, n = 184 Staphylococcus aureus 6.5% (2) 34% (13) 21.7% (15) 45% (46) 38.6% (32) 42.6% (78) Streptococci 16% (5) 16% (6) 15.9% (11) 16% (16) 20.5% (17) 18.0% (33) Coagulase-negative Staphylococci 6.5% (2) 0 2.9% (2) 14% (14) 8.4% (7) 11.5% (21) Gram negative rod 3.2% (1) 2.6% (1) 2.9% (2) 4% (4) 7.2% (6) 5.5% (10) Kingella kingae 6.5% (2) 5.2% (2) 5.8% (4) 0 0 0 Other 3.2% (1) 5.2% (2) 4.4% (3) 6% (6) 3.6% (3) 4.9% (9) Culture negative 55% (17) 37% (14) 44.9% (31) 13% (13) 18.1% (15) 15.3% (28) Culture not obtained 3.2% (1) 0 1.5% (1) 2% (2) 3.6% (3) 2.7% (5) Figures in parentheses are absolute no. of cares.

) 7.2% (6) 5.5% (10) Kingella kingae 6.5% (2) 5.2% (2) 5.8% (4) 0 0 0 Other 3.2% (1) 5.2% (2) 4.4% (3) 6% (6) 3.6% (3) 4.9% (9) Culture negative 55% (17) 37% (14) 44.9% (31) 13% (13) 18.1% (15) 15.3% (28) Culture not obtained 3.2% (1) 0 1.5% (1) 2% (2) 3.6% (3) 2.7% (5) Figures in parentheses are absolute no. of cares. A total of 33 of the adult cases were not identified by positive joint fluid culture, and of those, 4 had positive Gram stain of joint fluid and 4 had positive blood cultures. In all, 32 (46%) children were not identified by positive joint fluid culture, but one of those had positive Gram stain of joint fluid and three had positive blood culture. The clinical and laboratory characteristics of children and adults with SA diagnosed without positive joint fluid culture was similar to the culture positive SA. Specifically, median age, sex and joint involvement were comparable, as well as ESR, CRP and synovial fluid leukocyte count, duration of antibiotic therapy and hospital stay (data not shown). Of the 254 patients, 5 died in the hospital, (mean age, 70 years); 3 following infection with Staphylococcus aureus, 1 after pneumococcal infection, and 1 following streptococcal infection. The mortality rate was 1.8% when calculated from all cases of septic arthritis but 2,7% if limited to the adult population.

A total of 33 of the adult cases were not identified by positive joint fluid culture, and of those, 4 had positive Gram stain of joint fluid and 4 had positive blood cultures. In all, 32 (46%) children were not identified by positive joint fluid culture, but one of those had positive Gram stain of joint fluid and three had positive blood culture. The clinical and laboratory characteristics of children and adults with SA diagnosed without positive joint fluid culture was similar to the culture positive SA. Specifically, median age, sex and joint involvement were comparable, as well as ESR, CRP and synovial fluid leukocyte count, duration of antibiotic therapy and hospital stay (data not shown). Of the 254 patients, 5 died in the hospital, (mean age, 70 years); 3 following infection with Staphylococcus aureus, 1 after pneumococcal infection, and 1 following streptococcal infection. The mortality rate was 1.8% when calculated from all cases of septic arthritis but 2,7% if limited to the adult population. From 1990–2002 we identified five cases of arthritis due to mycobacteria other than tuberculosis; two Mycobacterium kansasii, two Mycobacterium avium-intracellulare and one Mycobacterium marinum infection. These infections affected three elbows and two wrists joints in four men and one woman, 34–49 years old. One of these patients was HIV positive but the other four had no identifiable immunodeficiency. These mycobacterial infections were not included in the overall analysis of the data.

Mycobacterium marinum infection. These infections affected three elbows and two wrists joints in four men and one woman, 34–49 years old. One of these patients was HIV positive but the other four had no identifiable immunodeficiency. These mycobacterial infections were not included in the overall analysis of the data. DISCUSSION In this paper we report a markedly increased rate of joint infections in adults in Iceland from 1990 to 2002. During the study period there was a mean annual increase of 0.61 cases/100 000 adult inhabitants. The incidence of 4.2 cases/100 000 adults in 1990–1994 is similar to the reported number in other studies3–6 16 but the incidence of 9.4/100 000 adults during 1998–2002 is considerably higher and statistically different (p<0.01). This change is primarily due to increase in iatrogenic infections following joint arthroscopies and arthrocentesis. It coincides with marked increase in the registered number of joint arthroscopies in Iceland and increased use of intra-articular steroids and joint-viscous supplements in Iceland during these years. However, this increase in SA following joint injections was not due to a higher number of arthrocentesis procedures performed by rheumatologists and orthopaedists. However, the increased use of intra-articular steroids and joint-viscous supplements suggests that the number of arthrocentesis procedures did indeed increase during this observational period. That scenario could be explained by higher numbers of arthrocentesis procedures being performed by other doctors. This is however purely speculative, since computerised documentation of arthrocentesis in primary care is mostly unavailable. Septic arthritis due to open joint surgeries did not increase over this time period.

scenario could be explained by higher numbers of arthrocentesis procedures being performed by other doctors. This is however purely speculative, since computerised documentation of arthrocentesis in primary care is mostly unavailable. Septic arthritis due to open joint surgeries did not increase over this time period. The frequency of post-arthroscopic SA has been reported 0.1–0.5% at specific medical centres9–12 but to our knowledge this frequency has not been previously estimated based on a nationwide analysis. The estimated frequency in Iceland of 0.14% is similar to these previous reports.

scenario could be explained by higher numbers of arthrocentesis procedures being performed by other doctors. This is however purely speculative, since computerised documentation of arthrocentesis in primary care is mostly unavailable. Septic arthritis due to open joint surgeries did not increase over this time period. The frequency of post-arthroscopic SA has been reported 0.1–0.5% at specific medical centres9–12 but to our knowledge this frequency has not been previously estimated based on a nationwide analysis. The estimated frequency in Iceland of 0.14% is similar to these previous reports. There is very limited data available in the medical literature about the actual risk of SA following arthrocentesis. In early reports this risk was estimated at 0.005 to 0.0002% per arthrocentesic procedure.13–15 These percentages are lower than can be estimated from our study. According to the SSI registry in Iceland the total number of documented ambulatory arthrocentesis in Iceland performed by rheumatologists and orthopaedists is 6900 per year (table 2). Additionally, an estimated maximum of 1000 arthrocentesis procedures are performed annually by general practitioners (L. Ólafsson, the Center of Health Care in Reykjavik, Iceland personal communication). Accordingly, the estimated risk of septic arthritis following arthrocentesis in Iceland is 3 infections/7900 procedures, or 0.037% per injection. Compared to other commonly performed procedures, an incidence of 0.037% per injection is not high. However, given the significant morbidity and 10–15% mortality previously reported in SA3–6 16 it is of utmost importance to avoid unnecessary infectious complications and to perform arthrocentesis according to the best standard of practice.

ommonly performed procedures, an incidence of 0.037% per injection is not high. However, given the significant morbidity and 10–15% mortality previously reported in SA3–6 16 it is of utmost importance to avoid unnecessary infectious complications and to perform arthrocentesis according to the best standard of practice. In our study 17,9% of SA in adults occurred post-arthrocentesis, which is alarmingly high compared to the frequency of 1.9–3% in previous studies.4 6 7 It has been suggested that intra-articular steroid17 or hyaluronan18 injection may increase the risk of joint infection. Thus, the generally increased use of steroids and hyaluronans in Iceland (table 2) may explain the increased incidence of SA in Iceland.

compared to the frequency of 1.9–3% in previous studies.4 6 7 It has been suggested that intra-articular steroid17 or hyaluronan18 injection may increase the risk of joint infection. Thus, the generally increased use of steroids and hyaluronans in Iceland (table 2) may explain the increased incidence of SA in Iceland. Presumably organisms frequently enter the joint during arthrocentesis. Skin fragments introduced into the joint during arthrocentesis contained bacterial genes (by polymerase chain reaction) one third of the time.19 and despite standard sterilisation of the skin surface, organisms could be cultured from needle tips in 14–28% of events.20 The best standard of practice for arthrocentesis has not been thoroughly studied and method of sterilisation varies considerably among doctors;21 alcohol swabs, pivodone iodine and chlorhexidine have been used in different reports, all apparently with satisfactory results, although chlorhexidine arguably has the edge over the other two methods. Chlorhexidine resulted in better sterilisation of needle tips compared to alcohol swabs20 and chlorhexidine use yielded significantly lower skin pathogen contamination of blood cultures compared with pivodone iodine.22 It may also be of practical importance in the clinical settings that full bacteriocidal effects from pivodone iodine takes over 1 min to develop.22 As far as we know, all three types of antiseptic techniques are in use in Iceland but no survey has been conducted to assess their prevalence among doctors. We were not able to correlate the type of antiseptic technique used to the rate of SA in this retrospective study. Irrespective of this, a panel of experts from the American College of Rheumatology emphasised that intra-articular corticosteroid injections are “safe and effective when administered by an experienced doctor”.23 This may indeed be the most important factor in avoiding iatrogenic SA following arthrocentesis.

tive study. Irrespective of this, a panel of experts from the American College of Rheumatology emphasised that intra-articular corticosteroid injections are “safe and effective when administered by an experienced doctor”.23 This may indeed be the most important factor in avoiding iatrogenic SA following arthrocentesis. The clinical characteristics of SA have been previously reported in children24–28 adults,3–6 in the elderly;29–32 and numerous reviews have been published on septic arthritis.33–36 Most previous studies in adults have been derived from selected regions or populations, although one report represented a whole communal region.5 The current report is the first study reporting a nationwide survey. We screened all microbiology laboratories in Iceland for positive joint fluid cultures, and all hospitals in Iceland for ICD-9 and-10 discharge diagnoses of bacterial arthritis. Our study has numerous clinical and laboratory findings similar to previous reports. Thus, the sex ratio of 1.7 (male/female), risk factors (including advancing age, RA, osteoarthritis (OA), DM, iatrogenic illness), joint involvement and pathogens cultured are quite comparable to these reports. In our study, the frequency of polyarticular involvement among adults was quite low (3.3%) compared to 8.4–19.5% in other studies7 8 and mortality among adults was only 2.7%, considerably lower than reported in most studies,3–6 19although similar to some.29 Our study agrees with previous findings that markers of systemic inflammation (fever, leukocytosis, elevated ESR or CRP) can be normal on presentation in SA. In fact normal temperature, WBC count, ESR and CRP were observed in 42%, 46%, 12% and 15%, respectively, in adults (table 3). Furthermore, although SA is usually associated with high synovial leukocyte count we observed counts less than 20 000 in 15% of adult cases. Our study emphasises the importance of obtaining blood cultures as well as synovial fluid cultures, and highlights the potential for false negative results if the patient has received oral antibiotics prior to obtaining cultures. Thus, the diagnosis of SA in adults according to our study should primarily be derived from thorough medical history and physical examination, supported by joint fluid WBC analysis and confirmed by synovial/blood cultures in 88% cases. Antibiotic exposure prior to synovial culture may decrease the culture yield to 63%.

Thus, the diagnosis of SA in adults according to our study should primarily be derived from thorough medical history and physical examination, supported by joint fluid WBC analysis and confirmed by synovial/blood cultures in 88% cases. Antibiotic exposure prior to synovial culture may decrease the culture yield to 63%. Surprisingly, 39% of children with clinical picture of septic arthritis had negative synovial fluid and blood cultures. The clinical and laboratory characteristics were similar in culture positive and negative children. These results are identical to numerous other reports24–28 and to date no reasonable explanation has been put forward. In summary, the incidence of septic arthritis has markedly increased in recent years due to increased number of arthroscopies and therapeutic joint injections. Although the frequency of SA per procedure has not changed, these results emphasise the importance of sterile technique and firm indications for diagnostic and therapeutic joint procedures. We thank Helgi Sigvaldason for statistical assistance and to Dr Magnús Gottfredsson for reviewing the manuscript. Funding: This study was suppported by a grant from the Wyeth Rheumatology Foundation of Iceland. Competing interests: None declared. Ethics approval: The study was approved by the Data Protection Authorities in Iceland and the National Bioethics Committee.

Psoriatic arthritis (PsA) is commonly defined as “an inflammatory arthritis (IA) associated with psoriasis, which is usually negative for rheumatoid factor (RF)”.1 A strong genetic component to susceptibility is suggested by the sibling recurrence risk ratio (λs), which is estimated to be 27.2 3 Part of the genetic predisposition is likely to be explained by genes within the major histocompatibility complex (MHC) region. For example, human leucocyte antigen (HLA) associations with psoriasis and IA in the form of rheumatoid arthritis (RA) are well characterised and, in each case, the MHC genes involved are recognised as the major disease susceptibility locus. Psoriasis has two distinct ages of onset: type I, early onset disease occurring at ⩽40 years of age and type II, late onset occurring at >40 years of age. Carriage of the HLA class I allele, HLA-Cw*06 is associated with type I but not type II psoriasis.4–13 In contrast, RA is associated with carriage of the shared epitope (SE) of the HLA class II DRB1 gene (a group of DRB1 alleles sharing a conserved amino acid motif in the third hypervariable region of the DRβ chain).14 HLA-Cw*06 is not found on haplotypes encoding the SE.

iated with type I but not type II psoriasis.4–13 In contrast, RA is associated with carriage of the shared epitope (SE) of the HLA class II DRB1 gene (a group of DRB1 alleles sharing a conserved amino acid motif in the third hypervariable region of the DRβ chain).14 HLA-Cw*06 is not found on haplotypes encoding the SE. For PsA disease susceptibility, separating the HLA-related genetic contribution from the contribution to psoriasis and IA alone is a challenge. There is evidence to suggest that carriage of the HLA-Cw*06 allele is associated with patients with PsA with type I psoriasis.12 15–19 Studies of the HLA class II DRB1 gene have reported that the HLA-DRB1*07 phenotype is associated with peripheral arthritis in PsA while HLA-DRB1*04 is associated with a subgroup of patients with PsA with polyarthritis mimicking that of RA.20–26 In one study of the HLA-DRB1 locus, HLA-DRB1*0402 (which is not a SE allele) was found to occur more frequently in patients with PsA compared with RA or healthy controls, whereas HLA-DRB*0401 (which is a SE allele) occurred less frequently.27 Conflicting results have been reported with regard to the HLA-DRB1*02 (HLA-DRB*15 or HLA-DRB*16) and PsA susceptibility with a decrease found in a UK study, but not in a cohort from Toronto.28 29 More recently, a UK study showed no overall difference in the frequency of the SE between PsA cases and controls.28

ently.27 Conflicting results have been reported with regard to the HLA-DRB1*02 (HLA-DRB*15 or HLA-DRB*16) and PsA susceptibility with a decrease found in a UK study, but not in a cohort from Toronto.28 29 More recently, a UK study showed no overall difference in the frequency of the SE between PsA cases and controls.28 Of the associations reported, however, it is still unclear whether the primary association is with PsA itself or secondary to one of the two constituent components, ie, IA or psoriasis. In previous studies, all appropriate comparison groups have not been analysed and some studies have also lacked power to address this question. The aim of this study was to compare the association between HLA-Cw*06 and HLA-DRB1, including the SE, with PsA susceptibility by comparing these phenotypes in patients with PsA, early undifferentiated inflammatory arthritis (UIA) alone, psoriasis alone and healthy controls.

also lacked power to address this question. The aim of this study was to compare the association between HLA-Cw*06 and HLA-DRB1, including the SE, with PsA susceptibility by comparing these phenotypes in patients with PsA, early undifferentiated inflammatory arthritis (UIA) alone, psoriasis alone and healthy controls. METHODS Overview For this study, we did not investigate HLA-Cw*06 carriage in UIA patients or HLA-DRB1 allele carriage in patients with psoriasis as previous studies have excluded a role for these genes in the respective conditions.12 30–33 A case–control association study was performed to investigate the role of HLA-Cw*06 in determining susceptibility to PsA by comparing allele and phenotype frequencies between patients with PsA, psoriasis alone and healthy controls. For the HLA-DRB1 gene variants, HLA-DRB1*04, HLA-DRB1*07, HLA-DRB1*02 (HLA-DRB*15 or HLA-DRB*16) and SE phenotype frequencies were compared in patients with PsA, UIA alone and healthy controls. Stratification analyses were performed by subdividing patients with PsA into type I and type II psoriasis according to the age at onset of their skin disease (type I = age of onset ⩽40 years and type II = >40 years of age). Analyses were also repeated in subsets of patients with PsA stratified by the presence of RF.

s. Stratification analyses were performed by subdividing patients with PsA into type I and type II psoriasis according to the age at onset of their skin disease (type I = age of onset ⩽40 years and type II = >40 years of age). Analyses were also repeated in subsets of patients with PsA stratified by the presence of RF. Subjects Patients with psoriatic arthritis The recruitment of patients with PsA for this study has been described previously.34 In brief, patients with PsA (n = 480) under active follow-up by hospital rheumatologists were recruited from throughout the UK with the majority of them coming from north-west England. All patients satisfied the inclusion criteria of having both clinically documented inflammatory synovitis and psoriasis regardless of their RF status. A trained research nurse interviewed the patients and completed a standardised clinical history and examination protocol. Detailed demographic and clinical information were obtained and whole blood was taken for the measurement of RF status, DNA extraction and subsequent genetic analysis. Patients with psoriasis As described previously,35 patients with type I psoriasis (age of onset ⩽40 years) (n = 611) were recruited via the Dermatology Centre at Hope Hospital in Manchester. Some of the patients with psoriasis may have an IA, but this was not documented for the majority. A subset (n = 229) underwent an examination to exclude inflammatory joint involvement.

ents with type I psoriasis (age of onset ⩽40 years) (n = 611) were recruited via the Dermatology Centre at Hope Hospital in Manchester. Some of the patients with psoriasis may have an IA, but this was not documented for the majority. A subset (n = 229) underwent an examination to exclude inflammatory joint involvement. Undifferentiated inflammatory arthritis Patients with early UIA were recruited from the Norfolk Arthritis Register (NOAR) as described previously.36 This is a primary-care-based inception cohort of subjects with primary UIA. Patients were aged ⩾16 years with two or more inflamed peripheral joints lasting at least 4 weeks. For the purpose of this study, all patients with HLA-DRB1 data available were included (n = 1621). Population controls Control subjects without a history of IA or psoriasis were recruited from blood donors and general practice registers (n = 537). All patients and controls were white Caucasians of British descent. They were recruited with ethical committee approval (MREC 99/8/84 (PsA samples); LREC 00089 (psoriasis samples), LREC 2003-075 (NOAR samples)) and provided written informed consent.

Population controls Control subjects without a history of IA or psoriasis were recruited from blood donors and general practice registers (n = 537). All patients and controls were white Caucasians of British descent. They were recruited with ethical committee approval (MREC 99/8/84 (PsA samples); LREC 00089 (psoriasis samples), LREC 2003-075 (NOAR samples)) and provided written informed consent. HLA typing Both broad HLA genotyping (HLA-Cw and HLA-DR) and subtyping to define SE alleles were performed using 50 ng of genomic DNA amplified with the Dynal RELI SSO HLA-Cw typing and HLA-DRB1 kits (http://www.dynalbiotech.com) using a third of the specified volumes for the polymerase chain reaction reagents in a 20 μl reaction instead of 60 μl as described previously.34 Alleles were assigned using the Pattern Matching Program provided by Dynal (Invitrogen Ltd, Paisley, UK). Statistical analysis Allele and phenotype frequencies for HLA-Cw*06 were compared between PsA cases, psoriasis cases and controls using the χ2 test implemented in STATA 8. For the HLA-DRB1gene, HLA-DRB1*04, HLA-DRB1*07, HLA-DRB1*02 (HLA-DRB*15 or HLA-DRB*16) and SE phenotype frequencies were compared between cases with PsA, UIA and controls using the χ2 test implemented in STATA 8. The SE was defined by the presence of any of the following alleles: HLA-DRB1*0101, HLA-DRB1*0102, HLA-DRB1*0104, HLA-DRB1*0401, HLA-DRB1*0404, HLA-DRB1*0405, HLA-DRB1*0408 and HLA-DRB1*1001.

16) and SE phenotype frequencies were compared between cases with PsA, UIA and controls using the χ2 test implemented in STATA 8. The SE was defined by the presence of any of the following alleles: HLA-DRB1*0101, HLA-DRB1*0102, HLA-DRB1*0104, HLA-DRB1*0401, HLA-DRB1*0404, HLA-DRB1*0405, HLA-DRB1*0408 and HLA-DRB1*1001. The PsA cohort was divided into patients with type I and type II psoriasis by their age at onset of their psoriasis (⩽40 years or >40 years, respectively) and the analyses were repeated for each subset of PsA. Similar analysis was undertaken after stratifying the PsA cohort by their RF status. Linkage disequilibrium analysis Pairwise linkage disequilibrium (LD) measures (both D′ and r2) were investigated between HLA-Cw*06 and HLA-DRB1*07 using HelixTree (Golden Helix Inc, Bozenian, MT, USA). These data has been reported previously.34 RESULTS A summary of the samples used for the HLA-Cw*06 and HLA-DRB1 analysis is provided in table 1. Table 1 Number of subjects in study (where data are available) Subjects HLA-Cw*06 information: n HLA-DRB1 information: n PsA whole cohort 453 465 PsA with type I psoriasis 335 342 PsA with type II psoriasis 115 120 Psoriasis (all type I) 611 NA UIA NA 1621 Population controls 166 537 UIA, early undifferentiated inflammatory arthritis; PsA, psoriatic arthritis; NA, not available. For three PsA cases, data were not available as to the type of psoriasis present.

Table 1 Number of subjects in study (where data are available) Subjects HLA-Cw*06 information: n HLA-DRB1 information: n PsA whole cohort 453 465 PsA with type I psoriasis 335 342 PsA with type II psoriasis 115 120 Psoriasis (all type I) 611 NA UIA NA 1621 Population controls 166 537 UIA, early undifferentiated inflammatory arthritis; PsA, psoriatic arthritis; NA, not available. For three PsA cases, data were not available as to the type of psoriasis present. Patient characteristics Psoriatic arthritis The characteristics of the PsA cohort have been described previously.34 There was an almost equal gender distribution with 57% being female and 74% having type I psoriasis. The median duration of psoriasis was 19 years (interquartile range (IQR) 9–33) and the median duration of joint disease was 10 years (IQR 5–19). A majority (63%) developed psoriasis before the onset of joint disease. RF was present in 17% (titre >1:40), 81% had nail involvement, 57% had five or more damaged joints (polyarthritis subgroup) and the median HAQ score was 1.25. As shown previously, patients with PsA with type I psoriasis have a stronger family history of both skin and joint disease and tend to develop arthritis after the onset of psoriasis.19 37 In addition, patients with PsA with type I psoriasis had a longer duration of joint disease and more nail involvement, but a lower median HAQ score and fewer involved joints compared with those with type II psoriasis.

istory of both skin and joint disease and tend to develop arthritis after the onset of psoriasis.19 37 In addition, patients with PsA with type I psoriasis had a longer duration of joint disease and more nail involvement, but a lower median HAQ score and fewer involved joints compared with those with type II psoriasis. Psoriasis All patients had type I psoriasis with 46% (283 of 611) being female. The median age of onset of psoriasis was 19 years (IQR 13–27). Some of these patients may have an unrecognised IA, but a subset (n = 229) have been specifically examined by a dermatologist to exclude an IA. All 611 patients with psoriasis were included in the analysis. No significant differences in clinical, demographic or HLA-Cw*06 carriage data were observed between those patients with psoriasis in whom PsA had been specifically excluded and those where it had not (HLA-Cw*06 carriage was 43% in those without PsA versus 40% in the remainder, p = 0.44). Early undifferentiated inflammatory arthritis Within the UIA cohort, 1053 (65%) were female. At baseline, the median disease duration was 6 months (IQR 3–12); RF was present at a titre >1:40 in 452/1433 (31.5%) and 743 (48.3%) satisfied the American College of Rheumatology criteria for RA.38 By year 5, 11% of the patients were recorded to have psoriasis in addition to their IA and by year 10 of follow-up, 12% developed psoriasis. Controls In the population control cohort, gender information was available for 268 subjects of whom 119 (44%) were female. HLA-Cw*06 data were available for 166, while HLA-DRB1 data were available for 573 subjects.

Early undifferentiated inflammatory arthritis Within the UIA cohort, 1053 (65%) were female. At baseline, the median disease duration was 6 months (IQR 3–12); RF was present at a titre >1:40 in 452/1433 (31.5%) and 743 (48.3%) satisfied the American College of Rheumatology criteria for RA.38 By year 5, 11% of the patients were recorded to have psoriasis in addition to their IA and by year 10 of follow-up, 12% developed psoriasis. Controls In the population control cohort, gender information was available for 268 subjects of whom 119 (44%) were female. HLA-Cw*06 data were available for 166, while HLA-DRB1 data were available for 573 subjects. HLA-Cw*06 The frequency of the HLA-Cw*06 phenotype in the population controls tested in the current study was within the range reported previously15 16 25 26 39–41 (table 2). Table 2 Comparison of HLA-Cw*06 in PsA cases, psoriasis and controls HLA-Cw*06 PsA cases Population controls n = 166 Type I psoriasis n = 611 Total cohort n = 453 Type I n = 335 Type II n = 115 0 262 (57.8) 166 (49.6) 94 (81.7) 138 (83.1) 254 (41.6) 1 182 (40.2) 162 (48.4) 19 (16.5) 28 (16.9) 323 (52.9) 2 9 (2) 7 (2) 2 (1.8) 0 34 (5.5) Phenotype 191 (42.2) 169 (50.4) 21 (18.3) 28 (16.9) 357 (58.4) p-Value* 5.5×10−9 4.39×10−13 0.76 2.15×10−21 PsA, psoriatic arthritis. Type I, patients with PsA with type I psoriasis; type II, patients with PsA with type II psoriasis. *Comparison of phenotype (carriage of one or two alleles) with population controls using the χ2 test. Data shown in n (%). Data were not available for three patients as to the type of psoriasis present.

Table 2 Comparison of HLA-Cw*06 in PsA cases, psoriasis and controls HLA-Cw*06 PsA cases Population controls n = 166 Type I psoriasis n = 611 Total cohort n = 453 Type I n = 335 Type II n = 115 0 262 (57.8) 166 (49.6) 94 (81.7) 138 (83.1) 254 (41.6) 1 182 (40.2) 162 (48.4) 19 (16.5) 28 (16.9) 323 (52.9) 2 9 (2) 7 (2) 2 (1.8) 0 34 (5.5) Phenotype 191 (42.2) 169 (50.4) 21 (18.3) 28 (16.9) 357 (58.4) p-Value* 5.5×10−9 4.39×10−13 0.76 2.15×10−21 PsA, psoriatic arthritis. Type I, patients with PsA with type I psoriasis; type II, patients with PsA with type II psoriasis. *Comparison of phenotype (carriage of one or two alleles) with population controls using the χ2 test. Data shown in n (%). Data were not available for three patients as to the type of psoriasis present. When compared with controls, the HLA-Cw*06 phenotype was shown to be strongly associated with PsA (odds ratio (OR) 3.6, 95% CI 2.3, 5.8 and p = 5.5×10−9) (table 2 and fig 1). Stratification analysis in the PsA cohort by RF status made no difference to the result (OR in the RF-negative PsA subgroup 3.6, 95% CI 2.3, 5.9, p = 9.6×10−9). However, the association with HLA-Cw*06 was confined to the subgroup of patients with PsA with type I psoriasis (OR 5.0, 95% CI 3.2, 7.9, p = 4.39×10−13) and was not observed in patients with PsA with type II psoriasis (OR 1.1, 95% CI 0.6, 2.1, p = 0.76).

he RF-negative PsA subgroup 3.6, 95% CI 2.3, 5.9, p = 9.6×10−9). However, the association with HLA-Cw*06 was confined to the subgroup of patients with PsA with type I psoriasis (OR 5.0, 95% CI 3.2, 7.9, p = 4.39×10−13) and was not observed in patients with PsA with type II psoriasis (OR 1.1, 95% CI 0.6, 2.1, p = 0.76). Figure 1 Comparison of HLA-Cw*06 phenotype in psoriatic arthritis (PsA) cases, psoriasis and controls. Odds ratios and 95% confidence intervals are shown on a log scale. For the psoriasis cohort, as expected, HLA-Cw*06 was strongly associated with type I psoriasis compared with controls (OR 6.9 95% CI 4.4 to 11.1, p = 2.15×10−21) (table 2 and fig 1). To determine whether HLA-Cw*06 is associated with PsA itself or primarily with psoriasis, we compared the HLA-Cw*06 phenotype frequencies in those patients with PsA with type I psoriasis and patients with type I psoriasis alone. A much weaker association was noted (PsA with type I psoriasis versus type I psoriasis, OR 0.72, 95% CI 0.55, 0.96, p = 0.02), suggesting that the primary association of HLA-Cw*06 is with type I psoriasis and not PsA per se.

cies in those patients with PsA with type I psoriasis and patients with type I psoriasis alone. A much weaker association was noted (PsA with type I psoriasis versus type I psoriasis, OR 0.72, 95% CI 0.55, 0.96, p = 0.02), suggesting that the primary association of HLA-Cw*06 is with type I psoriasis and not PsA per se. HLA-DRB1 Stratification analysis showed that the HLA-DRB1*07 phenotype was strongly associated in those patients with PsA with type I psoriasis (OR 2.7, 95% CI 2.1, 3.7 and p<0.00001) compared with controls. Although, HLA-DRB1*07 occurred significantly more frequently in PsA cases than controls, we have previously reported that this allele exhibits considerable LD with HLA-Cw*06 (correlation (r2) = 0.46).34 Therefore, it was not unexpected that, after adjusting for the presence of HLA-Cw*06 phenotype, the association of HLA-DRB1*07 with PsA as a whole group (OR 1.38, 95% CI 0.88, 2.17, p = 0.16) or in the subgroup with type I psoriasis compared with controls (OR 1.63, 95% CI 0.96, 2.78, p = 0.07) was no longer statistically significant. However, the association of PsA with HLA-Cw*06 remained similar after adjusting for the presence of the HLA-DRB1*07 (OR 3.2, 95% CI 2.0 to 5.3), confirming that the primary association is with HLA-Cw*06 and not HLA-DRB1*07.

ed with controls (OR 1.63, 95% CI 0.96, 2.78, p = 0.07) was no longer statistically significant. However, the association of PsA with HLA-Cw*06 remained similar after adjusting for the presence of the HLA-DRB1*07 (OR 3.2, 95% CI 2.0 to 5.3), confirming that the primary association is with HLA-Cw*06 and not HLA-DRB1*07. Patients with PsA negative for RF were less likely to carry the HLA-DRB1*04 phenotype compared with population controls (OR 0.74, 95% CI 0.55, 0.99, p = 0.03), but no difference was observed between those patients with PsA with a positive RF compared with controls (OR 0.96, 95% CI 0.56, 1.61, p = 0.88). In addition, the HLA-DRB1*04 phenotype occurred less frequently in patients with PsA with type I psoriasis compared with population controls (p = 0.004), but no difference was observed in those patients with PsA with type II psoriasis compared with controls (p = 0.45). Within patients with PsA, when the HLA-DRB1*04 phenotype was present, it occurred more commonly in patients with PsA with type II psoriasis compared with those with type I psoriasis (OR 1.81, 95% CI 1.14, 2.86, p = 0.007). No association was detected with those patients with PsA having ⩾5 damaged (poly-damaged) or ⩾5 involved (poly-involved) joints with the HLA-DRB1*04 phenotype compared with population controls (OR 0.87, 95% CI 0.6, 1.2, p = 0.39 and OR 0.81, 95% CI 0.60, 1.08, p = 0.14, respectively).

81, 95% CI 1.14, 2.86, p = 0.007). No association was detected with those patients with PsA having ⩾5 damaged (poly-damaged) or ⩾5 involved (poly-involved) joints with the HLA-DRB1*04 phenotype compared with population controls (OR 0.87, 95% CI 0.6, 1.2, p = 0.39 and OR 0.81, 95% CI 0.60, 1.08, p = 0.14, respectively). Table 3 shows that in the UIA cohort, the HLA-DRB1*04 phenotype was more common than population controls (OR 1.40, 95% CI 1.14, 1.72 and p = 0.001), while the HLA-DRB1*07 phenotype was more common in the PsA cohort compared with population controls (OR 2.15, 95% CI 1.62, 2.84, p = 2.6×10−8). However, when comparing patients with PsA with type II psoriasis with UIA subjects, no difference was observed between these two cohorts for either the HLA-DRB1*04 or the HLA-DRB1*07 phenotypes (p = 0.35 and p = 0.45 respectively). When compared with patients with PsA with type I psoriasis, the frequency of the HLA-DRB1*04 phenotype was significantly higher in UIA subjects (OR 2.17, 95% CI 1.66, 2.84, p<0.0001). Conversely, the HLA-DRB1*07 phenotype was significantly higher in patients with PsA with type I psoriasis compared with UIA (OR 3.23, 95% CI 2.51, 4.14, p<0.0001).

with PsA with type I psoriasis, the frequency of the HLA-DRB1*04 phenotype was significantly higher in UIA subjects (OR 2.17, 95% CI 1.66, 2.84, p<0.0001). Conversely, the HLA-DRB1*07 phenotype was significantly higher in patients with PsA with type I psoriasis compared with UIA (OR 3.23, 95% CI 2.51, 4.14, p<0.0001). Table 3 HLA-DRB1 phenotypes in PsA cases, UIA and controls (where data available) HLA-DRB1 PsA cases Controls n = 537 p value (PsA whole cohort compared with controls) UIA n = 1621 p Value (UIA compared with controls) Whole cohort n = 465 Type I n = 342 Type II n = 120 DRB1*02 118 (25) 81 (24) 36 (30) 145 (27) 0.57 342 (21) 0.006 DRB1*04 142 (31) 92 (27) 48 (40) 195 (36) 0.06 719 (44) 0.001 DRB1*07 188 (40) 159 (46) 29 (24) 129 (24) 3.09×10−8 344 (21) 0.21 PsA, psoriatic arthritis; UIA, early undifferentiated inflammatory arthritis; type I, patients with PsA with type I psoriasis; type II, patients with PsA with type II psoriasis. Data shown in n (%) unless stated otherwise. Previous studies have reported conflicting results with regard to the HLA-DRB1*02 (HLA-DRB*15 or HLA-DRB*16) and PsA susceptibility with a decrease found in a UK study, but not in a cohort from Toronto.28 29 We did not observe a difference in our cohort.

Table 3 HLA-DRB1 phenotypes in PsA cases, UIA and controls (where data available) HLA-DRB1 PsA cases Controls n = 537 p value (PsA whole cohort compared with controls) UIA n = 1621 p Value (UIA compared with controls) Whole cohort n = 465 Type I n = 342 Type II n = 120 DRB1*02 118 (25) 81 (24) 36 (30) 145 (27) 0.57 342 (21) 0.006 DRB1*04 142 (31) 92 (27) 48 (40) 195 (36) 0.06 719 (44) 0.001 DRB1*07 188 (40) 159 (46) 29 (24) 129 (24) 3.09×10−8 344 (21) 0.21 PsA, psoriatic arthritis; UIA, early undifferentiated inflammatory arthritis; type I, patients with PsA with type I psoriasis; type II, patients with PsA with type II psoriasis. Data shown in n (%) unless stated otherwise. Previous studies have reported conflicting results with regard to the HLA-DRB1*02 (HLA-DRB*15 or HLA-DRB*16) and PsA susceptibility with a decrease found in a UK study, but not in a cohort from Toronto.28 29 We did not observe a difference in our cohort. Shared epitope When comparing patients with PsA and controls, no association was detected with SE allele carriage and PsA (table 4 and fig 2) either in the entire cohort (p = 0.94) or after stratifying the patients with PsA by type I or type II psoriasis or by their RF status. For example, no association was detected when comparing SE phenotype between patients with PsA with type I psoriasis and population controls (OR 0.90, 95% CI 0.68, 1.19, p = 0.43) or when comparing patients with PsA with type II psoriasis with population controls (OR of 1.24, 95% CI 0.82, 1.88, p = 0.28).

ir RF status. For example, no association was detected when comparing SE phenotype between patients with PsA with type I psoriasis and population controls (OR 0.90, 95% CI 0.68, 1.19, p = 0.43) or when comparing patients with PsA with type II psoriasis with population controls (OR of 1.24, 95% CI 0.82, 1.88, p = 0.28). Figure 2 Shared epitope phenotype in psoriatic arthritis (PsA) cases, undifferentiated inflammatory arthritis and controls. Odds ratios and 95% confidence intervals are shown on a log scale. Table 4 Shared epitope frequency in PsA cases, UIA and controls Shared epitope UIA n = 1621 PsA cases Population controls n = 537 Whole cohort n = 467 Type I n = 344 Type II n = 120 0 634 (39.1) 256 (54.8) 197 (57.3) 59 (49.2) 293 (54.6) 1 740 (45.7) 184 (39.4) 126 (36.6) 55 (45.8) 203 (37.8) 2 247 (15.2) 27 (5.8) 21 (6.1) 6 (5.0) 41 (7.6) Phenotype 987 (60.9) 211 (45.2) 147 (42.7) 61 (50.8) 244 (45.4) p Value* 3.63×10−10 0.94 0.43 0.28 UIA, early undifferentiated inflammatory arthritis; PsA, psoriatic arthritis; type I, patients with PsA with type I psoriasis; type II, patients with PsA with type II psoriasis. *Comparison of phenotype (carriage of one or two alleles) with population controls. Data shown in n (%) unless stated otherwise.

Figure 2 Shared epitope phenotype in psoriatic arthritis (PsA) cases, undifferentiated inflammatory arthritis and controls. Odds ratios and 95% confidence intervals are shown on a log scale. Table 4 Shared epitope frequency in PsA cases, UIA and controls Shared epitope UIA n = 1621 PsA cases Population controls n = 537 Whole cohort n = 467 Type I n = 344 Type II n = 120 0 634 (39.1) 256 (54.8) 197 (57.3) 59 (49.2) 293 (54.6) 1 740 (45.7) 184 (39.4) 126 (36.6) 55 (45.8) 203 (37.8) 2 247 (15.2) 27 (5.8) 21 (6.1) 6 (5.0) 41 (7.6) Phenotype 987 (60.9) 211 (45.2) 147 (42.7) 61 (50.8) 244 (45.4) p Value* 3.63×10−10 0.94 0.43 0.28 UIA, early undifferentiated inflammatory arthritis; PsA, psoriatic arthritis; type I, patients with PsA with type I psoriasis; type II, patients with PsA with type II psoriasis. *Comparison of phenotype (carriage of one or two alleles) with population controls. Data shown in n (%) unless stated otherwise. As expected, the SE was significantly associated with UIA compared with controls (OR 1.88, 95% CI 1.53, 2.29, p = 3.63×10−10; table 4 and fig 2). When comparing the UIA and PsA subgroups directly, the frequency of the SE phenotype was significantly higher in UIA subjects compared with those patients with PsA with type I psoriasis (OR 2.08, 95% CI 1.64, 2.66, p = 7.1×10−10), but the effect size was lower when comparing UIA subjects to patients with PsA with type II psoriasis (OR 1.51, 95% CI 1.02, 2.22, p = 0.03).

he frequency of the SE phenotype was significantly higher in UIA subjects compared with those patients with PsA with type I psoriasis (OR 2.08, 95% CI 1.64, 2.66, p = 7.1×10−10), but the effect size was lower when comparing UIA subjects to patients with PsA with type II psoriasis (OR 1.51, 95% CI 1.02, 2.22, p = 0.03). DISCUSSION In this large association study of patients with PsA, we have shown that both HLA-Cw*06 and HLA-DRB1*07 are associated with PsA susceptibility in the subgroup of patients with PsA with type I psoriasis but not in those with type II psoriasis, suggesting that the primary association is with type I psoriasis. Our data also confirms that the association with the HLA-DRB1*07 is because this allele is in LD with HLA-Cw*06. We can find no evidence for association of the SE with PsA susceptibility.

PsA with type I psoriasis but not in those with type II psoriasis, suggesting that the primary association is with type I psoriasis. Our data also confirms that the association with the HLA-DRB1*07 is because this allele is in LD with HLA-Cw*06. We can find no evidence for association of the SE with PsA susceptibility. The simultaneous investigation of two HLA genes in large cohorts of patients with PsA and appropriate control groups has enabled us to try and dissect out the contribution to PsA over and above that to UIA alone and psoriasis alone. We did not investigate HLA-Cw*06 in UIA patients or HLA-DRB1 in patients with psoriasis as previous studies have excluded a role for these genes in the respective conditions.12 30–33 Previous studies on patients with psoriasis have shown that both HLA-Cw*06 and HLA-DRB1*07 are associated with type I psoriasis, but not with type II psoriasis.12 30 31 42 We have confirmed that, although both phenotypes show association, the HLA-DRB1*07 result has arisen due to LD with HLA-Cw*06. Furthermore, analysis of HLA-Cw*06 phenotype in patients with type I psoriasis, patients with PsA (stratified according to their types of psoriasis) as well as population controls has allowed us to conclude that the association is primarily with type I psoriasis rather than PsA itself.

arisen due to LD with HLA-Cw*06. Furthermore, analysis of HLA-Cw*06 phenotype in patients with type I psoriasis, patients with PsA (stratified according to their types of psoriasis) as well as population controls has allowed us to conclude that the association is primarily with type I psoriasis rather than PsA itself. Unsurprisingly, HLA-DRB1*04 was found to occur more frequently in the UIA cohort, the majority of whom satisfied American College of Rheumatology classification criteria for RA by 5 years. In contrast, it occurred less commonly in those patients with PsA who were negative for RF and those patients with PsA with type I psoriasis compared with controls. When HLA-DRB1*04 was present in patients with PsA, it occurred more frequently in those with type II psoriasis compared with patients with PsA with type I psoriasis. This may suggest that the HLA-DRB1*04 allele is protective for type I PsA. Alternatively, it may simply reflect the fact that if one allele is increased in frequency (HLA-DR*07 allele frequency increased in patients with PsA with type I psoriasis) then the frequency of others must be reduced.

PsA with type I psoriasis. This may suggest that the HLA-DRB1*04 allele is protective for type I PsA. Alternatively, it may simply reflect the fact that if one allele is increased in frequency (HLA-DR*07 allele frequency increased in patients with PsA with type I psoriasis) then the frequency of others must be reduced. The broad inclusion criteria for PsA used in this study may have led to the misclassification of some patients who have true RA and coincidental psoriasis being classified as PsA. In this situation, one may have expected to see an association with HLA-DRB1*04 in either the RF-positive subgroup of patients with PsA or those with polyarticular disease but we did not find that to be the case. An advantage of not excluding RF-positive patients is that we have been able to stratify the patients with PsA by their RF status to explore whether RF is an important co-factor in PsA susceptibility. However, in no situation did this stratification change the conclusions of an analysis, suggesting that it is not. Our genetic findings, therefore, accord with recent clinical data suggesting that polyarticular PsA is more similar to oligoarticular PsA than to RA.43

r RF is an important co-factor in PsA susceptibility. However, in no situation did this stratification change the conclusions of an analysis, suggesting that it is not. Our genetic findings, therefore, accord with recent clinical data suggesting that polyarticular PsA is more similar to oligoarticular PsA than to RA.43 We have also confirmed the finding of others that SE was not associated with PsA susceptibility.28 29 Unsurprisingly, SE was found to be strongly associated with UIA susceptibility and occurred significantly more frequently in UIA subjects compared with patients with PsA with type I psoriasis. The smaller number of patients with PsA with type II psoriasis may have limited the interpretation of this analysis but the odds of carrying SE was also significantly higher in UIA subjects compared with patients with PsA with type II psoriasis. The findings confirm that, although patients with PsA with type II psoriasis appear more genetically similar to UIA subjects than do patients with PsA with type I psoriasis, UIA and patients with PsA with type II psoriasis are sufficiently different, in terms of their genetic susceptibility, to be viewed as distinct entities.

dings confirm that, although patients with PsA with type II psoriasis appear more genetically similar to UIA subjects than do patients with PsA with type I psoriasis, UIA and patients with PsA with type II psoriasis are sufficiently different, in terms of their genetic susceptibility, to be viewed as distinct entities. In summary, our study confirms the established strong association of HLA-Cw*06 with type I psoriasis susceptibility. The association of PsA with HLA-Cw*06 is of similar strength and is confined to those patients with PsA with type I psoriasis. We conclude, therefore, that HLA-Cw*06 does not confer additional susceptibility to IA in patients with psoriasis. We also note that the association between HLA-DRB1*07 and PsA is due to its significant LD with HLA-Cw*06. No independent association was detected with HLA-DRB1*02, HLA-DRB1*04, HLA-DRB1*07 or with the SE and PsA. Our study suggests that the genetic susceptibility of PsA cannot be explained by the HLA-Cw*06 or HLA-DRB1 loci and confirms the importance of choosing the appropriate control populations when studying this condition. The findings also suggest that adjustment for HLA-Cw*06 is of central importance when attempting to dissect the genetic susceptibility of IA in patients with psoriasis. Finally, our findings suggest that patients with PsA with type I and type II psoriasis have different genetic susceptibility factors and that patients with PsA with type I psoriasis are more genetically similar to type I patients with psoriasis at least at the HLA-Cw*06 locus. This implies that in future genetic studies, it may be important to stratify patients with PsA according to the age of onset of psoriasis.

erent genetic susceptibility factors and that patients with PsA with type I psoriasis are more genetically similar to type I patients with psoriasis at least at the HLA-Cw*06 locus. This implies that in future genetic studies, it may be important to stratify patients with PsA according to the age of onset of psoriasis. This work has been supported by the Arthritis Research Campaign, the Wellcome Trust (Entry Level Fellowships for PYPCH and HSY; Advanced Fellowship for AB) and the Medical Research Council UK (Clinical Training Fellowship for PYPCH). Competing interests: None.

Polymyositis and dermatomyositis are chronic muscle disorders characterised by muscle weakness and fatigue and by skin involvement in the case of dermatomyositis.1 These diseases are also characterised by infiltration of inflammatory cells in skeletal muscle tissue, muscle fibre degeneration and regeneration. The pathogenesis of myositis has not been well characterised yet, but pro-inflammatory cytokines have been consistently found in the inflamed muscle and are implicated in the pathogenesis.2–4

also characterised by infiltration of inflammatory cells in skeletal muscle tissue, muscle fibre degeneration and regeneration. The pathogenesis of myositis has not been well characterised yet, but pro-inflammatory cytokines have been consistently found in the inflamed muscle and are implicated in the pathogenesis.2–4 Arachidonic acid metabolites such as prostaglandins (PG) might also contribute to the pathogenesis of inflammatory myositis. Human skeletal muscles have a considerable capacity to produce PGE2, PGD2, PGF2α and PGI2.5 PGE2 appears to be involved in a number of biological processes, including protein turnover and myogenesis, and is a potent mediator of muscular pain and inflammation.6–10 Interleukin (IL)1β and tumour necrosis factor (TNF), which are markedly expressed in myositis muscle tissue, stimulate PGE2 production in skeletal muscles.2 3 11 12 In the PGE2 biosynthetic pathway, cyclooxygenase (COX)-1 and COX-2 catalyse the conversion of arachidonic acid into PGH2 (fig 1). Recently, enhanced expression of COX-1 and COX-2 mRNA was demonstrated in inflamed muscle tissue from patients with myositis, suggesting a role for them in this disease.13 Three terminal PGE synthases (PGES) catalyse the formation of PGE2 from PGH2 Microsomal PGE synthase 1 (mPGES-1) is strongly induced by proinflammatory stimuli in various cells and preferentially couples with COX-2 contributing to significant PGE2 release.14–16 Cytosolic PGES (cPGES) and mPGES-2 are constitutively expressed and likely to function in the basal production of PGE217 18 Studies of mPGES-1 –/– knock-out mice have demonstrated a critical role for mPGES-1 in the development of pain, fever and inflammation.19 20 mPGES-1 is upregulated in a range of inflammatory diseases and considered a new target for therapeutic strategies to control induced PGE2 synthesis.21–23 However, the expression of mPGES-1 in muscle tissue from patients with myositis has not been studied.

mPGES-1 in the development of pain, fever and inflammation.19 20 mPGES-1 is upregulated in a range of inflammatory diseases and considered a new target for therapeutic strategies to control induced PGE2 synthesis.21–23 However, the expression of mPGES-1 in muscle tissue from patients with myositis has not been studied. Figure 1 A schematic overview of the prostaglandin (PG) biosynthesis cascade. Despite clinical improvement with conventional immunosuppressive treatment including high doses of glucocorticoids (GC), many patients with myositis experience persistent muscle weakness. There are also reports of persisting inflammatory cells and increased expression of IL1 in muscle tissue despite long-time treatment with high doses of GC.3 24 Intra-articular treatment with GC significantly reduces mPGES-1, COX-1 and COX-2 expression in the synovial tissue from patients with rheumatoid arthritis and is associated with clinical improvement.25 Whether mPGES-1 and COX expression in skeletal muscle from patients with myositis is affected by immunosuppressive treatment has not been investigated to date. In the present study we examined the expression and localisation of mPGES-1 and COX in skeletal muscle tissue from patients with polymyositis or dermatomyositis and from healthy subjects. In addition we studied the effects of immunosuppressive treatment on mPGES-1 and COX expression in skeletal muscle tissue from patients with myositis in relation to the effects on clinical function.

f mPGES-1 and COX in skeletal muscle tissue from patients with polymyositis or dermatomyositis and from healthy subjects. In addition we studied the effects of immunosuppressive treatment on mPGES-1 and COX expression in skeletal muscle tissue from patients with myositis in relation to the effects on clinical function. PATIENTS AND METHODS Patients and muscle biopsies In the first cohort, nine patients with recently diagnosed polymyositis or dermatomyositis and three with treatment-resistant myositis (disease duration 3–6 years) meeting the Bohan and Peter criteria were included (median age 54 years, range 44–76 years).26 27 Clinical data of the patients are presented in tables 1 and 2. Detailed clinical data for these patients have been reported previously.28

atomyositis and three with treatment-resistant myositis (disease duration 3–6 years) meeting the Bohan and Peter criteria were included (median age 54 years, range 44–76 years).26 27 Clinical data of the patients are presented in tables 1 and 2. Detailed clinical data for these patients have been reported previously.28 Table 1 Clinical data on the patients at the time of biopsies, cohort 1: patients with recently diagnosed and with treatment-resistant myositis (marked with *) Patient Age Sex Bohan and Peter diagnostic criteria Biopsy site Treatment at time of biopsy, mg/day 1 53 F Probable polymyositis Mus vastus lat None 2 59 F Probable polymyositis Mus deltoideus Pred 5, NSAID 3 76 F Definite polymyositis Mus vastus lat None 4 44 F Definite polymyositis Mus vastus lat None 5 65 F Definite dermatomyositis Mus vastus lat Pred 15 6 72 F Definite polymyositis Mus vastus lat Pred 15–40, AZA 7 60 M Probable polymyositis Mus vastus lat NSAID 8 56 M Definite polymyositis Mus vastus lat NSAID 9 55 M Probable dermatomyositis Mus vastus lat None 10* 67 M Definite dermatomyositis Mus vastus lat AZA, MTX, Cs 11* 54 F Probable polymyositis Mus tib ant AZA, MTX, Cs 12* 61 F Definite polymyositis Mus vastus lat AZA, MTX, Cs Ant, anterior; AZA, azathioprine; Cs, ciclosporine; F, female; lat, lateralis; M, male; Mus, musculus; MTX, methotrexate; NSAID, non-steroidal anti-inflammatory drugs; tib, tibialis; Pred, prednisolone.

Cs 11* 54 F Probable polymyositis Mus tib ant AZA, MTX, Cs 12* 61 F Definite polymyositis Mus vastus lat AZA, MTX, Cs Ant, anterior; AZA, azathioprine; Cs, ciclosporine; F, female; lat, lateralis; M, male; Mus, musculus; MTX, methotrexate; NSAID, non-steroidal anti-inflammatory drugs; tib, tibialis; Pred, prednisolone. Table 2 Clinical data on the patients at the time of biopsies, cohort 2: myositis patients before and after conventional immunosuppressive treatment Patient Age Sex Bohan and Peter diagnostic criteria Treatment at time of biopsy 1, mg/day Treatment at time of biopsy 2, mg/day 1 60 F Definite polymyositis None Pred 12.5, MTX 2 23 F Probable dermatomyositis None Pred 5, AZA 3 49 F Definite dermatomyositis NSAID Pred 15, NSAID 4 53 F Definite polymyositis None Pred 20, AZA 5 41 F Probable dermatomyositis None Pred 30 6 44 F Probable dermatomyositis None Pred 7.5 7 67 M Probable polymyositis None Pred 10 8 88 F Definite dermatomyositis None Pred 7.5 9 60 M Probable polymyositis NSAID Pred 30 10 62 M Definite dermatomyositis None Pred 10, AZA AZA, azathioprine; F, female; M, male; MTX, methotrexate; NSAID, non-steroidal anti-inflammatory drugs; Pred, prednisolone.

None Pred 7.5 7 67 M Probable polymyositis None Pred 10 8 88 F Definite dermatomyositis None Pred 7.5 9 60 M Probable polymyositis NSAID Pred 30 10 62 M Definite dermatomyositis None Pred 10, AZA AZA, azathioprine; F, female; M, male; MTX, methotrexate; NSAID, non-steroidal anti-inflammatory drugs; Pred, prednisolone. To investigate the effects of immunosuppressive treatment a second cohort of 10 patients with recently diagnosed polymyositis or dermatomyositis with available follow-up biopsies was included (median age 56.5 years, range 23–88 years). The patients were initially treated with oral prednisolone (40–60 mg/day) with slowly tapering doses and all patients (except two) received an additional immunosuppressive agent (tables 1 and 2). Two patients were treated with non-steroidal anti-inflammatory drugs (NSAIDs) at the time of the first biopsy. Muscle tissue biopsies were taken from m. vastus lateralis before and after a median of 8.5 months (range 4–11 months) with immunosuppressive treatment. Muscle tissue biopsies were obtained under local anaesthesia using the semi-open biopsy technique.28 29 Muscle biopsies from seven healthy individuals (four women and three men, median age 46 years, range 38–50 years) without clinical or histopathological signs of muscle disease were included as controls. Muscle biopsies were obtained from musculus vastus lateralis in six individuals and from musculus tibialis anterior in one individual.

rom seven healthy individuals (four women and three men, median age 46 years, range 38–50 years) without clinical or histopathological signs of muscle disease were included as controls. Muscle biopsies were obtained from musculus vastus lateralis in six individuals and from musculus tibialis anterior in one individual. The approval was granted by the Ethics Committee at the Karolinska University Hospital, Stockholm and all patients and controls gave their informed consent to participate in the study. Clinical assessment of patients Muscle performance (functional index (FI) of myositis) was assessed by the number of repetitions performed in defined muscle groups before treatment and at the time of the second biopsy.30 The individual total score is presented as percentage of the maximal score 64 (mean values of left and right side). The responder criterion was set to 20% improvement.

(FI) of myositis) was assessed by the number of repetitions performed in defined muscle groups before treatment and at the time of the second biopsy.30 The individual total score is presented as percentage of the maximal score 64 (mean values of left and right side). The responder criterion was set to 20% improvement. Immunohistochemical analysis Staining of serial cryostat sections with mouse monoclonal anti-CD3 (BD Biosciences, San José, California, USA), anti-CD68 (marker of monocyte/macrophage lineage, KP-1 clone, Dako Cytomation, Glostrup, Denmark) and anti-CD163 (resident tissue macrophage marker, BerMac3 clone, Dako Cytomation) antibodies was performed using a standard protocol.31 Staining with rabbit polyclonal anti-human mPGES-1 antiserum,22 polyclonal anti-cPGES, anti-COX-2 and anti-COX-1 (Cayman Chemicals, Ann Arbor, Michigan, USA) and mouse monoclonal anti-COX-1 antibodies (Wako Chemicals, Neuss, Germany) was performed as previously described.32 Staining in skeletal muscle tissue was halted by preincubation of anti-mPGES-1 serum with mPGES-1 protein and by preincubation of commercial antibodies with respective blocking peptides (Cayman Chemical). Isotype-matched irrelevant antibodies were used as negative controls. The first and the last sections from each series of consecutive sections were stained with haematoxylin and eosin to evaluate the number of inflammatory infiltrates.

bation of commercial antibodies with respective blocking peptides (Cayman Chemical). Isotype-matched irrelevant antibodies were used as negative controls. The first and the last sections from each series of consecutive sections were stained with haematoxylin and eosin to evaluate the number of inflammatory infiltrates. Stained tissue sections were examined using a Polyvar II microscope (Reichert-Jung, Vienna, Austria) and photographed with a digital Leica camera 300F (Leica, Cambridge, UK). The number of PGES and COX positive cells was assessed by conventional microscopy measurements of the entire tissue section (2–9 mm2) using semi-quantitative scale: 0, no staining; 1, a few stained scattered cells; 2, many stained scattered cells; 3, many stained scattered cells and cells in one infiltrate; 4, strong staining in many scattered cells and several infiltrates. Evaluation of coded sections was performed by two independent observers. The mean scores from the two assessments were used for statistical analysis. Expression of CD3, CD68 and CD163 was assessed quantitatively using computer-assisted image analysis. The images were analysed with a Quantimet 600 image analyser (Leica) and positive staining was expressed as percentage of total counterstained tissue area. Double immunofluorescence was performed using anti-human mPGES-1 antiserum and mouse monoclonal anti-CD3, anti-CD68, anti-CD163 and anti-prolyl-4-hydroxylase (fibroblast marker, 5B5 clone, Dako Cytomation) antibodies as published previously.22

Expression of CD3, CD68 and CD163 was assessed quantitatively using computer-assisted image analysis. The images were analysed with a Quantimet 600 image analyser (Leica) and positive staining was expressed as percentage of total counterstained tissue area. Double immunofluorescence was performed using anti-human mPGES-1 antiserum and mouse monoclonal anti-CD3, anti-CD68, anti-CD163 and anti-prolyl-4-hydroxylase (fibroblast marker, 5B5 clone, Dako Cytomation) antibodies as published previously.22 Statistical analysis Data were analysed using the Mann–Whitney U test and Wilcoxon signed rank test and Bonferroni corrections for multiple comparisons. p Values <0.05 were considered statistically significant. Correlation between muscle FI and enzyme expression in muscle tissue was analysed using Spearman rank correlation test.

l analysis Data were analysed using the Mann–Whitney U test and Wilcoxon signed rank test and Bonferroni corrections for multiple comparisons. p Values <0.05 were considered statistically significant. Correlation between muscle FI and enzyme expression in muscle tissue was analysed using Spearman rank correlation test. RESULTS Expression of PGES and COX A marked mPGES-1 staining localised in many scattered mononuclear cells and in mononuclear cells in infiltrates was observed in all patients with some interindividual variations (fig 2A). For 6 patients mPGES-1 staining was additionally localised to smooth muscle cells in large vessels, and for 10 patients to capillaries as well. In muscle tissue from healthy individuals, weak mPGES-1 staining was detected in a few scattered mononuclear cells and capillaries (fig 2B), as well as in smooth muscle cells in large vessels in two individuals. Using conventional microscopic assessment, we found that mPGES-1 expression in muscle tissue from patients with myositis (n = 12) was significantly higher (p<0.01) when compared to healthy individual tissue (n = 7) (fig 3).

and capillaries (fig 2B), as well as in smooth muscle cells in large vessels in two individuals. Using conventional microscopic assessment, we found that mPGES-1 expression in muscle tissue from patients with myositis (n = 12) was significantly higher (p<0.01) when compared to healthy individual tissue (n = 7) (fig 3). Figure 2 Immunohistochemical staining (brown) for (A, B) microsomal prostaglandin E synthase (mPGES)-1, (C, D) cyclooxygenase (COX)-1 and (E, F) COX-2 in representative muscle tissue sections counterstained with haematoxylin (A, C, E) from patients with polymyositis and (B, D, F) from healthy individuals (original magnification ×500). Figure 3 Expression of microsomal prostaglandin E synthase (mPGES)-1, cytosolic PGES (cPGES), cyclooxygenase (COX)-1 and COX-2 in muscle tissue from patients with polymyositis or dermatomyositis and from healthy individuals. Results are expressed as positive cell score mean (standard error of mean (SEM)). *p<0.05, patients vs healthy individuals. Staining of cPGES was observed in scattered mononuclear cells, cells surrounding large vessels and muscle fibres in myositis and healthy muscle tissues with a similar distribution pattern (data not shown). In two patient biopsies, cPGES positive cells were detected in inflammatory cells in infiltrates. The staining for cPGES in the myositis muscle was not significantly different when compared to healthy controls (fig 3).

e fibres in myositis and healthy muscle tissues with a similar distribution pattern (data not shown). In two patient biopsies, cPGES positive cells were detected in inflammatory cells in infiltrates. The staining for cPGES in the myositis muscle was not significantly different when compared to healthy controls (fig 3). COX-1 expression was detected in muscle fibres, blood vessels and scattered mononuclear cells in patients and healthy subjects. In patients a strong COX-1 staining was additionally observed in mononuclear cells within the inflammatory infiltrates (fig 2C, D). COX-2 was expressed in macrophage-like cells within inflammatory infiltrates, in scattered mononuclear cells and in some large vessels in myositis muscle tissue. By contrast, in healthy muscle tissue COX-2 staining was only detected in few scattered mononuclear cells and in some vessels (fig 2E, F). COX-1 and COX-2 expression were significantly enhanced (p<0.01) in myositis muscle tissue when compared to healthy control (fig 3). Double immunofluorescence revealed the expression of mPGES-1 in CD163-positive (fig 4A–C) and of CD68-positive macrophages (data not shown). However, we could not detect any mPGES-1 staining in T lymphocytes or fibroblasts (data not shown).

COX-1 expression was detected in muscle fibres, blood vessels and scattered mononuclear cells in patients and healthy subjects. In patients a strong COX-1 staining was additionally observed in mononuclear cells within the inflammatory infiltrates (fig 2C, D). COX-2 was expressed in macrophage-like cells within inflammatory infiltrates, in scattered mononuclear cells and in some large vessels in myositis muscle tissue. By contrast, in healthy muscle tissue COX-2 staining was only detected in few scattered mononuclear cells and in some vessels (fig 2E, F). COX-1 and COX-2 expression were significantly enhanced (p<0.01) in myositis muscle tissue when compared to healthy control (fig 3). Double immunofluorescence revealed the expression of mPGES-1 in CD163-positive (fig 4A–C) and of CD68-positive macrophages (data not shown). However, we could not detect any mPGES-1 staining in T lymphocytes or fibroblasts (data not shown). Figure 4 Double fluorescence staining demonstrating cellular localisation of microsomal prostaglandin E synthase (mPGES)-1: (A) mPGES-1 positive (green) cells, (B) CD163 positive (red) cells and (C) double stained (yellow) cells in representative muscle tissue section from the patient with polymyositis (original magnification ×250). Effects of immunosuppressive treatment on PGES and COX expression The expression of mPGES-1 in myositis muscle tissue was not altered by the treatment (n = 10) (fig 5A). In patients without infiltrates after treatment (4 out of 10), mPGES-1 staining was still apparent in scattered mononuclear cells. Likewise, the distribution pattern or score for cPGES and COX-1 positive cells remained unchanged (fig 5A). By contrast, the score for COX-2 positive cells in muscle tissue was significantly reduced (p<0.01) after treatment (fig 5A, B). There was no significant difference between polymyositis and dermatomyositis regarding the expression pattern of these enzymes before or after treatment.

ve cells remained unchanged (fig 5A). By contrast, the score for COX-2 positive cells in muscle tissue was significantly reduced (p<0.01) after treatment (fig 5A, B). There was no significant difference between polymyositis and dermatomyositis regarding the expression pattern of these enzymes before or after treatment. Figure 5 A. Expression of microsomal prostaglandin E synthase (mPGES)-1 and related enzymes in muscle tissues of patients with polymyositis or dermatomyositis before and after conventional treatment. Results are expressed as positive cell score mean (standard error of mean (SEM)). B. Expression of cyclooxygenase (COX)-2 in muscle tissues of patients with polymyositis or dermatomyositis before and after conventional treatment. Results are expressed as positive cell score. C. Number of inflammatory cells in muscle tissue from patients with polymyositis and dermatomyositis before and after conventional treatment. Results are expressed as percentage of the total area of counterstained tissue. D. Effects of conventional treatment on muscle function in patients with polymyositis or dermatomyositis. Results are expressed as percentage of functional index (FI) for each patient. *p<0.05, before vs after treatment. Effects of immunosuppressive treatment on muscle histopathology Before treatment, 8 out of 10 patients presented infiltrates of mononuclear cells. After 4–11 months of immunosuppressive treatment the number of inflammatory infiltrates tended to be lower (median 0.52, range 0–2.6 infiltrates/mm2 vs median 0.22, range 0–2.4 infiltrates/mm2 before and after treatment, respectively), but this difference did not reach statistical significance.

es of mononuclear cells. After 4–11 months of immunosuppressive treatment the number of inflammatory infiltrates tended to be lower (median 0.52, range 0–2.6 infiltrates/mm2 vs median 0.22, range 0–2.4 infiltrates/mm2 before and after treatment, respectively), but this difference did not reach statistical significance. Positive staining for CD3, CD68 and CD163 markers was detected in infiltrates, in scattered cells and in cells surrounding large vessels. The total number of monocytes/macrophages (CD68 positive cells) was significantly reduced after the treatment (p<0.01), while the positively stained area was not altered for T lymphocytes or CD163-positive resident macrophages (fig 5C). Effects of immunosuppressive treatment on muscle function Before treatment, muscle weakness in patients was confirmed by a reduced FI (median 38.4%, range 22–65%), performed in all but one patient (n = 9). After 4–11 months of immunosuppressive treatment, a significantly increased FI was recorded (median 86.3%, range 37.5–100%, p<0.05), reflecting improved muscle function (fig 5D). On an individual basis, 8 out of 10 patients had improved by more than 20% in FI score. For one patient a decreased FI score was noted and for another the results of the first FI was not available. Despite the marked improvement in muscle function during the treatment period, a majority of the patients still had FI scores below the maximal value. Only one patient improved up to maximal score (100%).

core. For one patient a decreased FI score was noted and for another the results of the first FI was not available. Despite the marked improvement in muscle function during the treatment period, a majority of the patients still had FI scores below the maximal value. Only one patient improved up to maximal score (100%). There was no correlation between enzyme expression in muscle tissue and FI before or after treatment and the changes in enzyme expression did not correlate with changes in FI after treatment. In the patient who deteriorated clinically, the score for COX-2 expression was decreased after the treatment (from 2.5 to 1.5), in similar fashion to other patients who clinically improved.

tissue and FI before or after treatment and the changes in enzyme expression did not correlate with changes in FI after treatment. In the patient who deteriorated clinically, the score for COX-2 expression was decreased after the treatment (from 2.5 to 1.5), in similar fashion to other patients who clinically improved. DISCUSSION The present study demonstrates significantly increased expression of mPGES-1, COX-2 and COX-1 in skeletal muscle tissue from patients with polymyositis or dermatomyositis compared to that from healthy individuals. In healthy muscle tissue COX-1 and cPGES proteins were detected in scattered mononuclear cells, in vessels and in muscle fibres, suggesting that these enzymes account for the basal production of PGE2. In healthy muscle tissue we also observed mPGES-1 and COX-2 staining in few scattered mononuclear cells and blood vessels, indicating constitutive expression of these enzymes in muscle tissue. Constitutive expression of mPGES-1 and COX-2 has been reported in other tissues, eg, kidney, suggesting their possible role for basal PGE2 production under non-pathological conditions.33 34 Moreover, increased expression of COX-2 and mPGES-1 has been demonstrated in response to non-inflammatory stimuli, such as mechanical stretch and mechanical stress in certain cells.8 35 36

rted in other tissues, eg, kidney, suggesting their possible role for basal PGE2 production under non-pathological conditions.33 34 Moreover, increased expression of COX-2 and mPGES-1 has been demonstrated in response to non-inflammatory stimuli, such as mechanical stretch and mechanical stress in certain cells.8 35 36 The observed enhanced expression of COX-1 and COX-2 in muscle tissue from patients with myositis is in accordance with previously published data demonstrating increased levels of COX-1 and COX-2 mRNA in inflammatory cells and vessels in muscle tissue from patients with myositis.13 Notably there have been no clinical trials addressing specifically the impact of selective or non-selective NSAIDs in patients with myositis. In addition, mPGES-1 expression was enhanced in myositis muscle tissue, as an increased number of positively stained scattered mononuclear cells and as positively stained cells within inflammatory infiltrates compared to low constitutive expression in healthy muscle tissue. Using double staining, we identified CD68 and CD163-positive macrophages as the major cell types expressing mPGES-1. IL1β and TNF are strongly expressed in myositis skeletal muscle tissue2 3 and known to maintain mPGES-1 expression in macrophages, consequently contributing to the enhanced release of PGE2 and inflammation in the muscle tissue.

e identified CD68 and CD163-positive macrophages as the major cell types expressing mPGES-1. IL1β and TNF are strongly expressed in myositis skeletal muscle tissue2 3 and known to maintain mPGES-1 expression in macrophages, consequently contributing to the enhanced release of PGE2 and inflammation in the muscle tissue. We also examined the effects of conventional immunosuppressive treatment on muscle functional activity, muscle histopathology and expression of mPGES-1/COX in muscle tissues. Treatment resulted in significant improved muscle function in the majority of the patients. However, most patients still had impaired muscle function at the time of the second biopsy. Significant reduction in the total number of macrophages (CD68 positive cells) confirmed anti-inflammatory effects of the treatment. After treatment the COX-2 positive cells in muscle tissue were significantly decreased, while the expression of mPGES-1, cPGES or COX-1 was not suppressed. Downregulation of COX-2 expression is one of the expected anti-inflammatory effects of GC and is associated with suppression of PGE2 biosynthesis. However, there was no correlation between COX-2 expression in muscle tissue and FI before or after treatment, probably due to the small number of observations.

We also examined the effects of conventional immunosuppressive treatment on muscle functional activity, muscle histopathology and expression of mPGES-1/COX in muscle tissues. Treatment resulted in significant improved muscle function in the majority of the patients. However, most patients still had impaired muscle function at the time of the second biopsy. Significant reduction in the total number of macrophages (CD68 positive cells) confirmed anti-inflammatory effects of the treatment. After treatment the COX-2 positive cells in muscle tissue were significantly decreased, while the expression of mPGES-1, cPGES or COX-1 was not suppressed. Downregulation of COX-2 expression is one of the expected anti-inflammatory effects of GC and is associated with suppression of PGE2 biosynthesis. However, there was no correlation between COX-2 expression in muscle tissue and FI before or after treatment, probably due to the small number of observations. The reduced number of COX-2 positive cells could be explained by GC dependent downregulation of COX-2 expression and/or a significant reduction in total number of macrophages (CD68 positive cells). By contrast, resident tissue macrophages (CD163-positive cells) did not decrease significantly after treatment. It is tempting to speculate that the population of CD68 cells that decreased as a result of the treatment constitutes cells that did not express CD163 or mPGES-1, as these molecules were not changed after treatment. A similar relative persistence of synovial CD163-positive resident tissue macrophages compared to infiltrating macrophages during anti-TNF treatment has been demonstrated in chronic autoimmune arthritis.37 In addition, immunosuppressive treatment did not affect the number of T lymphocytes in myositis muscle providing a basis for persisting immune reaction that targets muscle fibres.

t tissue macrophages compared to infiltrating macrophages during anti-TNF treatment has been demonstrated in chronic autoimmune arthritis.37 In addition, immunosuppressive treatment did not affect the number of T lymphocytes in myositis muscle providing a basis for persisting immune reaction that targets muscle fibres. Interestingly, in some conditions mPGES-1 functionally utilises PGH2 generated by COX-1.34 38 39 The persisting expression of mPGES-1 and COX-1 in inflamed muscle tissue despite treatment might preserve PGE2 production and contribute to chronic muscle inflammation. However, the role of the mPGES-1/COX-1 pathway in overall PGE2 production in muscle tissue remains to be elucidated. Recent data suggest that COX-2–dependent PG synthesis is important for skeletal muscle regeneration.40–42 While COX-2 inhibition reduces inflammation to a large extent due to the suppression of PGE2 formation, it might impede the functional recovery of muscles via the suppression of other PGs. In this context mPGES-1 may constitute a more selective and safe therapeutic target than COX-2. Selective inhibition of mPGES-1 will allow for intact baseline PGE2 production as well as intact production of other PG important for muscle regeneration and, ultimately, constitute a more preferable anti-inflammatory treatment than the currently used systemic GCs.

more selective and safe therapeutic target than COX-2. Selective inhibition of mPGES-1 will allow for intact baseline PGE2 production as well as intact production of other PG important for muscle regeneration and, ultimately, constitute a more preferable anti-inflammatory treatment than the currently used systemic GCs. In conclusion, we have demonstrated a significantly enhanced expression of mPGES-1, COX-2 and COX-1 in patients with polymyositis or dermatomyositis compared to healthy controls, suggesting its role in the pathogenesis of these diseases. Moreover, we have shown for the first time that conventional immunosuppressive treatment led to a significant downregulation of COX-2 in myositis muscle tissue, while the expression of mPGES-1 and COX-1 remained unchanged. This persisting expression of mPGES-1 and COX-1 may have a role in the chronicity of myositis. The authors would like to thank Professor Kristian Borg for providing muscle biopsy samples from healthy individuals, and Drs Maryam Dastmalchi, Mikael Heimbürger and Christina Ståhl-Hallengren for providing patient samples and clinical data from patients. Funding: This study was supported by grants from King Gustaf V 80 Years Foundation, The Swedish Society of Medicine, The Swedish Rheumatism Association, The Swedish Research Council 2005-74X-14045-02A and 2004–5259, Professor Nanna Svartz Foundation, Magnus Bergvall Foundation, Börje Dahlin Foundation, Karolinska Institutet Foundation and Foundation Clas Groschinskys Minnesfond. Competing interests: None.

Funding: This study was supported by grants from King Gustaf V 80 Years Foundation, The Swedish Society of Medicine, The Swedish Rheumatism Association, The Swedish Research Council 2005-74X-14045-02A and 2004–5259, Professor Nanna Svartz Foundation, Magnus Bergvall Foundation, Börje Dahlin Foundation, Karolinska Institutet Foundation and Foundation Clas Groschinskys Minnesfond. Competing interests: None. Ethics approval: The approval was granted by the Ethics Committee at the Karolinska University Hospital, Stockholm and all patients and controls gave their informed consent to participate in the study.

Interleukin 27 is a member of the IL6/IL12 family and is composed of a p28 subunit and Epstein-Barr virus-induced gene 3, polypeptides structurally related to p35 and p40 of IL12, respectively.1 IL27 is produced by activated antigen-presenting cells and induces proliferation of and T bet expression in naïve CD4+ T cells.1 2 WSX-1, which was cloned as a homologue of gp130 of the IL6 receptor,3 constitutes a functional signal-transducting receptor for IL27 with gp130.4 WSX-1 is highly expressed in CD4+ T cells as well as in natural killer (NK)/natural killer T (NKT) cells and macrophages.3 5 6 Analysis of mice deficient for WSX-1 infected with Leishmania major revealed the critical role of WSX-1 in the initial mounting of proper Th1 responses.6 In infection with Trichuris muris, a nematode whose clearance depends on Th2 responses, WSX-1-deficient mice showed impaired Th1 responses with augmented Th2 responses resulting in more efficient expulsion of the worms than that in wild type (WT) mice, confirming its role for Th1 development.7 8

proper Th1 responses.6 In infection with Trichuris muris, a nematode whose clearance depends on Th2 responses, WSX-1-deficient mice showed impaired Th1 responses with augmented Th2 responses resulting in more efficient expulsion of the worms than that in wild type (WT) mice, confirming its role for Th1 development.7 8 Recent lines of evidence, however, have shown a distinct role for WSX-1 and its ligand, IL27, as an attenuator of inflammatory responses. In Toxoplasma gondii or Trypanosoma cruzi infection, CD4+ T cells as well as NKT cells and macrophages in WSX-1-deficient mice overproduced several inflammatory cytokines, resulting in devastating inflammation in the liver and other organs.9 10 The suppressive role of WSX-1 was also observed in various experimental settings such as concanavalin A (Con A)-induced hepatitis, Mycobacterium tuberculosis infection, an allergic asthma model and experimental autoimmune encephalomyelitis.11–15 These data clearly demonstrated that IL27/WSX-1 plays an inhibitory role by regulating cell activation and cytokine production.16

ous experimental settings such as concanavalin A (Con A)-induced hepatitis, Mycobacterium tuberculosis infection, an allergic asthma model and experimental autoimmune encephalomyelitis.11–15 These data clearly demonstrated that IL27/WSX-1 plays an inhibitory role by regulating cell activation and cytokine production.16 Systemic lupus erythaematosus (SLE) is a multi-system disease that is caused by tissue damage resulting from autoantibody and complement-fixing immune complex deposition. Lupus nephritis manifests considerable heterogeneity in phenotype and histology. In particular, diffuse proliferative glomerulonephritis (DPGN) and membranous glomerulonephritis (MGN) represent two histological forms that are polar opposites.17 18 The pathogenesis of DPGN is associated with predominance of Th1 cytokines,19 while that of MGN with predominantly Th2 cytokine response.20 MRL/lpr mice develop a systemic autoimmune disease, which is reminiscent of SLE in humans. In MRL/lpr mice, Fas-mediated apoptosis of activated lymphocytes was severely impaired, and T cell-dependent production of autoantibodies results in immune complex-mediated glomerulonephritis and vasculitis.21 22 Kidney disease in MRL/lpr mouse is a particularly suitable model of DPGN. Intriguingly, disruption of the WSX-1 gene changed the pathophysiology of glomerulonephritis developing in MRL/lpr (WT) mice. WSX-1–/– MRL/lpr mice developed a disease resembling human MGN with augmented Th2 responses, confirming that the Th1/Th2 cytokine balance is a key to the pathogenesis of differential types of glomerulonephritis.23 In this study, we generated lines of WSX-1 transgenic MRL/lpr mice to further investigate roles of IL27/WSX-1 in the development of autoimmune disorders in MRL/lpr mice.

with augmented Th2 responses, confirming that the Th1/Th2 cytokine balance is a key to the pathogenesis of differential types of glomerulonephritis.23 In this study, we generated lines of WSX-1 transgenic MRL/lpr mice to further investigate roles of IL27/WSX-1 in the development of autoimmune disorders in MRL/lpr mice. METHODS Generation of WSX-1 transgenic MRL/lpr mice WSX-1 transgenic mice in the MRL/lpr background were produced by crossing WSX-1 transgenic BALB/c mice24 into the MRL/lpr background more than six times (continual backcrossing: 98.44% in MRL/lpr background). Genotyping for lpr alleles was performed by PCR as described previously.23 We generated two strains of WSX-1 transgenic mice in the MRL/lpr background (transgenic high (TgH) and low (TgL)) depending on different expression levels of WSX-1. Female mice from the same litters were used in the present study. Mice were maintained in the Laboratory of Animal Experiments of Kyushu University. All experiments were approved by the Institutional Animal Research Committee of Kyushu University and conformed to the animal care guidelines of the American Physiologic Society.

mice from the same litters were used in the present study. Mice were maintained in the Laboratory of Animal Experiments of Kyushu University. All experiments were approved by the Institutional Animal Research Committee of Kyushu University and conformed to the animal care guidelines of the American Physiologic Society. Western blotting We evaluated the production of WSX-1 protein in the transgenic mice using anti-T cell lymphocyte cytokine receptor (TCCR) (WSX-1) antibody (Abcam, Cambridge, Massachusetts, USA), anti-β-actin antibody (Sigma, St Louis, Missouri, USA), and anti-mouse IgG-horseradish peroxidase (HRP) antibodies (Amersham Biosciences, Piscataway, New Jersey, USA). They were visualised with an electrochemical luminescence (ECL) detection system (Amersham Biosciences).

ntibody (Abcam, Cambridge, Massachusetts, USA), anti-β-actin antibody (Sigma, St Louis, Missouri, USA), and anti-mouse IgG-horseradish peroxidase (HRP) antibodies (Amersham Biosciences, Piscataway, New Jersey, USA). They were visualised with an electrochemical luminescence (ECL) detection system (Amersham Biosciences). Laboratory assessments For serum chemistry, total protein, blood urea nitrogen (BUN) and creatinine (Cr)8 levels were assessed in the sera from 10 mice in each group at 24 weeks. Urinary protein:urinary Cr ratios were also determined. Anti-nuclear antibodies (ANA) were detected by indirect immunofluorescence using HEp-2 substrate slides (Orgentec, Mainz, Germany) with fluorescein isothiocyanate-conjugated AffiniPure donkey anti-mouse IgG (Jackson ImmunoResearch, West Grove, Pennsylvania, USA).25 26 Serum anti-double-stranded DNA (anti-dsDNA) antibodies (Abs) were analysed by ELISA (Shibayagi, Gunma, Japan). For serum Ig, determination ELISA was performed using the following antibodies: rat anti-mouse IgG1 (Zymed Laboratories, San Francisco, California, USA), rat anti-mouse IgG1-HRP (BioSource International, Camarillo, California, USA), goat anti-mouse IgG2a (Bethyl Laboratories, Montgomery, Alabama, USA), rabbit anti-mouse IgG2a-HRP (Cappel Lab, Durham, North Carolina, USA) and goat anti-mouse IgE Ab (Bethyl Laboratories).

med Laboratories, San Francisco, California, USA), rat anti-mouse IgG1-HRP (BioSource International, Camarillo, California, USA), goat anti-mouse IgG2a (Bethyl Laboratories, Montgomery, Alabama, USA), rabbit anti-mouse IgG2a-HRP (Cappel Lab, Durham, North Carolina, USA) and goat anti-mouse IgE Ab (Bethyl Laboratories). Histopathological and immunohistopathological studies of kidneys The severity of glomerulonephritis was evaluated as described previously.23 For immunohistochemical staining, kidneys were snap frozen in optimal cutting temperature compound (Sakura, Osaka, Japan). To detect immune complex (IC) deposits, cryostat sections (2 µm) were fixed in chilled acetone and stained with fluorescein isothiocyanate (FITC)-conjugated goat polyclonal anti-mouse IgG Abs (Organon Teknika, Scarborough, Maine, USA), a FITC-conjugated goat anti-mouse IgG1 Ab and a FITC-conjugated goat anti-mouse IgG2a Ab (Southern Biotechnology Associates, Birmingham, Alabama, USA). For negative controls, sections were treated with normal goat IgG (Santa Cruz Biotechnology, Santa Cruz, California, USA). The fluorescence strength was analysed using scion image (Scion Cooperation, Frederick, Maryland, USA).27

ted goat anti-mouse IgG2a Ab (Southern Biotechnology Associates, Birmingham, Alabama, USA). For negative controls, sections were treated with normal goat IgG (Santa Cruz Biotechnology, Santa Cruz, California, USA). The fluorescence strength was analysed using scion image (Scion Cooperation, Frederick, Maryland, USA).27 Real-time quantitative PCR and TaqMan primers and probes Expression levels of interferon (IFN)γ and IL4 in CD4+ T cells were determined using TaqMan PCR and an ABI prism 7700 sequence detection system (Applied Biosystems Japan, Tokyo, Japan). The relative expression of each mRNA was determined and normalised to the expression of the internal housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Primer and probe sequences are described previously.23

an PCR and an ABI prism 7700 sequence detection system (Applied Biosystems Japan, Tokyo, Japan). The relative expression of each mRNA was determined and normalised to the expression of the internal housekeeping gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH). Primer and probe sequences are described previously.23 Activation of CD4+ T cells CD4+ T cells were purified from splenic extracts using magnetic beads (Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). Purified CD4+ T cells were activated with plate-bound anti-CD3 Ab (1 μg/ml) plus soluble anti-CD28 Ab (1 μg/ml) (BD Biosciences, San Jose, California, USA) for 2 days, transferred to a new plate without antibodies and additionally cultured for 5 days as a total of 7 days either in the presence or absence of IL27. Culture supernatants containing recombinant murine IL27 were prepared as described previously.16 The cells were then washed and restimulated either with anti-CD3 Ab plus anti-CD28 Ab for cytokine production or with IL27 for signal transducer and activator of transcription (STAT) activation. Anti-STAT1, anti-phosphotyrosine (pY)-STAT1 and anti-pY-STAT3 Abs were purchased from New England BioLabs (Beverly, Massachusetts, USA). Anti-STAT3 Ab was purchased from Santa Cruz Biotechnology.

us anti-CD28 Ab for cytokine production or with IL27 for signal transducer and activator of transcription (STAT) activation. Anti-STAT1, anti-phosphotyrosine (pY)-STAT1 and anti-pY-STAT3 Abs were purchased from New England BioLabs (Beverly, Massachusetts, USA). Anti-STAT3 Ab was purchased from Santa Cruz Biotechnology. Measurement of cytokines Cytokines in culture supernatants of CD4+ T cells were analysed using a micro bead-based ELISA system (Multiplex Antibody Bead Kits, Biosource) according to the manufacturer’s directions with Luminex 100 (Luminex, Austin, Texas, USA). Cytokines in the sera were measured by ELISA kits (Genzyme, R&D Systems, Abingdon, UK and eBioscience, Los Angeles, California, USA) for detection of IFNγ, IL4, IL17A and IL2. Statistical analyses For survival of mice, Kaplan–Meier analysis was carried out using the Statview software package (SAS Institute Japan, Tokyo, Japan). Other quantitative data were expressed as the mean (SD). The Mann–Whitney U rank sum test was performed to analyse the difference between two groups, while for individual comparisons among the three groups the Kruskal–Wallis test, followed by the Scheffe test was performed. All tests were two-tailed. A p value of less than 0.05 was considered statistically significant.

(SD). The Mann–Whitney U rank sum test was performed to analyse the difference between two groups, while for individual comparisons among the three groups the Kruskal–Wallis test, followed by the Scheffe test was performed. All tests were two-tailed. A p value of less than 0.05 was considered statistically significant. RESULTS Prolonged survival in TgH mice, not in TgL mice The MRL/lpr mice develop a rapid and fluminant autoimmune nephritis with 50% mortality at 6 months of age.28 In this study MRL/lpr mice (WT mice) died after birth at a rate described above. Although TgL mice died at the similar rate as WT mice, TgH mice showed significantly extended survival rates (0% dead at 24 weeks after birth; p<0.001 over WT mice) (fig 1A).

nt autoimmune nephritis with 50% mortality at 6 months of age.28 In this study MRL/lpr mice (WT mice) died after birth at a rate described above. Although TgL mice died at the similar rate as WT mice, TgH mice showed significantly extended survival rates (0% dead at 24 weeks after birth; p<0.001 over WT mice) (fig 1A). Figure 1 Prolonged survival of MRL/lpr mice with high expression of WSX-1. A. The cumulative survivals of wild type (WT), transgenic low (TgL) and high (TgH) mice were monitored weekly (n = 20 per group). *p<0.05 by the Kaplan–Meier method. There was no difference in the survival rate between WT and TgL mice. B. Western blot analysis of the transgenic expression of WSX-1 in CD4+ T cells from WT mice and two Tg-positive MRL/lpr mouse lines, TgL and TgH. The lysates of purified splenic CD4+ T cells from 16-week-old WT, TgL and TgH mice were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) under reducing condition and transferred to membranes. Expression levels for β-actin protein was visualised for each sample as a loading control. The transgene products demonstrate robust, moderate and faint expression of WSX-1 in TgH, TgL and WT mice T cells, respectively. Improvement of clinical features and parameters in TgH mice Given the striking improvement of survival in TgH mice, clinical features and serum chemistry parameters of 24-week-old mice were examined. TgH mice showed significantly lower BUN, compared with WT and TgL mice (table 1). Additionally, the urinary protein:creatinine ratio was also lower in TgH mice than in TgL or WT mice. These data indicated amelioration of kidney function in TgH mice. The parameters of 36-week-old TgH mice were comparable with those of 24-week-old TgH mice (data not shown). Of note, splenomegaly and lymphadenopathy were also significantly reduced in TgH mice.

ne ratio was also lower in TgH mice than in TgL or WT mice. These data indicated amelioration of kidney function in TgH mice. The parameters of 36-week-old TgH mice were comparable with those of 24-week-old TgH mice (data not shown). Of note, splenomegaly and lymphadenopathy were also significantly reduced in TgH mice. Table 1 Clinical manifestations and serum chemistry in female wild type (WT), transgenic low (TgL) and high (TgH) mice MRL/lpr 24W WT TgL TgH Body weight, g 49.0 (6.10) 43.1 (4.40) 40.4 (3.00) Spleen weight, g 1.31 (0.56) 0.90 (0.21) 0.33 (0.07)* Total lymph node weight, g 6.14 (1.71) 3.43 (3.31) 1.51 (0.53)* Urinary protein:creatinine ratio 17.4 (5.80) 15.98 (10.6) 5.24 (2.22)* Serum protein, g/dl 6.83 (0.67) 6.93 (0.66) 7.68 (0.77) Blood urea nitrogen, mg/dl 57.2 (24.9) 57.9 (19.2) 28.4 (6.00)* Serum creatinine, mg/dl 0.27 (0.12) 0.20 (0.09) 0.36 (0.14) Values are mean (SD). *p<0.05 vs WT Amelioration of glomerulonephritis in TgH mice Since kidney dysfunction is the primary cause of death in MRL/lpr mice, histopathological examination of kidneys from the three groups of 24-week-old mice was performed. Typical histological features of DPGN were observed in WT and also in TgL mice, including inflammatory cell infiltration, glomerular sclerosis, mesangial proliferation and crescent formation (fig 2A,B). By striking contrast, TgH mice showed drastic attenuation of inflammatory and proliferative changes (fig 2C). The score of glomerular proliferative activity of TgH was significantly decreased compared to that of WT (fig 2M. p = 0.0074 TgH vs WT).

lar sclerosis, mesangial proliferation and crescent formation (fig 2A,B). By striking contrast, TgH mice showed drastic attenuation of inflammatory and proliferative changes (fig 2C). The score of glomerular proliferative activity of TgH was significantly decreased compared to that of WT (fig 2M. p = 0.0074 TgH vs WT). Figure 2 Amelioration of glomerulonephritis and decreased Ig deposition in transgenic high (TgH) mice. The kidneys were fixed in 10% formalin for 24 h at 4°C. Paraffin sections (4 µm) were stained either with H&E, periodic acid Schiff stain, or periodic acid-methenamine silver (PAM). Microscopic examination of the kidney glomerulus of a 24-week-old wild type (WT) mouse (A), transgenic low (TgL) mouse (B) and TgH mouse (C) (periodic acid Schiff staining; original magnification, ×400). IgG (D, E and F), IgG1 (G, H and I) and IgG2a (J, K and L) deposits in the glomeruli of a 24-week-old WT, TgL and TgH mice were visualised using immunofluorescent anti-Ig staining (×400) and quantitative analysed by image software. WT: IgG (83.8 (7.25)), IgG1 (68.4 (6.4)) and IgG2a (76.8 (4.9)), TgL: IgG (81.0 (8.48)), IgG1 (71.4 (7.3)) and IgG2a (73.0 (4.6)), TgH:IgG (25.2 (9.65)), IgG1 (22.8 (6.45)), IgG2a (19.6 (3.2)). These representative data obtained from 30 glomerular cross sections per kidney (six mice per group) with similar staining patterns. M. The severity of glomerulonephritis was evaluated by the score of glomerular proliferative activity. The scores of glomerular proliferative activity of 24-week-old WT, TgL and TgH mice were 2.38 (0.356), 1.92 (0.564) and 0.74 (0.378), respectively. Deposition of immunoglobulin is one of the hallmarks of glomerulonephritis.15 We then performed immunofluorescent staining of the kidneys to detect Ig in glomeruli in the three groups of mice. An intense IgG deposition was detected in mesangial lesions and along the capillary walls of glomeruli in WT and TgL mice. The isotypes of the deposited Ig were mainly IgG2a and IgG1 in part (fig 2D,E,G,H,J,K). By contrast, IgG deposition was hardly observed in TgH mice, and the deposition of IgG1 and IgG2a was remarkably decreased (fig 2F,I,L), and the fluorescent strength score was significantly decreased (p = 0.0035 TgH vs WT).

gL mice. The isotypes of the deposited Ig were mainly IgG2a and IgG1 in part (fig 2D,E,G,H,J,K). By contrast, IgG deposition was hardly observed in TgH mice, and the deposition of IgG1 and IgG2a was remarkably decreased (fig 2F,I,L), and the fluorescent strength score was significantly decreased (p = 0.0035 TgH vs WT). Decreased production of ANA, anti-dsDNA antibodies and immunoglobulins To further examine the immunological changes in TgH mice, the level and the nature of serum Ig and autoantibodies were evaluated. Positive staining for ANA was detected in 16-week-old WT TgL mouse sera at a 1:200 dilution, as determined by indirect immunofluorescence (n = 10 per group), whereas sera from 16-week-old TgH mice were negative for ANA (n = 8) (fig 3A). The sera from 24- and 36-week-old TgH mice were also negative (data not shown). Production of anti-dsDNA Abs was significantly lower in TgH mice than those of WT and TgL mice (fig 3B). While there were no significant differences in the levels of IgG1 and IgE, the levels of total IgG and IgG2a were significantly lower in TgH mice than in other groups (fig 3C).

e were also negative (data not shown). Production of anti-dsDNA Abs was significantly lower in TgH mice than those of WT and TgL mice (fig 3B). While there were no significant differences in the levels of IgG1 and IgE, the levels of total IgG and IgG2a were significantly lower in TgH mice than in other groups (fig 3C). Figure 3 Decreased autoreactive immune responses in transgenic high (TgH) mice. A. Sera collected from 16-week-old wild type (WT), transgenic low (TgL) and TgH mice were examined for anti-nuclear antibodies (ANA) as described in Materials and methods. The same sera in (A) were measured for anti-dsDNA antibodies (B) or for Ig levels (C). D. The expression levels of interferon (IFN)γ and interleukin (IL)4 mRNA in CD4+ T cells and those of IL12 and IL10 mRNA in B220+CD3+ cell-depleted splenocytes were examined by quantitative PCR. Data shown are mean (SD) of 10 mice per group. *p<0.05 by unpaired Student t test. Reduced cytokine production by CD4+ T cells from TgH mice Expression of IFNγ and IL4 in splenic CD4+ T cells decreased in a manner dependent on the expression levels of WSX-1 (fig 3D). Similarly, expression of IL12b and IL10 in spleen cells depleted of B220+CD3+ cell was also decreased in WSX-1 transgenic mice.

t test. Reduced cytokine production by CD4+ T cells from TgH mice Expression of IFNγ and IL4 in splenic CD4+ T cells decreased in a manner dependent on the expression levels of WSX-1 (fig 3D). Similarly, expression of IL12b and IL10 in spleen cells depleted of B220+CD3+ cell was also decreased in WSX-1 transgenic mice. Decrease in CD3+B220+CD4–CD8– T cell, and increase in CD4+ and CD8+ T cell in the WSX-1 TgH MRL/lpr mouse Given the significant improvement of splenomegaly and lymphadenopathy in TgH mice (table 1), cellular composition in the spleen was analysed. The percentage of CD3+B220+CD4–CD8– T cells in TgH mouse was greatly diminished over WT mice and in TgL mice (fig 4A, and data not shown). Concomitantly, percentages of CD4+ and CD8+ T cells increased compared with WT mice.

galy and lymphadenopathy in TgH mice (table 1), cellular composition in the spleen was analysed. The percentage of CD3+B220+CD4–CD8– T cells in TgH mouse was greatly diminished over WT mice and in TgL mice (fig 4A, and data not shown). Concomitantly, percentages of CD4+ and CD8+ T cells increased compared with WT mice. Figure 4 Decrease in CD3+B220+CD4–CD8– T cell and upregulated expression of activation markers on transgenic high (TgH) CD4+ T cells. A. Spleens were removed from 16-week-old wild type (WT) and TgH mice, and single cell suspensions were stained for the expression of surface markers, followed by flow cytometry. Numbers are the percentage of CD3+B220+CD4–CD8– T cells, CD8+ T cells and CD4+ T cells against total cell populations. B. Surface expression levels of CD69, CD25, CD62L and CD44 were analysed by flow cytometry. Numbers are the percentage of the total CD4+ T cells. Expression of activation markers in CD4+ T cells in WSX-1 TgH MRL/lpr mice To clarify the activation status of CD4+ T cells in the mice, expression of several cell surface activation markers was evaluated (fig 4B). In WT and TgH mice, most of the CD4+ T cells expressed CD69 on their cell surface, due presumably to in their possible autoreactivity. More than 10% of CD4+ T cells were positive for CD25 expression in WT and TgH mice. While some 40% of CD4+ T cells did not express CD62L in WT mice, approximately 80% of CD4+ T cells did not express CD62L. Interestingly, however, less CD4+ T cells expressed CD44 in TgH mice than in WT mice. These data demonstrated that while most CD4+ T cells were activated by CD69 expression in WT and TgH mice, more cells in TgH mice showed the activated phenotype by CD62L expression. However, because much fewer cells expressed CD44 in TgH than in WT mice, the CD4+ T cells in TgH mice appeared to be in an activation status different from that in WT mice. Although the percentage of CD25+ cells in the CD4+ T cell population was higher in TgH mice than in WT mice, there was no significant difference in the expression levels of FoxP3 in CD4+ T cells between these mice (data not shown).

cells in TgH mice appeared to be in an activation status different from that in WT mice. Although the percentage of CD25+ cells in the CD4+ T cell population was higher in TgH mice than in WT mice, there was no significant difference in the expression levels of FoxP3 in CD4+ T cells between these mice (data not shown). CD4+ T cells from WSX-1 TgH mice were more subject to the IL27-mediated suppression of cytokine production To examine the effect of IL27 on lymphocyte activity in the mice, cytokine production by in vitro activated CD4+ T cells either in the presence or absence of IL27 was examined (fig 4A). CD4+ T cells from TgH mice produced more IFNγ and less IL4 than those from WT mice. Nonetheless, CD4+ T cells from TgH mice were more sensitive to IL27-mediated suppression of cytokine production. Although WT and TgH CD4+ T cells were subject to IL27-mediated suppression of IFNγ and IL4 production, the suppressive effect was prominent in TgH cells over WT cells. Production of IFNγ and IL4 was strikingly suppressed to barely detectable levels in TgH CD4+ T cells. IL17 production by TgH CD4+ T cells was lower than WT cells even without IL27 addition, and was suppressed to a barely detectable level in TgH cells. Interestingly, IL2 production was not affected by WSX-1 overexpression and CD4+ cells from WT and TgH mice were similarly subject to IL27-mediated suppression.

+ T cells. IL17 production by TgH CD4+ T cells was lower than WT cells even without IL27 addition, and was suppressed to a barely detectable level in TgH cells. Interestingly, IL2 production was not affected by WSX-1 overexpression and CD4+ cells from WT and TgH mice were similarly subject to IL27-mediated suppression. Downstream of the IL27R (WSX-1 plus gp130), STAT1 and STAT3 are activated.16 24 When CD4+ T cells were isolated and immediately examined for STAT1/3 phosphorylation, phosphorylation of STAT1 and STAT3 was apparent in TgH CD4+ T cells as compared with WT cells (fig 5B). When these cells were stimulated with IL27 for 1 h, further phosphorylation of STAT1 and STAT3 was observed in WT and Tg cells. While the levels of STAT1 phosphorylation were comparable between WT and Tg cells, STAT3 phosphorylation was higher in Tg cells than in WT cells.

4+ T cells as compared with WT cells (fig 5B). When these cells were stimulated with IL27 for 1 h, further phosphorylation of STAT1 and STAT3 was observed in WT and Tg cells. While the levels of STAT1 phosphorylation were comparable between WT and Tg cells, STAT3 phosphorylation was higher in Tg cells than in WT cells. Figure 5 IL27 suppression of cytokine production by activated CD4+ T cells. A. CD4+ T cells from wildtype (WT) and transgenic high (TgH) mice were stimulated with plate-bound anti-CD3 antibody plus soluble anti-CD28 antibody (1 μg/ml) in the absence or presence of interleukin (IL)27 for 24 h. Culture supernatants were measured for the production of respective cytokines. Data shown are mean (SD) of 10 mice per group. *p<0.05 by unpaired Student t test. B. CD4+ T cells from WT and TgH mice were stimulated as in Materials and methods. Phosphorylation of signal transducer and activator of transcription (STAT)1 or STAT3 in whole cell lysates in WT and TgH mice was analysed with anti-phosphotyrosine (α-pY) -STAT1 or anti-pY-STAT3 antibodies, respectively. The filter was stripped and re-probed with anti-STAT1 or anti-STAT3 antibodies to ensure the same amounts of samples were loaded. Experiments were repeated three times with similar results. DISCUSSION In this study we generated lines of Tg mice that overexpressed WSX-1 in T cells to examine the impact of WSX-1 overexpression on pathophysiology and autoimmune status of MRL/lpr mice. We demonstrated that overexpression of WSX-1 suppressed the development of autoimmune nephritis in WT mice, and ANA, the values of anti-dsDNA Ab, serum Ig and expression of various cytokines, significantly decreased in of the Tg mice. While CD4+ T cells in TgH mice were in a distinct activation status from those in WT mice, these cells were more subject to the effects of IL27 in vitro. These results strongly suggested that increased expression of WSX-1 suppressed the autoimmune reaction and the subsequent glomerulonephritis in WT mice.

the Tg mice. While CD4+ T cells in TgH mice were in a distinct activation status from those in WT mice, these cells were more subject to the effects of IL27 in vitro. These results strongly suggested that increased expression of WSX-1 suppressed the autoimmune reaction and the subsequent glomerulonephritis in WT mice. We originally demonstrated the pivotal role of WSX-1 in the initial mounting of Th1 differentiation via T bet induction,2 and in mice deficient in the WSX-1 gene, proper Th1 differentiation was impaired with Th2 skewing during protozoan infection.6 In line with these findings, we revealed that disruption of the WSX-1 gene drastically changed the histological features of glomerulonephritis developing in MRL/lpr mice from DPGN to MGN accompanied by impaired IFNγ production with predominance Th2-dependent IgG1 deposition and increased levels of IgG1 and IgE in the sera.23 T cells in WSX-1–/– MRL/lpr mice displayed spontaneous skewing of autoimmune responses toward Th2 type. Thus, our previous reports suggested that immune status, or more specifically, the balance between Th1 and Th2 responses, is a key determinant for the pathogenesis of the glomerulonephritis. Counterintutively, transgenic overexpression of WSX-1 gene resulted in amelioration of glomerulonephritis in MRL/lpr mice. It would be reasonable to assume this was the result of the suppressive effects of IL27 since the cytokine expression by CD4+ T cells as well as autoimmune reaction was suppressed in TgH mice.

tis. Counterintutively, transgenic overexpression of WSX-1 gene resulted in amelioration of glomerulonephritis in MRL/lpr mice. It would be reasonable to assume this was the result of the suppressive effects of IL27 since the cytokine expression by CD4+ T cells as well as autoimmune reaction was suppressed in TgH mice. CD4+ T cells from TgH mice activated in vitro produced more IFNγ and less IL4 than those from WT mice and showed Th1 phenotyping, which was in line with the Th1-promoting function of IL27. However, IL27 addition strikingly suppressed production of IFNγ and IL4 to barely detectable levels in Tg CD4+ T cells (fig 5A). These results confirmed the suppressive effect of IL27 on cytokine production and also revealed that the suppressive effect was much higher for CD4+ T cells of TgH than for those of WT mice. Such suppression of cellular response was largely consistent with the diminished pathophysiology of glomerulonephritis in TgH mice. We have recently reported that IL27 exerts its suppressive effects preferentially on activated CD4+ T cells, and that STAT3 activation in response to IL27 stimulation of activated T cells is, at least partially, responsible for the IL27-mediated suppression of cytokine production.16 This is quite consistent with our present finding that TgH mice CD4+ T cells were more sensitive to IL27 stimulation by STAT3 activation (fig 5B). Ohwaki et al reported the involvement of STAT1 in IL27-mediated suppression of IL2 production by naïve T cells,29 although Villarino et al reported that IL27-mediated suppression in activated T cells is independent of STAT1.30 The discrepancy may be ascribed to the activation status of the cells.

3 activation (fig 5B). Ohwaki et al reported the involvement of STAT1 in IL27-mediated suppression of IL2 production by naïve T cells,29 although Villarino et al reported that IL27-mediated suppression in activated T cells is independent of STAT1.30 The discrepancy may be ascribed to the activation status of the cells. In summary, we demonstrated that WSX-1 overexpression in the MRL/lpr background rendered the autoimmune prone mice protected from the development of autoimmune disease. Further elucidation of the molecular mechanisms underlying the IL27-mediated cytokine suppression and detailed dissection of the situations where IL27 differentially exerts its two roles will no doubt help development of new therapies against various diseases by suppressing excess of cell responses. The authors thank Ms H Noguchi and Ms Y Furukawa for their excellent technical assistance. We are grateful to Dr M Hanada for encouragement and support of our study. Funding: This study was supported in part by grants from the Ministry of Education, Science, Technology, Sports and Culture of Japan (to AY, SH and HY), from Japan Research Foundation for Clinical Pharmacology (to HY), from the Sumitomo Foundation, Grant for Basic Science Research Projects (to HY), from the Naito Foundation (to HY), and from the Takeda Science Foundation (to HY). This work was also supported by the president’s expenditure (research project expenditure) of Saga University (to HY). Competing interests: None declared.

Since the discovery of cortisone and its first use in patients with rheumatoid arthritis (RA)1 glucocorticoids have been extensively used to suppress synovial inflammation. However, in patients with established synovitis, glucocorticoids such as cortisol (hydrocortisone), prednisone and prednisolone do not cause permanent resolution of inflammation and long-term use has adverse effects on bone, skin and fat tissue.2 3 Endogenous glucocorticoids also have a role in suppressing disease activity in RA. Early morning stiffness is attributed to the nocturnal decrease in circulating cortisol levels. Administration of metyrapone to reduce endogenous corticosteroid production increases disease activity in RA.4 It is unclear, however, whether endogenous corticosteroid action contributes to susceptibility to, or severity of, RA. Subtle abnormalities of the hypothalamic-pituitary-adrenal axis have been seen in glucocorticoid-naive patients with RA5–7 but their origin remains unclear.8

oduction increases disease activity in RA.4 It is unclear, however, whether endogenous corticosteroid action contributes to susceptibility to, or severity of, RA. Subtle abnormalities of the hypothalamic-pituitary-adrenal axis have been seen in glucocorticoid-naive patients with RA5–7 but their origin remains unclear.8 We have previously hypothesised that periarticular osteopenia in RA is partly due to excessive local glucocorticoid activation through the 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) enzyme.9 This enzyme converts inactive steroids (cortisone and prednisone) to their active counterparts (cortisol and prednisolone).10 Although 11β-HSD1 is bidirectional, its predominant action in vivo is conversion of inactive to active glucocorticoids. Hepatic 11β-HSD1 is essential for activation of oral cortisone/prednisone—patients who lack this enzyme are unresponsive to cortisone and prednisone but respond to hydrocortisone and prednisolone.11 We have reported that synovial fibroblasts express 11β-HSD1 in vitro and in vivo.12 In osteoblasts and synovial cells 11β-HSD1 activity is upregulated by proinflammatory cytokines.9 12 This suggested that 11β-HSD1 might generate high levels of glucocorticoids within the joint and that this might contribute to periarticular osteopenia.

that synovial fibroblasts express 11β-HSD1 in vitro and in vivo.12 In osteoblasts and synovial cells 11β-HSD1 activity is upregulated by proinflammatory cytokines.9 12 This suggested that 11β-HSD1 might generate high levels of glucocorticoids within the joint and that this might contribute to periarticular osteopenia. By contrast, a related enzyme 11β-HSD2 solely inactivates steroids. This enzyme is expressed in mineralocorticoid target tissues, various developmental tissues and some tumours.13–15 Recent studies have reported expression of 11β-HSD2 in peripheral blood mononuclear cells (PBMCs) and synovium of patients with RA.16–18 We therefore examined glucocorticoid metabolism and function in synovial tissue from patients with RA using specific enzyme assays and inhibitors. In addition, we examined glucocorticoid concentrations in synovial fluid and compared the systemic metabolism of glucocorticoids in patients with RA and non-inflammatory joint conditions. PATIENTS AND METHODS Patients Biopsy specimens of matched synovium and skin were obtained during hip, knee or elbow arthroplasty from consenting patients who fulfilled the American College of Rheumatology criteria for RA and OA. Table 1 gives clinical details of the patients.

By contrast, a related enzyme 11β-HSD2 solely inactivates steroids. This enzyme is expressed in mineralocorticoid target tissues, various developmental tissues and some tumours.13–15 Recent studies have reported expression of 11β-HSD2 in peripheral blood mononuclear cells (PBMCs) and synovium of patients with RA.16–18 We therefore examined glucocorticoid metabolism and function in synovial tissue from patients with RA using specific enzyme assays and inhibitors. In addition, we examined glucocorticoid concentrations in synovial fluid and compared the systemic metabolism of glucocorticoids in patients with RA and non-inflammatory joint conditions. PATIENTS AND METHODS Patients Biopsy specimens of matched synovium and skin were obtained during hip, knee or elbow arthroplasty from consenting patients who fulfilled the American College of Rheumatology criteria for RA and OA. Table 1 gives clinical details of the patients. Table 1 Clinical characteristics of subjects for synovial tissue corticosteroid metabolism studies Patients Age (years)Mean (SD) F/M (n) Site of operation (n) Treatment (n) ESR(mm/1st h)Mean (SD) CRP(mg/l)Mean (SD) With RA (n = 12) 62 (10) 11/1 Hip (6) Knee (4) Elbow (2) Methotrexate (3) Prednisolone (3) Anti-TNF (2) Sulfasalazine (1) Hydroxychloroquine (1) Azathioprine (1) 39 (20) 27 (23) With OA (n = 8) 67 (7) 6/2 Hip (7) Knee (1) 14 (11)* *p<0.05 compared with patients with rheumatoid arthritis. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; OA, osteoarthritis; RA, rheumatoid arthritis; TNF, tumour necrosis factor.

Table 1 Clinical characteristics of subjects for synovial tissue corticosteroid metabolism studies Patients Age (years)Mean (SD) F/M (n) Site of operation (n) Treatment (n) ESR(mm/1st h)Mean (SD) CRP(mg/l)Mean (SD) With RA (n = 12) 62 (10) 11/1 Hip (6) Knee (4) Elbow (2) Methotrexate (3) Prednisolone (3) Anti-TNF (2) Sulfasalazine (1) Hydroxychloroquine (1) Azathioprine (1) 39 (20) 27 (23) With OA (n = 8) 67 (7) 6/2 Hip (7) Knee (1) 14 (11)* *p<0.05 compared with patients with rheumatoid arthritis. CRP, C-reactive protein; ESR, erythrocyte sedimentation rate; OA, osteoarthritis; RA, rheumatoid arthritis; TNF, tumour necrosis factor. Synovial tissue was taken on ice and prepared within 2 h by removing adherent non-synovial tissue. Tissue was divided into 100 mg sections for enzyme assay or ELISA. Skin tissue was prepared by removing subcutaneous fat and dividing into 100 mg pieces. Matched synovial fluid and serum samples were obtained from patients with active RA undergoing joint aspiration as part of routine care. Blood was drawn immediately before joint aspiration. Clinical details are given online in supplementary table 1. Urine samples for corticosteroid metabolite analysis were obtained from patients with newly presenting RA or non-inflammatory joint disease (localised OA (n = 5); trigger finger (n = 3); hypermobility (n = 1)). Clinical details are given in supplementary table 2. All studies had ethical approval from the local ethics committee and informed consent was obtained when samples were taken.

Urine samples for corticosteroid metabolite analysis were obtained from patients with newly presenting RA or non-inflammatory joint disease (localised OA (n = 5); trigger finger (n = 3); hypermobility (n = 1)). Clinical details are given in supplementary table 2. All studies had ethical approval from the local ethics committee and informed consent was obtained when samples were taken. 11β-Hydroxysteroid dehydrogenase enzyme assays Synovial or skin tissue (100 mg per assay) was incubated in RPMI-1640 medium containing 1% non-essential amino acids, 1% penicillin/streptomycin, 1% sodium pyruvate, 2 mM glutamine and 20% heat-inactivated fetal calf serum (FCS; Labtech International, Sussex, UK). Cortisol (100 nM; to measure glucocorticoid inactivation) or cortisone (to measure activation) along with tracer amounts of [3H]cortisol (specific activity 78.4 Ci/mmol; NEN Life Science Products, Hounslow, UK) or [3H]cortisone (generated as previously described19) were added and tissue incubated at 37°C for 18 h. Similar assays were performed using [3H]prednisolone or [3H]prednisone.20 Steroids were extracted from medium/tissue using dichloromethane (5–7 ml) and separated by thin-layer chromatography using ethanol:chloroform as the mobile phase. Thin-layer chromatography plates were analysed using a Bioscan imager (Bioscan, Washington DC, USA) and fractional steroid conversion calculated. Results were expressed as pmol product/mg tissue/h. Experiments were carried out in duplicate or triplicate.

ayer chromatography using ethanol:chloroform as the mobile phase. Thin-layer chromatography plates were analysed using a Bioscan imager (Bioscan, Washington DC, USA) and fractional steroid conversion calculated. Results were expressed as pmol product/mg tissue/h. Experiments were carried out in duplicate or triplicate. RNA extraction and reverse transcription RNA was extracted from synovium using a single-step method (TRI Reagent, Sigma, Poole, UK). Aliquots (1 μg) of RNA were reverse transcribed using random hexamers in a 20 μl volume according to the manufacturer’s protocol (Promega, Madison, USA). Real-time PCR Expression of mRNA for 11β-HSD1/2 was assessed by real-time PCR in an ABI 7500 system (Applied Biosytems, Warrington, UK). Reactions were performed in 25 μl aliquots on a 96-well plate (Sigma). Primers for 18S were used as an internal reference. Reactions contained TaqMan PCR master mix (Applied Biosytems), 900 nmol primers, 100–200 nmol TaqMan probe and 25–50 ng cDNA. Reactions were as follows: 50°C for 2 min, 95°C for 10 min, 44 cycles of 95°C for 15 s and 60°C for 1 min. Data were obtained as Ct values (cycle number at which logarithmic PCR plots cross a calculated threshold line) according to the manufacture’s guidelines, and used to determine ΔCt values (Ct of target gene – Ct of housekeeping gene) as raw data for gene expression. Probe and primer sequences were 11β-HSD1:

for 1 min. Data were obtained as Ct values (cycle number at which logarithmic PCR plots cross a calculated threshold line) according to the manufacture’s guidelines, and used to determine ΔCt values (Ct of target gene – Ct of housekeeping gene) as raw data for gene expression. Probe and primer sequences were 11β-HSD1: forward AGGAAAGCTCATGGGAGGACTAG, reverse ATGGTGAATATCATCATGAAAAAGATTC, probe CATGCTCATTCTCAACCACATCACCAACA; 11β-HSD2: forward CAGGTGTCCTAGTGCACATTGAC, reverse GTAGCCCACTCTCTCGTCCAA, probe AAGGCACGCCCTCCCAGCG. Immunohistochemistry Cryostat sections of synovial tissue were analysed with previously validated sheep polyclonal antibodies to 11β-HSD1 and 11β-HSD2 1/50 (The Binding Site, Birmingham, UK)14 21 with anti-sheep/goat biotin AB360 as secondary antibody 1/50. Immunohistochemistry was also carried out using polyclonal antisera to the endothelial marker von Willebrand’s factor 1/1000 (Dako, Ely, UK); the fibroblast marker ASO2 (CD90) 1/100 (Invitrogen, Paisley, UK); and the monocyte/macrophage marker CD68 1/100 (BD Biosciences, New Jersey, USA). Immunohistochemical analyses were carried out as described previously.12

using polyclonal antisera to the endothelial marker von Willebrand’s factor 1/1000 (Dako, Ely, UK); the fibroblast marker ASO2 (CD90) 1/100 (Invitrogen, Paisley, UK); and the monocyte/macrophage marker CD68 1/100 (BD Biosciences, New Jersey, USA). Immunohistochemical analyses were carried out as described previously.12 Measurement of corticosteroid levels in serum and synovial fluids Serum and synovial fluid cortisol and cortisone levels were measured using a specific cortisol ELISA (R&D Systems, Abingdon, UK) and a previously reported radioimmunoassay for cortisone.22 Since serum and synovial fluid samples contain cortisol-binding proteins these techniques only give information about total cortisol concentration (rather than unbound “free” levels).23 Because there is much less protein binding of cortisone, levels of cortisone ascertained by these methods are more likely to reflect “free” levels.22 Measurement of urinary corticosteroid metabolites The gas chromatography–mass spectrometry method was based on Palermo et al24 using a 5970 mass spectrometer (Hewlett-Packard, Houston, Texas, USA). Intra-assay and interassay coefficients of variation were <10% for cortisol and cortisone. 11β-HSD1 activity was calculated as the tetrahydrocortisol + allo-tetrahydrocortisol/tetrahydrocortisone ((THF+alloTHF)/THE) ratio. Renal 11β-HSD2 activity was calculated as the urinary-free cortisol/cortisone ratio (UFF/UFE). The sum of total corticosteroid metabolites (THF, alloTHF, THE, cortols, cortolones, UFF, UFE) was used as an index of total cortisol secretion.

ortisol + allo-tetrahydrocortisol/tetrahydrocortisone ((THF+alloTHF)/THE) ratio. Renal 11β-HSD2 activity was calculated as the urinary-free cortisol/cortisone ratio (UFF/UFE). The sum of total corticosteroid metabolites (THF, alloTHF, THE, cortols, cortolones, UFF, UFE) was used as an index of total cortisol secretion. Analysis of interleukin 6 (IL6) levels by ELISA Soluble IL6 in tissue supernatants was measured using a sandwich ELISA (BD Biosciences Pharmingen, Torreyana, San Diego, USA). The limit of detection was 2.2 pg/ml and intra-assay and interassay coefficients of variation were 5.2% and 9.3%. Data were expressed as pg IL6/mg tissue. Statistics Data were reported as the mean (SD) of replicate mean values for separate synovial explants unless otherwise stated. Regression analysis was performed using Microsoft Excel 2003. One-way analysis of variance was performed using SPSS Data Editor.

Analysis of interleukin 6 (IL6) levels by ELISA Soluble IL6 in tissue supernatants was measured using a sandwich ELISA (BD Biosciences Pharmingen, Torreyana, San Diego, USA). The limit of detection was 2.2 pg/ml and intra-assay and interassay coefficients of variation were 5.2% and 9.3%. Data were expressed as pg IL6/mg tissue. Statistics Data were reported as the mean (SD) of replicate mean values for separate synovial explants unless otherwise stated. Regression analysis was performed using Microsoft Excel 2003. One-way analysis of variance was performed using SPSS Data Editor. RESULTS Enzyme activity studies 11β-HSD enzyme activity was present in all synovial samples from patients with RA (fig 1). Activity was also present in synovium from patients with osteoarthritis (OA). Activity was bidirectional in all cases, although the relative amounts of activation (cortisone to cortisol conversion) to inactivation (cortisol to cortisone) varied between subjects. There was no significant difference between RA and OA synovium in overall levels of each activity or in the ratio of activation to inactivation. Compared with patient-matched skin samples synovial samples had greater mean (SE) activating capacity (14.1 (1.8) vs 8.9 (1.3) pmol/mg protein/h; p<0.001) and a higher ratio of activation to inactivation (ratio 1.7 (0.4) vs 0.6 (0.1); p<0.01), but there was no difference in inactivating capacity (16.1 (3.5) vs 20.1 (3.4); NS). For glucocorticoid activation in synovium there was a significant positive correlation with the preoperative erythrocyte sedimentation rate (ESR; fig 2A), but this relationship was not seen with CRP (r2 = 0.04, NS). Inactivation did not correlate with either inflammatory marker. Synovial tissue also metabolised prednisone and prednisolone in a manner indistinguishable from cortisone/cortisol (for 100 nM substrate activation assay: cortisone to cortisol 16.3 (4.7) pmol/mg protein/h, prednisone to prednisolone 23.0 (5.1); for inactivation assay; cortisol to cortisone 15.2 (6.1), prednisolone to prednisone 12.5 (4.7); n = 4, NS).

rednisone and prednisolone in a manner indistinguishable from cortisone/cortisol (for 100 nM substrate activation assay: cortisone to cortisol 16.3 (4.7) pmol/mg protein/h, prednisone to prednisolone 23.0 (5.1); for inactivation assay; cortisol to cortisone 15.2 (6.1), prednisolone to prednisone 12.5 (4.7); n = 4, NS). Figure 1 Glucocorticoid metabolism in synovial explants taken from patients with rheumatoid arthritis (RA) and osteoarthritis (OA). The ability of synovial tissue to interconvert cortisone and cortisol was examined using tissue freshly isolated after joint replacement surgery. Data are steroid generation for each sample adjusted for protein content and assay time and are expressed as mean (SD) of samples from each patient. Figure 2 Relationship between glucocorticoid activating capacity of synovial explants and the systemic inflammatory response and the effect of a specific inhibitor of 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) on glucocorticoid activation and inactivation. (A) A significant correlation was seen between glucocorticoid activation (oxoreductase activity) and the erythrocyte sedimentation rate (ESR) measured before surgery in patients with rheumatoid arthritis (RA). (B) The effect of PF-877423, a specific bidirectional inhibitor of 11β-HSD1 enzyme activity, on glucocorticoid activation (oxoreductase activity, cortisone to cortisol conversion) and inactivation (dehydrogenase activity, cortisol to cortisone conversion) in synovium from patients with RA or osteoarthritis (OA). The inhibitor reduced the capacity of synovium to activate glucocorticoids but was unable to block the inactivating capacity. In OA tissue inactivation increased with 11β-HSD1 inhibition. These results indicate that the inactivating capacity is not due to 11β-HSD1. Characterisation of glucocorticoid inactivating capacity The glucocorticoid activating (oxoreductase) capacity of synovial tissue appeared to be due to expression of 11β-HSD1 as this is the only enzyme known to convert cortisone to cortisol. Additionally, we have previously localised 11β-HSD1 in synovium by immunohistochemistry12 and RT-PCR on synovial tissue demonstrated abundant 11β-HSD1 mRNA expression (Ct value 32.0 (0.7) relative to 12.9 (1.9) for 18S).

be due to expression of 11β-HSD1 as this is the only enzyme known to convert cortisone to cortisol. Additionally, we have previously localised 11β-HSD1 in synovium by immunohistochemistry12 and RT-PCR on synovial tissue demonstrated abundant 11β-HSD1 mRNA expression (Ct value 32.0 (0.7) relative to 12.9 (1.9) for 18S). The glucocorticoid inactivating (dehydrogenase) activity could be due to either 11β-HSD1, 11β-HSD2 or a combination of both. The lack of correlation between activation and inactivation (p = 0.56) suggested the presence of more than one enzyme and a specific inhibitor of 11β-HSD1 (PF-877423; Pfizer, New York, USA) was used to clarify this further. PF-877423 blocks both activating (oxoreductase) and inactivating (dehydrogenase) directions of 11β-HSD1 but has no effect on 11β-HSD2 activity (Bujalska et al, submitted for publication). Enzyme activity data indicated that the inhibitor reduced glucocorticoid activation in synovial tissue from patients with RA and OA but increased glucocorticoid inactivation in OA tissue (fig 2B). There was also a trend towards higher inactivation in RA tissue treated with inhibitor (p = 0.07). These observations suggested that glucocorticoid inactivation in synovial tissue was due to expression of 11β-HSD2 rather than 11β-HSD1. 11β-HSD2 was detectable by RT-PCR on RNA extracted from synovial tissue (Ct 33.6 (1.0) relative to 12.1 (1.8) for 18S). Immunohistochemistry for 11β-HSD2 demonstrated expression of the enzyme in both RA and OA tissue, where it colocalised predominantly with CD68-positive macrophages (fig 3). This contrasted with 11β-HSD1 expression which was localised primarily to fibroblasts.

ovial tissue (Ct 33.6 (1.0) relative to 12.1 (1.8) for 18S). Immunohistochemistry for 11β-HSD2 demonstrated expression of the enzyme in both RA and OA tissue, where it colocalised predominantly with CD68-positive macrophages (fig 3). This contrasted with 11β-HSD1 expression which was localised primarily to fibroblasts. Figure 3 Immunohistological localisation of 11β-hydroxysteroid dehydrogenase (11β-HSD) enzymes in synovium. The expression of 11β-HSD1 and 11β-HSD2 within rheumatoid synovium were examined using specific antibodies. 11β-HSD1 expression colocalised predominantly with a fibroblast marker. Expression of 11β-HSD2 colocalised with expression of CD68. Functional consequences of enzyme activity High levels of IL6 were produced by synovial tissue. As expected, levels were higher in patients with RA than in those with OA. Incubation of synovial tissue with cortisol significantly reduced IL6 production, indicating tissue sensitivity to glucocorticoids. Incubation with cortisone also caused a significant reduction in IL6 expression, suggesting that synovial 11β-HSD1 was functionally active, converting cortisone to cortisol in an autocrine fashion (fig 4A). This was confirmed by studies using PF-877423, which blocked the suppressive effect of cortisone on IL6 (fig 4B).

Incubation with cortisone also caused a significant reduction in IL6 expression, suggesting that synovial 11β-HSD1 was functionally active, converting cortisone to cortisol in an autocrine fashion (fig 4A). This was confirmed by studies using PF-877423, which blocked the suppressive effect of cortisone on IL6 (fig 4B). Figure 4 The functional effect of glucocorticoid activation on interleukin 6 (IL6) synthesis in patients with rheumatoid arthritis (RA) and osteoarthritis (OA). (A) Synovial tissue was incubated for 24 h in the presence or absence of 100 nM cortisone. In both RA and OA synovium cortisone had a significant suppressive effect on IL6 production. (B) A specific 11β-HSD1 inhibitor (PF-877423) was able to block the decrease in IL6 production caused by cortisone treatment. Measurement of tissue glucocorticoid levels in vivo To evaluate net synovial tissue glucocorticoid metabolism within the joint we measured corticosteroid levels in paired serum and synovial fluid samples. To minimise the confounding effect of cortisol-binding proteins we measured serum-to-synovial fluid concentration gradients in cortisone since cortisone has limited binding to these proteins. Cortisone levels in synovial fluid obtained from patients with RA were significantly different from plasma levels (serum 56.7 (11.8) nmol/l, synovial fluid 20.9 (8.7) nmol/l; mean fall from serum to synovial fluid 63%, p<0.001). The low cortisone levels in synovial fluid suggested local net conversion of cortisone to cortisol.

e levels in synovial fluid obtained from patients with RA were significantly different from plasma levels (serum 56.7 (11.8) nmol/l, synovial fluid 20.9 (8.7) nmol/l; mean fall from serum to synovial fluid 63%, p<0.001). The low cortisone levels in synovial fluid suggested local net conversion of cortisone to cortisol. Measures of systemic glucocorticoid metabolism in inflammatory arthritis Systemic measures of 11β-HSD1 activity were assessed in subjects with untreated RA compared with patients with non-inflammatory joint disease. There was a significantly higher ratio of cortisol to cortisone metabolites in the urine of patients with RA, suggesting higher 11β-HSD1 activity (figs 5A and B). No change was seen in the ratio of urinary cortisol to cortisone indicating that renal 11β-HSD2 was unchanged. There was a positive correlation between 11β-HSD1 activity and ESR (but not CRP) in patients with RA (figs 5C and D). The urinary cortisol/cortisone ratio also correlated with ESR in patients with RA (r = 0.42, p<0.05) probably reflecting the contribution 11β-HSD1 makes to this ratio. There was no correlation between the urinary cortisol/cortisone ratio and CRP (r = 0.11, NS). There was no correlation between total urinary corticosteroid metabolites and ESR or CRP (r = 0.29 and 0.08, both NS).

R in patients with RA (r = 0.42, p<0.05) probably reflecting the contribution 11β-HSD1 makes to this ratio. There was no correlation between the urinary cortisol/cortisone ratio and CRP (r = 0.11, NS). There was no correlation between total urinary corticosteroid metabolites and ESR or CRP (r = 0.29 and 0.08, both NS). Figure 5 Systemic measures of local glucocorticoid metabolism in patients with rheumatoid arthritis (RA) and non-inflammatory joint disease. (A, B) Measurement of the balance of glucocorticoid activation/inactivation in non-inflammatory controls and patients with untreated RA. There was a significant increase in the tetrahydrocortisol + allo-tetrahydrocortisol/tetrahydrocortisone (THF+alloTHF)/THE ratio in patients with inflammatory arthritis, whereas the urinary free cortisol/cortisone ratio was unchanged. This indicated that 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1)-mediated glucocorticoid activation was enhanced in patients with RA. (C, D) In patients with RA the relationship between the (THF+alloTHF)/THE ratio, a systemic measure of 11β-HSD1 activity, and inflammatory markers was examined. The erythrocyte sedimentation rate (ESR), but not C-reactive protein (CRP), demonstrated a significant correlation with this measure, suggesting increasing 11β-HSD1-mediated glucocorticoid activation with increasing degrees of inflammation. DISCUSSION By a variety of measures, substantial glucocorticoid metabolism was identified in the joint. The net consequence was local glucocorticoid excess. In patients with RA, glucocorticoid generation within synovium, and systemic levels of active steroid, correlated with disease activity and had functional effects on the inflammatory response. Although this may form part of an endogenous mechanism for immunoregulation, a subset of synovial cells were able to inactivate glucocorticoids through expression of 11β-HSD2 and thus may be glucocorticoid resistant.

tive steroid, correlated with disease activity and had functional effects on the inflammatory response. Although this may form part of an endogenous mechanism for immunoregulation, a subset of synovial cells were able to inactivate glucocorticoids through expression of 11β-HSD2 and thus may be glucocorticoid resistant. Soon after the initial use of cortisone and hydrocortisone for RA it was realised that these steroids were interconverted in vivo and that 11β-HSD activity was essential for conversion of cortisone to cortisol.25 This suggested that synovial tissue itself might metabolise corticosteroids. Initially, it was thought that patients with RA would have increased ability to inactivate hydrocortisone.26 Although the understanding of tissue steroid metabolism was rudimentary, bidirectional 11β-HSD activity was noted in rheumatoid synovial tissue27 and significant amounts of cortisol were generated from cortisone injected into inflamed joints.28 Recently, interest has been renewed, with reports implicating 11β-HSD2 expression in RA. 11β-HSD2 was the most upregulated of >4300 genes in a study examining PBMCs from patients with recent onset RA compared with longstanding RA.18 11β-HSD2 was one of three (out of 20 000) significantly upregulated genes in PBMCs from identical twins discordant for RA.17 Analysis of rheumatoid synovium indicated that 11β-HSD2 was expressed predominantly in synovial macrophages. A separate study examined the capacity of synovial explants to inactivate steroids and described how this activity was higher in patients with RA.16 That study did not involve direct measurement of glucocorticoid activation and relied on 11β-HSD inhibitors with limited specificity. Furthermore, the association between local and systemic glucocorticoid metabolism was not assessed. To address this, we used specific enzyme assays and selective inhibitors to define the nature of glucocorticoid metabolising activity. We additionally examined possible consequences of corticosteroid metabolism on local and systemic levels of glucocorticoids.

cal and systemic glucocorticoid metabolism was not assessed. To address this, we used specific enzyme assays and selective inhibitors to define the nature of glucocorticoid metabolising activity. We additionally examined possible consequences of corticosteroid metabolism on local and systemic levels of glucocorticoids. The generation of substantial amounts of active glucocorticoids within the joint could account for the reason why therapeutic glucocorticoids are so effective at dampening flares of synovial inflammation. In the inflamed joint, cells will be exposed to both circulating glucocorticoid and glucocorticoids generated from circulating inactive precursors by cells expressing 11β-HSD1. Locally produced glucocorticoids will be free to diffuse into surrounding tissues. This could affect the integrity of adjacent connective tissue. The periarticular osteoporosis seen in RA is of multifactorial origin but high local glucocorticoid levels would suppress bone-forming ability and thus contribute to uncoupling of bone resorption from formation.29 High glucocorticoid levels would also be expected to have an immunosuppressive effect. This could contribute to the increased risk of septic arthritis in RA.30

factorial origin but high local glucocorticoid levels would suppress bone-forming ability and thus contribute to uncoupling of bone resorption from formation.29 High glucocorticoid levels would also be expected to have an immunosuppressive effect. This could contribute to the increased risk of septic arthritis in RA.30 The finding of 11β-HSD2 expression in synovium raises the possibility that some cells within synovium are resistant to steroids through expression of this enzyme. Further assessment of the functional implications of 11β-HSD2 expression is beyond the scope of this study but there is a need to define further the phenotype of 11β-HSD2-expressing synovial cells and to clarify the relationship these cells have with 11β-HSD2-positive PBMCs. An important and unexpected finding in this study is the presence of considerable glucocorticoid metabolising capacity in synovium from patients with OA. This tissue was initially used as a control for patients with RA with the expectation that glucocorticoid metabolism would be less prominent. However, net activity appeared similar to that of RA. Additionally, the relative decrease in IL6 production in response to cortisone treatment was similar. The inability of the inhibitor to block the capacity of OA synovium to convert cortisol to cortisone also suggests that 11β-HSD2 is expressed. 11β-HSD2 expression in OA has been identified in a previous study,16 which suggested that its expression is lower in OA than RA. The roles of 11β-HSD1/2 in OA remain to be defined.

inability of the inhibitor to block the capacity of OA synovium to convert cortisol to cortisone also suggests that 11β-HSD2 is expressed. 11β-HSD2 expression in OA has been identified in a previous study,16 which suggested that its expression is lower in OA than RA. The roles of 11β-HSD1/2 in OA remain to be defined. Glucocorticoid activation was associated with ESR but not CRP measurements. A potential explanation for this is the confounding effect glucocorticoid activation had on synovial IL6 production. CRP synthesis is primarily an IL6-driven process and thus the relationship between joint inflammation and CRP will be complicated by the inhibitory effect of glucocorticoids on IL6. The ESR is less dependent on IL6 and so less likely to be directly influenced by local glucocorticoid metabolism.

synovial IL6 production. CRP synthesis is primarily an IL6-driven process and thus the relationship between joint inflammation and CRP will be complicated by the inhibitory effect of glucocorticoids on IL6. The ESR is less dependent on IL6 and so less likely to be directly influenced by local glucocorticoid metabolism. The data presented here have several limitations. First, synovial tissue was obtained from patients with RA treated with various disease-modifying drugs. The small number of patients studied makes it difficult to comment about potential effects of antirheumatic treatment on steroid metabolising enzyme activity. Second, synovial tissue was examined in patients who had established RA. It will be of interest to study temporal changes in glucocorticoid metabolism in patients with RA at different disease stages and to assess whether findings are specific to RA or a general feature of persistent synovial inflammation. Although urinary measures of systemic steroid metabolism correlated well with measures in synovial tissue and fluid, an intrinsic limitation is that these might reflect altered steroid metabolism in tissues other than synovium. We have previously reported that inflammatory cytokines increase 11β-HSD1 expression in fibroblasts from several tissues12 and 11β-HSD1 expression has been reported in subpopulations of peritoneal macrophages31 and T cells.32 Regardless of the extent of altered glucocorticoid metabolism in RA, the functional consequences of this deserve further study.

inflammatory cytokines increase 11β-HSD1 expression in fibroblasts from several tissues12 and 11β-HSD1 expression has been reported in subpopulations of peritoneal macrophages31 and T cells.32 Regardless of the extent of altered glucocorticoid metabolism in RA, the functional consequences of this deserve further study. We thank Mr Andrew Thomas, Dr Shinner and the operating theatre staff of the Royal Orthopaedic Hospital in Birmingham for their help in sample collection. Sue Hughes and Beverley Hughes carried out enzyme assays and biochemical measurements. Debbie Hardie performed immunohistochemical analysis on synovial tissue. Funding: This study was funded by the Arthritis Research Campaign (project grant number 18081) and the Medical Research Council, UK. Competing interests: None. Ethics approval: Approved by the local ethics committee.

Rheumatoid arthritis (RA) affects almost 1% of the adult population.1 Affected subjects experience considerable morbidity, including a rapid loss of function,2 3 joint destruction, permanent deformity and reduced life expectancy.4 5 Based on published guidelines,6 7 treatment of RA should focus on preserving function, preventing or controlling joint damage and achieving remission of disease activity. Progressive radiographic damage and deterioration in physical function occur in patients with RA who do not receive timely effective treatment. Several long-term prospective studies have looked at the relationship between radiographic damage and physical function in patients with RA.8–10 A prospective 6-year follow-up study by Kuper et al10 showed that large joint damage occurs early in the disease, with at least 20% of patients having damage in at least one joint within 1 year of disease onset and 50% of patients having some damage within 6 years of disease onset. Radiographic changes were significantly related to damage in the hands and feet, the physical disability index and cumulative disease activity.10 In a review of the links between joint damage and disability in patients with RA, Scott et al reported that joint damage progresses continuously over the first 20 years of RA and that this damage accounts for 25% of the disability seen in patients.9 Radiographic damage is thus clearly a major determinant of long-term physical function in patients with established RA.

disability in patients with RA, Scott et al reported that joint damage progresses continuously over the first 20 years of RA and that this damage accounts for 25% of the disability seen in patients.9 Radiographic damage is thus clearly a major determinant of long-term physical function in patients with established RA. Although studies have shown that a relationship between physical function and radiographic damage exists, they have mostly examined this association looking at radiographic damage over the long term. It is not known whether progression of radiographic damage over short periods is associated with immediate impairment in physical function. This question is particularly relevant today because of the availability of powerful treatments that can prevent radiographic progression if started early enough. For example, in clinical trials, etanercept, either alone or in combination with methotrexate, has been shown to reduce radiographic progression in patients with recent and long-term RA.11–13 Our study used data from the Trial of Etanercept and Methotrexate with Radiographic Patient Outcomes (TEMPO),11 14 to investigate the longitudinal relationship between physical function and both level of radiographic damage and radiographic progression over a 2-year period, in a cohort of patients with early or advanced active RA.

ed data from the Trial of Etanercept and Methotrexate with Radiographic Patient Outcomes (TEMPO),11 14 to investigate the longitudinal relationship between physical function and both level of radiographic damage and radiographic progression over a 2-year period, in a cohort of patients with early or advanced active RA. METHODS TEMPO was a multicentre, double-blind study that compared the effect of etanercept plus methotrexate versus etanercept alone versus methotrexate alone in 686 patients with active RA of 6 months to 20 years’ duration.11 14 Details of this trial are published elsewhere.11 14–16 All patients were randomly assigned to one of three treatment groups: etanercept (25 mg twice-weekly subcutaneous doses), methotrexate (7.5–20 mg weekly oral doses) or combination therapy with etanercept and methotrexate. Physical function was measured by the Health Assessment Questionnaire (HAQ) at baseline and at protocol-specified intervals throughout the study; scores at baseline and 1 and 2 years were used in the analysis. The HAQ is scored on a scale of 0–3, with higher scores indicating increased disability.

All patients were randomly assigned to one of three treatment groups: etanercept (25 mg twice-weekly subcutaneous doses), methotrexate (7.5–20 mg weekly oral doses) or combination therapy with etanercept and methotrexate. Physical function was measured by the Health Assessment Questionnaire (HAQ) at baseline and at protocol-specified intervals throughout the study; scores at baseline and 1 and 2 years were used in the analysis. The HAQ is scored on a scale of 0–3, with higher scores indicating increased disability. Radiographs of the hands, wrists and forefeet were taken at baseline, 6 months, 1 year and 2 years. Digitised images were scored by two readers, using the van der Heijde-modified Sharp method.17 Radiographic progression was determined for each interval by subtracting status scores of two time points. Radiographic progression was divided into four categories (negative, <0; zero, 0–1; minor, 1–5; and greater progression, >5) for comparisons with HAQ scores. The radiographic population consisted of patients who had an acceptable baseline and at least one acceptable post-baseline film.

status scores of two time points. Radiographic progression was divided into four categories (negative, <0; zero, 0–1; minor, 1–5; and greater progression, >5) for comparisons with HAQ scores. The radiographic population consisted of patients who had an acceptable baseline and at least one acceptable post-baseline film. Statistical analysis To control for within-patient correlation, generalised mixed linear modelling was used to model the dependent variable HAQ score by the absolute Sharp score (damage score) or the interval change in Sharp score, with age, sex, disease duration, treatment, Disease Activity Score (DAS) and C-reactive protein (CRP) levels as covariates. If interval change in Sharp score was entered as an independent variable, the HAQ score at the end of the interval was chosen as the dependent variable in the model and the HAQ score at baseline was omitted. A random intercept and a compound symmetry covariance structure seemed to best fit the data. Results are expressed as estimated marginal means. In separate analyses, the interaction of disease duration and (change in) Sharp score with respect to HAQ score was tested.

the model and the HAQ score at baseline was omitted. A random intercept and a compound symmetry covariance structure seemed to best fit the data. Results are expressed as estimated marginal means. In separate analyses, the interaction of disease duration and (change in) Sharp score with respect to HAQ score was tested. RESULTS Of the 686 randomly assigned patients, 622 (91%) had a baseline and at least one follow-up film and were included in the radiographic analysis. No statistically significant differences were found among the three treatment groups in baseline demographic and disease characteristics. Patients included in the analysis were predominantly women (76.5%), with mean disease duration of 6.35 years, mean HAQ score of 1.72 and mean total Sharp score (TSS) of 33.0 at baseline. In the linear mixed model, which was applied here in order to aggregate data of baseline, year 1 and year 2 data under simultaneous adjustment for within-patient correlation, the TSS was significantly positively associated with the HAQ score independently of the DAS (table 1). A number of additional variables were independently associated with the HAQ score, including age (positive correlation), sex (women had higher HAQ scores) and the DAS (positive correlation).

ithin-patient correlation, the TSS was significantly positively associated with the HAQ score independently of the DAS (table 1). A number of additional variables were independently associated with the HAQ score, including age (positive correlation), sex (women had higher HAQ scores) and the DAS (positive correlation). Table 1 Sharp score is positively associated with Health Assessment Questionnaire (HAQ) score independently of Disease Activity Score (DAS): generalised mixed linear modelling* Parameter Regression coefficient p Value 95% CI of the regression coefficient Lower bound Upper bound Intercept −0.125 0.26 −0.342 0.092 Sex (F) −0.243 <0.001 −0.333 −0.152 Age (years) 0.0103 <0.001 0.0073 0.0133 Disease duration (years) 0.0059 0.174 −0.0026 0.0143 CRP (mg/dl) 0.0029 <0.001 0.0014 0.0044 Treatment 0.0104 0.67 −0.0369 0.0577 DAS 0.255 <0.001 0.240 0.2688 Sharp score (Sharp units) 0.0019 <0.001 0.0011 0.0027 *Dependent variable: HAQ. CRP, C-reactive protein.

Table 1 Sharp score is positively associated with Health Assessment Questionnaire (HAQ) score independently of Disease Activity Score (DAS): generalised mixed linear modelling* Parameter Regression coefficient p Value 95% CI of the regression coefficient Lower bound Upper bound Intercept −0.125 0.26 −0.342 0.092 Sex (F) −0.243 <0.001 −0.333 −0.152 Age (years) 0.0103 <0.001 0.0073 0.0133 Disease duration (years) 0.0059 0.174 −0.0026 0.0143 CRP (mg/dl) 0.0029 <0.001 0.0014 0.0044 Treatment 0.0104 0.67 −0.0369 0.0577 DAS 0.255 <0.001 0.240 0.2688 Sharp score (Sharp units) 0.0019 <0.001 0.0011 0.0027 *Dependent variable: HAQ. CRP, C-reactive protein. The HAQ score is primarily determined by disease activity. In the context of a clinical trial, it is expected that the disease activity would show important variation that would easily obscure a contributory and independent association with radiographic damage or progression. We first investigated the relationship between the DAS and the HAQ score. After adjustment for age and sex, the relationship between DAS and the HAQ score was almost linear (fig 1). This linear relationship justifies why in every subsequent analysis to determine the contribution of radiographic damage to explaining physical function we have adjusted for the known relationship between the DAS and the HAQ score.

r age and sex, the relationship between DAS and the HAQ score was almost linear (fig 1). This linear relationship justifies why in every subsequent analysis to determine the contribution of radiographic damage to explaining physical function we have adjusted for the known relationship between the DAS and the HAQ score. Figure 1 Marginal means for the Health Assessment Questionnaire (HAQ) score, adjusted for age, sex, disease duration, treatment, Sharp score and time as a function of disease activity score (DAS) in categories of one unit (error bars reflect standard error). To visualise the adjusted relationship between the HAQ score and the Sharp score, Sharp scores were divided into six categories of 10 Sharp units each and estimated marginal means were calculated. Figure 2 shows estimated marginal means of the fitted model (mean HAQ scores adjusted for age, sex, DAS, treatment and time) for each category of 10 Sharp units and clearly points to an increasing trend. A separate analysis, in which the Sharp score was replaced by the change in Sharp score, also indicated that the change in Sharp score (interval progression; Sharp units per year) was significantly and independently associated with the HAQ score (p = 0.001; table 2). This implies that patients who show joint damage progression have worse physical function, independent of disease activity and the type of treatment they use, although the effects were small.

erval progression; Sharp units per year) was significantly and independently associated with the HAQ score (p = 0.001; table 2). This implies that patients who show joint damage progression have worse physical function, independent of disease activity and the type of treatment they use, although the effects were small. Figure 2 Marginal means for the Health Assessment Questionnaire (HAQ) scores adjusted for age, sex, disease duration, treatment, Disease Activity Score (DAS), Sharp score and time as a function of increasing Sharp scores (in categories of 10 units; error bars reflect standard error). Table 2 Health Assessment Questionnaire (HAQ) score is positively associated with radiographic progression rate independently of disease activity (Disease Activity Score (DAS) and C-reactive protein (CRP)): generalised mixed linear modelling* Parameter Regression coefficient p Value 95% CI of the regression coefficient Lower bound Upper bound Intercept −0.0142 0.922 −0.3007 0.2723 Sex (F) −0.3118 <0.001 −0.4307 −0.1928 Age (years) 0.0120 <0.001 0.0080 0.0159 Disease duration (years) 0.0113 0.057 −0.0015 0.0241 CRP (mg/dl) 0.0024 0.085 0.0003 0.0051 Treatment 0.0026 0.934 −0.0997 0.0649 DAS 0.2313 <0.001 0.1969 0.2658 Change in Sharp score (Sharp units/year) 0.0088640 0.001 0.0035 0.0143 *Dependent variable: HAQ

<0.001 −0.4307 −0.1928 Age (years) 0.0120 <0.001 0.0080 0.0159 Disease duration (years) 0.0113 0.057 −0.0015 0.0241 CRP (mg/dl) 0.0024 0.085 0.0003 0.0051 Treatment 0.0026 0.934 −0.0997 0.0649 DAS 0.2313 <0.001 0.1969 0.2658 Change in Sharp score (Sharp units/year) 0.0088640 0.001 0.0035 0.0143 *Dependent variable: HAQ Figure 3 shows the effect of changes in radiographic damage on physical function, with progression of radiographic damage stratified into four categories (negative, zero, minor and greater progression). HAQ scores tended to increase with increasing radiographic progression, although the differences were small and probably not clinically meaningful. Figure 3 also suggests that patients with negative progression scores have lower HAQ scores than patients with positive progression scores, but this effect was not statistically significant.

to increase with increasing radiographic progression, although the differences were small and probably not clinically meaningful. Figure 3 also suggests that patients with negative progression scores have lower HAQ scores than patients with positive progression scores, but this effect was not statistically significant. Figure 3 Marginal means for the Health Assessment Questionnaire (HAQ) score, adjusted for age, sex, disease duration, treatment, disease activity score (DAS), Sharp score and time, as a function of change in Sharp score (four categories of progression; error bars reflect standard error). In an attempt to determine whether the documented association between the Sharp score and the HAQ score, as well as the association between the change in Sharp score and HAQ score, was dependent on disease duration, we tested the following interactions: Sharp score and disease duration and change in Sharp score and disease duration, with disease duration dichotomised at a cut-off level of 3 years, with respect to explaining variation in HAQ score. Both interactions were not statistically significant (results not shown)

e duration, we tested the following interactions: Sharp score and disease duration and change in Sharp score and disease duration, with disease duration dichotomised at a cut-off level of 3 years, with respect to explaining variation in HAQ score. Both interactions were not statistically significant (results not shown) DISCUSSION This analysis provides further evidence for the longitudinal relationship between radiographic damage and physical function. As shown, progression of radiographic damage over short periods of time is associated with worse physical function, which is independent of the effects of inflammatory disease activity (DAS). In addition, patients who had negative radiographic progression tended to have better physical function than patients with zero or low positive progression scores. This finding gains importance in the light of the low radiographic progression rates that were found in the 2-year TEMPO trial.14 Despite these low progression rates, this (sensitive) longitudinal analysis suggests that any type of radiographic progression has repercussions for physical functioning. It may be argued that DAS and CRP inappropriately reflect disease activity and that the proposed association between radiographic progression and physical function would disappear if a variable that better reflects inflammatory activity were used. This is, however, highly unlikely. The contribution of radiographic progression to the HAQ score was similarly high before and after adjustment for DAS and CRP, so that both effects are truly separated (data not shown).

hysical function would disappear if a variable that better reflects inflammatory activity were used. This is, however, highly unlikely. The contribution of radiographic progression to the HAQ score was similarly high before and after adjustment for DAS and CRP, so that both effects are truly separated (data not shown). In a 10-year longitudinal study,18 we have previously observed that radiographic damage and disease activity independently contribute to changes in physical function in RA regardless of disease duration. Others have reported the longitudinal relationship between joint damage and physical function after adjustment for disease activity.19 In that particular study, performed on an inception cohort of patients with RA followed up for many years while receiving standard conventional antirheumatic treatment, the main driver of physical function was determined to be disease activity. But Welsing et al also found a significant contribution of joint damage (measured by van der Heijde’s modification of the Sharp score) to explain the variation in physical function.19

ile receiving standard conventional antirheumatic treatment, the main driver of physical function was determined to be disease activity. But Welsing et al also found a significant contribution of joint damage (measured by van der Heijde’s modification of the Sharp score) to explain the variation in physical function.19 There are important differences between these previous studies18 19 and our current study: First, we analysed a trial population using highly effective treatments. In our study, radiographic progression was minimal, although the level of baseline damage was, on average, substantial. Apart from that, the active trial treatment resulted in very low levels of disease activity in the majority of the trial population. Even with such a low level of variation in the determinants of function (disease activity and radiographic damage/progression) a relationship between radiographic damage progression and the HAQ score was suggested by this analysis. Second, in our study, we investigated the contribution of radiographic progression itself, rather than radiographic damage at any particular time point. Our analysis suggests that both the level of damage, and the progression rate, although very low, contributed to variation in physical function. Third, our observations were made in a 2-year follow-up study, whereas the previous studies18 19 investigated function over a period of 9–12 years, in which deterioration of physical function is likely to be far more substantial. Taken together, however, the longitudinal studies provide evidence that physical function in patients with RA is determined not only by disease activity but also by radiographic damage and the rate at which radiographic damage increases over time. Fluctuations in the progression rate and in the disease activity cause demonstrable effects on physical function within very short time intervals. These data add to the ever-growing evidence that it is important not only to suppress disease activity but also to halt radiographic progression to maintain physical function.

uctuations in the progression rate and in the disease activity cause demonstrable effects on physical function within very short time intervals. These data add to the ever-growing evidence that it is important not only to suppress disease activity but also to halt radiographic progression to maintain physical function. It is not a coincidence that studies using longitudinal data analyses are consistent with the existence of a relationship between radiographic damage and physical function. It has been hypothesised that this relationship exists, but it has turned out to be rather difficult to confirm it statistically in studies using cross-sectional analytical approaches.20 21 A longitudinal data analysis provides far more statistical power because it uses all available data of a prospectively followed cohort in an aggregated manner, while adjusting for spurious intra-patient correlation. Apart from that, longitudinal data analysis provides insight into how changes in disease activity and radiographic progression—for example, as a result of effective treatments, influence physical function.

f a prospectively followed cohort in an aggregated manner, while adjusting for spurious intra-patient correlation. Apart from that, longitudinal data analysis provides insight into how changes in disease activity and radiographic progression—for example, as a result of effective treatments, influence physical function. Physical function is often considered to be a rather static outcome. However, recent experience from randomised clinical trials evaluating highly effective treatments has shown that impairment of physical function in patients with RA is reversible to some extent. In a recent analysis, Aletaha et al reported that the irreversible part of the HAQ score was between 0% and 33%, increasing with the duration of RA.22 We hypothesise that this irreversible impairment is caused by structural damage, while the reversible impairment is primarily inflammation. If true, early effective intervention to limit structural damage may be helpful in decreasing irreversible physical impairment.

en 0% and 33%, increasing with the duration of RA.22 We hypothesise that this irreversible impairment is caused by structural damage, while the reversible impairment is primarily inflammation. If true, early effective intervention to limit structural damage may be helpful in decreasing irreversible physical impairment. Many studies support the use of early treatment of RA. Landewé et al showed that the best way to prevent radiographic damage is to start treatment before such damage has occurred.23 If this is not possible, aggressive treatment should be started within 12–24 months after the diagnosis of RA to limit radiographic progression over time.23 Anderson et al showed that the rate of radiological progression is set during the early stages of RA and suggested resetting this progression rate through pharmacological treatment as soon as possible after the disease is recognised.24 Other studies support the finding that radiographic damage occurs early in the onset of RA and progresses throughout the disease.2 3 25–27 A recent meta-analysis showed that starting treatment early has a long-term benefit on inhibition of structural damage.28 This analysis demonstrated that the association between radiographic progression and physical function was similarly important in relatively early RA (disease duration ⩽3 years) as compared with more advanced RA (disease duration >3 years), which emphasises the importance of minimising radiographic progression at all stages of the disease.

rated that the association between radiographic progression and physical function was similarly important in relatively early RA (disease duration ⩽3 years) as compared with more advanced RA (disease duration >3 years), which emphasises the importance of minimising radiographic progression at all stages of the disease. Rheumatologists frequently question whether the small differences in radiographic progression between trial arms are clinically meaningful. This analysis supports the view that subtle differences in radiographic progression have a measurable impact on physical function, although most probably at a level that is not truly clinically meaningful for the patient over a relatively short period. The question, however, is how physical function would be influenced if mild progression were allowed for years and years, using the argument that small fluctuations are not clinically meaningful. CONCLUSIONS This analysis suggests that in patients with RA, greater radiographic damage and recent radiographic progression correlate with a higher degree of disability, after adjustment for age, sex, disease duration and disease activity. We thank Laini Dubach and Ruth Pereira for professional writing support. Funding: Professional writing support was funded by Wyeth Pharmaceuticals Competing interests: DvdH, RL and LK have received consulting fees from Wyeth, manufacturer of etanercept. RvV has declared no competing interests. SF is an employee and stockholder of Wyeth. Ethics approval: Ethics committee approval obtained.

Pulmonary arterial hypertension (PAH) is a progressive and often fatal complication of connective tissue diseases (CTDs) such as systemic sclerosis (SSc), systemic lupus erythaematosus (SLE), and overlap or mixed connective tissue disease (MCTD).1–3 CTDs are disorders characterised by a wide range of vascular, inflammatory, and fibrotic manifestations in many organs including lung, kidney, and skin. Over the past decade, advances in medical treatment have improved the management of the complications associated with CTDs. Patients with SLE have benefited from immunosuppressive treatments,4 while improved management of the specific complications associated with SSc and MCTD (eg, scleroderma renal crisis), has improved prognosis.5 However, PAH remains a major cause of long-term morbidity and mortality. The reported symptomatic PAH prevalence measured by right heart catheterisation is 8–12% in patients with SSc,1 6 6–11% in patients with SLE,7 8 and up to 10–45% in patients with MCTD3 as measured by echocardiography and/or right heart catheterisation.

er, PAH remains a major cause of long-term morbidity and mortality. The reported symptomatic PAH prevalence measured by right heart catheterisation is 8–12% in patients with SSc,1 6 6–11% in patients with SLE,7 8 and up to 10–45% in patients with MCTD3 as measured by echocardiography and/or right heart catheterisation. In these patients, early detection of PAH and a multidisciplinary approach to diagnosis and treatment in specialised PAH and/or CTD centres may improve clinical outcome.9 10 Therapeutic approaches for PAH–CTD are based on those used for treating idiopathic PAH (iPAH).11 Anticoagulation, diuretics, and oxygen supplementation are often used although the benefit of this supportive therapy has not been demonstrated in PAH–CTD.1 Prostacyclin analogues may improve exercise capacity and pulmonary haemodynamics in these patients.12–15 However, despite treatment, patients with PAH–CTDs are functionally impaired with a decreased health status and a poor prognosis. In the absence of concomitant PAH, survival of patients with SSc exceeds 90% at 1 year16 17 but once PAH has been diagnosed, it decreases to 50%,18 19 which is worse than for patients with iPAH (84%).19 The risk of death from PAH related to SSc is threefold higher than from iPAH.19

d health status and a poor prognosis. In the absence of concomitant PAH, survival of patients with SSc exceeds 90% at 1 year16 17 but once PAH has been diagnosed, it decreases to 50%,18 19 which is worse than for patients with iPAH (84%).19 The risk of death from PAH related to SSc is threefold higher than from iPAH.19 Bosentan is an oral dual (ETA and ETB) endothelin-1 receptor antagonist. In placebo-controlled clinical trials and in long-term extension studies,20–22 bosentan was well tolerated, improved exercise capacity and haemodynamics, and delayed the time to clinical worsening in patients with iPAH and PAH–CTD. Survival estimates at 1 and 2 years were 86% and 73%, respectively, in a subgroup analysis of patients with PAH–CTD.23 Improvement in quality of life (SF-36 instrument) has been reported after 3 and 6 months of bosentan treatment in patients with iPAH and PAH–CTD (59% and 41%, respectively) participating in the VITAL study.24 However, changes in health-related quality of life have not been assessed together with survival. Since the concomitant assessment of these two aspects is critical to appreciate overall outcome, the present multi-centre European study was designed to investigate changes in health-related quality of life together with survival over a 48-week observation period in patients with PAH exclusively related to CTD.

rvival. Since the concomitant assessment of these two aspects is critical to appreciate overall outcome, the present multi-centre European study was designed to investigate changes in health-related quality of life together with survival over a 48-week observation period in patients with PAH exclusively related to CTD. PATIENTS AND METHODS Patients Included patients (over 18 years of age) had PAH in World Health Organization (WHO) functional class III25 related to diffuse or limited SSc, MCTD, or SLE (other CTDs were excluded). PAH was confirmed in all patients by right heart catheterisation requiring mean pulmonary artery pressure ⩾25 mmHg at rest, pulmonary vascular resistance >3 Wood units, and pulmonary capillary wedge pressure <15 mmHg.26 This catheterisation was performed within 6 months prior to the start of bosentan therapy. Signs of right heart failure, if present at baseline, were required to be stable and patients were required to have received adequate diuretics treatment prior to bosentan initiation. Total lung capacity (TLC) was required to be above 50% of predicted, to exclude patients with severe interstitial lung disease. Patients were also excluded if they had received any PAH treatments (except anticoagulants) within 1 month of screening, if they were receiving or were expected to receive epoprostrenol or prostacyclin analogues for more than 2 consecutive weeks, or if they had received glibenclamide, cyclosporin A, or tacrolimus within 1 week of screening. Selective phosphodiesterase inhibitors and endothelin receptor antagonists other than bosentan were not allowed during the study. Disease-modifying antirheumatic drugs (DMARDs) were allowed provided the patient had been stable on treatment for 3 months prior to bosentan initiation.

acrolimus within 1 week of screening. Selective phosphodiesterase inhibitors and endothelin receptor antagonists other than bosentan were not allowed during the study. Disease-modifying antirheumatic drugs (DMARDs) were allowed provided the patient had been stable on treatment for 3 months prior to bosentan initiation. The study was conducted according to the most recent amendments to the Declaration of Helsinki, and in adherence to Good Clinical Practice guidelines. Local institutional review boards or independent ethics committees approved the protocol. Written informed consent was obtained from all patients. Study design and procedures The study was a prospective single-arm trial and was conducted in 23 centres in 8 European countries. Patients received bosentan 62.5 mg twice a day for 4 weeks followed by the 125 mg twice a day target dose for 44 weeks, in addition to stable antirheumatic treatment. Patients who did not tolerate the 125 mg twice a day target dose were down titrated to the starting dose. Patients were evaluated on an outpatient basis at baseline and at weeks 4, 8, 16, 24, 36, and 48 or at premature withdrawal. Efficacy assessments included the change from baseline to weeks 16 and 48 in WHO functional class, the time from baseline to clinical worsening (defined as the combined endpoint of death or hospitalisation due to PAH complications, use of epoprostenol or prostacyclin analogues for worsening of PAH, lung transplantation, discontinuation due to worsening of PAH), and the time to death.

and 48 in WHO functional class, the time from baseline to clinical worsening (defined as the combined endpoint of death or hospitalisation due to PAH complications, use of epoprostenol or prostacyclin analogues for worsening of PAH, lung transplantation, discontinuation due to worsening of PAH), and the time to death. Health-related quality of life and disability were evaluated at baseline and week 48, with a generic instrument (the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36))27 and a disease-specific instrument (the scleroderma modified Health Assessment Questionnaire (HAQ)).28 Both instruments have been validated in a variety of chronic diseases including SSc28–32 and SLE.33 34

line and week 48, with a generic instrument (the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36))27 and a disease-specific instrument (the scleroderma modified Health Assessment Questionnaire (HAQ)).28 Both instruments have been validated in a variety of chronic diseases including SSc28–32 and SLE.33 34 In the SF-36 questionnaire, 35 items cover 8 domains of health: physical functioning, role limitation caused by physical functioning, bodily pain, general health perceptions, vitality, social functioning, role limitation caused by emotional problems, and mental health. The patient’s responses are first reported on a scale from 0 to 100 (higher score indicates better health-related quality of life) for each domain following item weighing and additive scaling, these eight scale scores (0–100) are finally transformed (using a linear z-score transformation) to correspond to a mean of 50 and standard deviation of 10 in the 1998 general US population (norm-based scale scores). A 36th item (health transition) asks respondents about any health changes over the past year using five assessment categories, “much better” (category 1), “somewhat better” (category 2), “about the same” (category 3), “somewhat worse” (category 4), and “much worse” (category 5).

US population (norm-based scale scores). A 36th item (health transition) asks respondents about any health changes over the past year using five assessment categories, “much better” (category 1), “somewhat better” (category 2), “about the same” (category 3), “somewhat worse” (category 4), and “much worse” (category 5). The HAQ assessment includes 20 questions in 8 domains of functional activities: dressing, rising, eating, walking, hygiene, reach, grip, and usual activities. The patient’s responses are reported on an ordinal scale from 0 (no disability) to 3 (complete disability). The HAQ disability index is the average score across the eight domains. In addition, six visual analogue scales (VAS) evaluate disease-specific organ system symptoms: pain, digital ulcer, gastrointestinal, vascular, and pulmonary involvement, and overall disease severity, with scores standardised to a continuous scale from 0 (no symptoms) to 3 (worst symptoms). Safety was assessed by the reporting of adverse events up to 1 day after study drug discontinuation and serious adverse events up to 28 days after study drug discontinuation. Liver function tests were performed at monthly intervals. Any marked laboratory abnormality was reported as an adverse event.

The HAQ assessment includes 20 questions in 8 domains of functional activities: dressing, rising, eating, walking, hygiene, reach, grip, and usual activities. The patient’s responses are reported on an ordinal scale from 0 (no disability) to 3 (complete disability). The HAQ disability index is the average score across the eight domains. In addition, six visual analogue scales (VAS) evaluate disease-specific organ system symptoms: pain, digital ulcer, gastrointestinal, vascular, and pulmonary involvement, and overall disease severity, with scores standardised to a continuous scale from 0 (no symptoms) to 3 (worst symptoms). Safety was assessed by the reporting of adverse events up to 1 day after study drug discontinuation and serious adverse events up to 28 days after study drug discontinuation. Liver function tests were performed at monthly intervals. Any marked laboratory abnormality was reported as an adverse event. Statistical methods Sample size selection was empirical for this open, single arm study. Statistical analyses were performed in an exploratory fashion. For numerical endpoints, the change from baseline is presented with 95% two-sided confidence intervals, based on asymptotic normality assumption. For dichotomous endpoints, the proportion of patients is provided with 95% two-sided CI, based on the binomial exact distribution. For time-to-event endpoints, Kaplan–Meier estimates are presented with 95% two-sided CI calculated from Greenwood’s formula. All treated patients were used for the survival analysis.

ption. For dichotomous endpoints, the proportion of patients is provided with 95% two-sided CI, based on the binomial exact distribution. For time-to-event endpoints, Kaplan–Meier estimates are presented with 95% two-sided CI calculated from Greenwood’s formula. All treated patients were used for the survival analysis. Values for missing assessments of WHO functional class, SF-36, and HAQ/VAS were derived by carrying forward the last observed post-baseline assessment. For patients who died, underwent lung transplantation, or discontinued study medication due to worsening of PAH prior to the considered timepoint (week 16 or week 48), the most conservative approach was used in case of missing assessment. The missing assessment was replaced with the worst value out of: (1) the last available post-baseline value observed for this patient, or (2) the worst value observed at the considered timepoint (ie, week 16 or week 48) over all other patients. RESULTS A total of 53 patients were treated over a 1-year period and received bosentan from July 2003 to August 2005. The median exposure to bosentan was 48.6 weeks (range 0.7 to 56.7 weeks). During the study, 17 patients (32%) were prematurely discontinued from study treatment because of an adverse event (n = 14, 26%), sudden death (n = 1, 2%), disease progression (n = 1, 2%), or loss to follow-up (n = 1, 2%). In total, 36 patients out of 53 completed the study.

an was 48.6 weeks (range 0.7 to 56.7 weeks). During the study, 17 patients (32%) were prematurely discontinued from study treatment because of an adverse event (n = 14, 26%), sudden death (n = 1, 2%), disease progression (n = 1, 2%), or loss to follow-up (n = 1, 2%). In total, 36 patients out of 53 completed the study. Patient demographics and disease characteristics Patient demographics and disease characteristics are presented in table 1. The majority of patients were female and the mean (SD) age was 63 (13) years. Of 53 patients, 42 had SSc, 29 patients had limited and 13 patients had diffuse SSc. There were six patients with MCTD and five with SLE. Twelve patients had a history or evidence of lung fibrosis at baseline. In all but one of these cases was FVC greater than 60% of predicted.

and the mean (SD) age was 63 (13) years. Of 53 patients, 42 had SSc, 29 patients had limited and 13 patients had diffuse SSc. There were six patients with MCTD and five with SLE. Twelve patients had a history or evidence of lung fibrosis at baseline. In all but one of these cases was FVC greater than 60% of predicted. Table 1 Demographics and patient characteristics at baseline Parameter Value Male/female, n (%) 9 (17%)/44 (83%) Age, years (mean (SD)) 63 (13) (range 22–79) Weight, kg (mean (SD)) 67 (13) (range 40–99) Caucasian/Asian 51 (96%)/2 (4%) Aetiology, n (%): Limited systemic sclerosis 29 (55%) Diffuse systemic sclerosis 13 (25%) Mixed connective tissue disease 6 (11%) Systemic lupus erythaematosus 5 (9%) Time from CTD diagnosis, weeks (mean (SD)) 431 (503) (range 0–2227) Time from PAH diagnosis, weeks (mean (SD)) 45 (66) (range 1–236) Signs of right heart failure, n (%) 8 (15%) Patients with at least one digital ulcer, n (%) (n = 50) 15 (30%) Right heart catheterisation: Mean pulmonary arterial pressure, mmHg (mean (SD)) 39.5 (12.6) Cardiac index, litres/min/m2 (mean (SD); n = 49) 2.9 (0.9) Mean pulmonary capillary wedge pressure, mmHg (mean (SD); n = 51) 10.1 (4.4) Pulmonary vascular resistance, dyn/s/cm–5 (mean (SD); n = 47) 559.4 (371.5) Total lung capacity, % (mean (SD); n = 46) 80.8 (18.3) Forced vital capacity, % of predicted (mean (SD)) 85.9 (23.6) Concomitant treatment, n (%): Antithrombotic agents* 41 (77%) Antacids/drug for treatment of peptic ulcer and flatulence 40 (75%) Calcium channel blockers 32 (60%) Corticosteroids† 30 (57%) Diuretics 28 (53%) Unless otherwise stated, n = 53.

46) 80.8 (18.3) Forced vital capacity, % of predicted (mean (SD)) 85.9 (23.6) Concomitant treatment, n (%): Antithrombotic agents* 41 (77%) Antacids/drug for treatment of peptic ulcer and flatulence 40 (75%) Calcium channel blockers 32 (60%) Corticosteroids† 30 (57%) Diuretics 28 (53%) Unless otherwise stated, n = 53. *Mostly acenocoumarol, warfarin, or acetylsalicylic acid. †Mostly prednisolone. CTD, connective tissue disease; PAH, pulmonary arterial hypertension. Concomitant corticosteroid therapy The numbers (percentages) of patients treated with prednisone or prednisolone included: 7 (54%) patients with diffuse SSc (mean dose: 12.4 mg; range: 5–40 mg); 12 (41%) patients with limited SSc (mean dose: 10.0 mg; range: 2.5–40 mg); 2 (40%) patients with SLE (mean dose: 11.3 mg, range: 5–30 mg); and 6 (100%) patients with MCTD (mean dose: 9.1 mg, range: 2.5–25 mg). WHO functional class All patients were in WHO functional class III at baseline. At week 16, WHO class improved in 12 out of 51 patients (24%), remained stable in 35 patients (69%), and worsened in 4 patients (8%). At week 48, it improved in 14 out of 51 patients (27%), remained stable in 29 patients (57%), and worsened in 8 patients (16%) (fig 1).

ents were in WHO functional class III at baseline. At week 16, WHO class improved in 12 out of 51 patients (24%), remained stable in 35 patients (69%), and worsened in 4 patients (8%). At week 48, it improved in 14 out of 51 patients (27%), remained stable in 29 patients (57%), and worsened in 8 patients (16%) (fig 1). Figure 1 Improvement/worsening in World Health Organization (WHO) functional class at week 16 and week 48. Two patients at week 16 and three at week 48 were discontinued because of pulmonary arterial hypertension (PAH) worsening/death and were assigned WHO functional class IV at week 16 and week 48, respectively, as per protocol. (n = 51). Confidence intervals (95%) are indicated in brackets. Clinical worsening Summary statistics on clinical worsening are shown in fig 2. The Kaplan–Meier estimate for the absence of clinical worsening was 88% at week 16 and 68% at week 48. Interestingly, there was no clinical worsening among the five patients with SLE at week 48 whereas the estimate for the absence of clinical worsening was 75%, 67%, and 61% for the patients with diffuse SSc, MCTD, and limited SSc, respectively. One patient received intravenous epoprostenol for 10 days during the course of the study.

, there was no clinical worsening among the five patients with SLE at week 48 whereas the estimate for the absence of clinical worsening was 75%, 67%, and 61% for the patients with diffuse SSc, MCTD, and limited SSc, respectively. One patient received intravenous epoprostenol for 10 days during the course of the study. Figure 2 Kaplan–Meier estimates for time to clinical worsening, defined as the combined endpoint of death, hospitalisation due to pulmonary arterial hypertension (PAH) complications, use of epoprostenol or prostacyclin analogues for worsening of PAH, lung transplantation, or discontinuation due to worsening of PAH. The Kaplan–Meier estimate was 88% (95% CI 79–97%) at week 16, and 68% (95% CI 55–82%) at week 48. Survival The Kaplan–Meier estimates for the observed survival are presented in fig 3. Survival was 92% at week 48. Four deaths were reported during the study: (1) sudden death, (2) staphylococcal sepsis as a result of ischemic colitis, (3) hyponatremia as a complication of high dose diuretics, including furosemide, and (4) exacerbated dyspnoea from worsening pulmonary fibrosis. These four patients had been diagnosed with limited SSc (n = 2), diffuse SSc (n = 1), and MCTD (n = 1). No patient with SLE died during the study.

result of ischemic colitis, (3) hyponatremia as a complication of high dose diuretics, including furosemide, and (4) exacerbated dyspnoea from worsening pulmonary fibrosis. These four patients had been diagnosed with limited SSc (n = 2), diffuse SSc (n = 1), and MCTD (n = 1). No patient with SLE died during the study. Figure 3 Kaplan–Meier estimates of survival (deaths occurring after treatment discontinuation for an adverse event were counted as an event). The Kaplan–Meier estimate was 92% (95% CI 85–100%) at week 48. SF-36 and HAQ/VAS scores The SF-36 domain scores decreased minimally from baseline to week 48. The self-evaluated health transition item at week 48 showed slightly more patients reporting improvement than patients reporting deterioration, in contrast to the baseline status. Accordingly, change in mean (SEM) value for this item was –0.83 (0.22) (95% CI –1.27 to –0.39) (table 2). The HAQ scores presented an overall increase at week 48 (table 3). The disability index was 1.17 (0.11) at baseline and tended to increase at week 48 (0.22 (0.11)). A similar mean increase was reported for all VAS scores at week 48, except for the lung score, which was stable on average (table 3).

to –0.39) (table 2). The HAQ scores presented an overall increase at week 48 (table 3). The disability index was 1.17 (0.11) at baseline and tended to increase at week 48 (0.22 (0.11)). A similar mean increase was reported for all VAS scores at week 48, except for the lung score, which was stable on average (table 3). Table 2 Changes from baseline to week 48 in the Short Form Health Survey (SF-36) domain scores and health transition item Baseline Week 48 Change 95% CI Domain scores (norm-based) Physical functioning* 28.76 (1.24) 27.72 (1.48) –1.04 (1.37) –3.79 to 1.71 Role, physical* 27.47 (2.15) 26.19 (1.89) –1.28 (2.25) –5.81 to 3.25 Pain 43.00 (1.55) 41.92 (1.74) –1.08 (1.60) –4.29 to 2.14 General health perception* 33.89 (1.24) 32.42 (1.24) –1.47 (1.35) –4.19 to 1.25 Vitality* 40.63 (1.41) 40.15 (1.47) –0.49 (1.32) –3.15 to 2.17 Social functioning 39.67 (1.81) 38.51 (2.13) –1.16 (2.26) –5.70 to 3.38 Role, emotional† 32.73 (3.03) 30.65 (3.10) –2.07 (3.65) –9.43 to 5.29 Mental health* 45.11 (1.65) 42.94 (1.98) –2.17 (1.63) –5.45 to 1.11 Health transition item 3.81 (0.15) 2.98 (0.18) –0.83 (0.22) –1.27 to –0.39 Health transition item (n) 1: Much better 3 4 2: Somewhat better 2 17 3: About the same 6 9 4: Somewhat worse 26 10 5: Much worse 10 7 Values are mean (SEM). Four patients were discontinued because of pulmonary arterial hypertension (PAH) worsening/death and were assigned the worst value observed over the analysis set at week 48, as per protocol. The health transition item is reporting comparison to 1 year before. A decrease (negative change) in a domain score corresponds to deterioration.

re discontinued because of pulmonary arterial hypertension (PAH) worsening/death and were assigned the worst value observed over the analysis set at week 48, as per protocol. The health transition item is reporting comparison to 1 year before. A decrease (negative change) in a domain score corresponds to deterioration. n = 47, *n = 46, †n = 45.

re discontinued because of pulmonary arterial hypertension (PAH) worsening/death and were assigned the worst value observed over the analysis set at week 48, as per protocol. The health transition item is reporting comparison to 1 year before. A decrease (negative change) in a domain score corresponds to deterioration. n = 47, *n = 46, †n = 45. Table 3 Changes from baseline to week 48 in the Health Assessment Questionnaire (HAQ) and visual analogue scale (VAS) scores on a 0–3 scale Baseline Week 48 Change 95% CI HAQ scores Dressing 1.15 (0.16) 1.49 (0.19) 0.34 (0.16) 0.01 to 0.67 Arising 0.77 (0.13) 1.09 (0.17) 0.32 (0.20) –0.08 to 0.72 Eating 0.81 (0.13) 0.98 (0.17) 0.17 (0.16) –0.16 to 0.50 Walking 1.32 (0.14) 1.60 (0.18) 0.28 (0.16) –0.05 to 0.60 Hygiene 0.98 (0.15) 1.30 (0.19) 0.32 (0.18) –0.04 to 0.68 Reach 1.28 (0.16) 1.62 (0.18) 0.34 (0.16) 0.01 to 0.67 Grip 1.15 (0.17) 1.28 (0.18) 0.13 (0.17) –0.22 to 0.48 Activity 1.94 (0.14) 1.98 (0.17) 0.04 (0.14) –0.23 to 0.32 HAQ disability index 1.17 (0.11) 1.39 (0.14) 0.22 (0.11) –0.01 to 0.44 VAS scores Pain 0.87 (0.11) 1.06 (0.13) 0.19 (0.16) –0.13 to 0.51 Gastrointestinal 0.44 (0.10) 0.67 (0.11) 0.23 (0.11) 0.0 to 0.46 Lung 1.71 (0.11) 1.70 (0.14) –0.02 (0.14) –0.29 to 0.26 Vascular 1.11 (0.13) 1.32 (0.15) 0.21 (0.19) –0.18 to 0.60 Digital ulcer 0.58 (0.12) 0.86 (0.16) 0.28 (0.16) –0.03 to 0.59 Disease 1.39 (0.13) 1.56 (0.14) 0.17 (0.13) –0.10 to 0.44 Values are mean (SEM), n = 47. Five patients were discontinued because of pulmonary arterial hypertension (PAH) worsening/death and were assigned the worst value observed over the analysis set at week 48, as per protocol. A negative change corresponds to an improvement of the HAQ and VAS scores.

17 (0.13) –0.10 to 0.44 Values are mean (SEM), n = 47. Five patients were discontinued because of pulmonary arterial hypertension (PAH) worsening/death and were assigned the worst value observed over the analysis set at week 48, as per protocol. A negative change corresponds to an improvement of the HAQ and VAS scores. Safety The most frequent adverse events (% patients) observed during the study were peripheral oedema (17%), liver enzyme elevations (17%, of which 11% were specified as aminotransferase increases), diarrhoea (13%), exacerbated dyspnoea (13%), and nausea (13%). At least one serious adverse event was reported in 45% of patients (37 events in 24 patients). The most frequent serious adverse events were exacerbated dyspnoea (8%) and pneumonia (8%). All serious adverse events were judged unrelated to study medication by the investigators. Overall, 14 patients (26%) discontinued study medication because of an adverse event or serious adverse event. Discontinuations most often involved exacerbated dyspnoea (6%), general physical health deterioration (6%), and liver enzyme increase (6%). DISCUSSION This is the first large multi-centre prospective single-arm study of survival and quality of life in PAH–CTD confirmed by right heart catheterisation. The results represent an important source of data on survival and quality of life in patients with PAH–CTD treated with bosentan.

Overall, 14 patients (26%) discontinued study medication because of an adverse event or serious adverse event. Discontinuations most often involved exacerbated dyspnoea (6%), general physical health deterioration (6%), and liver enzyme increase (6%). DISCUSSION This is the first large multi-centre prospective single-arm study of survival and quality of life in PAH–CTD confirmed by right heart catheterisation. The results represent an important source of data on survival and quality of life in patients with PAH–CTD treated with bosentan. The survival estimate of 92% with only slight alteration in the patients’ quality of life can be considered as a positive outcome in PAH–CTD, a progressive disease with very poor outcome in the absence of specific therapy. Treatment of PAH with bosentan was indicated in the patients included in this study. Hence, a placebo therapy for 48 weeks would have been unethical. The absence of a placebo group may be considered a limitation of the study, with concern that some of the improvements were due to “placebo” effect, rather than drug efficacy. However, non-subjective parameters, such as survival, cannot be explained by a placebo effect. Previously reported placebo-controlled PAH studies of bosentan or other PAH-specific treatments have not shown any clinically relevant improvements in placebo groups, in fact, they have generally shown a decline.13 14 20 21 35–37 In addition, a historical comparison shows that the observed estimate for absence of clinical worsening at week 16 is similar to the rates reported for the PAH–CTD patients treated with bosentan in the placebo-controlled BREATHE-1 study.21

n placebo groups, in fact, they have generally shown a decline.13 14 20 21 35–37 In addition, a historical comparison shows that the observed estimate for absence of clinical worsening at week 16 is similar to the rates reported for the PAH–CTD patients treated with bosentan in the placebo-controlled BREATHE-1 study.21 The survival results of the study are in line with published data in patients treated with bosentan. In particular, the observed survival rate of 92% is comparable with the 1-year survival of 86% reported by Denton et al23 in a subgroup analysis of 64 patients with PAH–CTD (SSc, SLE, MCTD), who were enrolled in the two bosentan placebo-controlled PAH trials and the open-label extensions. Similarly, Williams et al38 reported a survival at 1 year of 81% in a cohort of 45 patients with PAH related to SSc in class III–IV who were treated with bosentan as first line therapy. In contrast, the survival reported by these authors for the historical control (47 patients treated with basic therapy with (n = 27) or without (n = 20) prostanoids) was 68%.38 Data on the effects of injectable prostanoids and sildenafil on haemodynamics, exercise capacity and symptoms in PAH associated with connective tissue disease have been presented.39 40 However, their effect has not been studied in a specific long-term cohort in this patient population.41 42 The results of the study also suggest that bosentan improved or stabilised the clinical status of most patients with PAH–CTD. In 84% of patients, the WHO functional class improved (27%), or remained unchanged (57%) at week 48.

The survival results of the study are in line with published data in patients treated with bosentan. In particular, the observed survival rate of 92% is comparable with the 1-year survival of 86% reported by Denton et al23 in a subgroup analysis of 64 patients with PAH–CTD (SSc, SLE, MCTD), who were enrolled in the two bosentan placebo-controlled PAH trials and the open-label extensions. Similarly, Williams et al38 reported a survival at 1 year of 81% in a cohort of 45 patients with PAH related to SSc in class III–IV who were treated with bosentan as first line therapy. In contrast, the survival reported by these authors for the historical control (47 patients treated with basic therapy with (n = 27) or without (n = 20) prostanoids) was 68%.38 Data on the effects of injectable prostanoids and sildenafil on haemodynamics, exercise capacity and symptoms in PAH associated with connective tissue disease have been presented.39 40 However, their effect has not been studied in a specific long-term cohort in this patient population.41 42 The results of the study also suggest that bosentan improved or stabilised the clinical status of most patients with PAH–CTD. In 84% of patients, the WHO functional class improved (27%), or remained unchanged (57%) at week 48. The self-evaluated SF-36 health transition item showed more improvements than deteriorations in overall health perception at the end of the study, as well as an improvement in this overall perception compared to the year preceding the study. The decrease in SF-36 domain scores was lower than that generally considered as the minimal clinically important difference.43 The apparent discrepancy between the encouraging results of the SF-36 health transition item and the decrease in the domain scores can be explained by the different nature of these measures. Whereas the health transition item is a general and relative (compared to 1 year ago) measure, the domain scores are specific and absolute (providing a value at the date of assessment) measures.

F-36 health transition item and the decrease in the domain scores can be explained by the different nature of these measures. Whereas the health transition item is a general and relative (compared to 1 year ago) measure, the domain scores are specific and absolute (providing a value at the date of assessment) measures. Patients included in the current study had moderate-to-severe baseline disabilities44 (HAQ disability index  = 1.17 (0.11)). At week 48, the observed mean increases in the HAQ disability index (+0.22, 95% CI –0.01 to 0.44) reached the minimal clinical important difference value reported for patients with SSc (0.10–0.22)45 46 and the VAS scores increased by less than 0.3 units, except for the lung score which was stable. Interpretation of these findings is limited by the absence of a control group. In a different cohort of SSc patients without PAH followed for a mean of 1.8 years, the HAQ disability index increased by 0.4 units and VAS scores by 0.5–0.6 despite state of the art therapy.47 Hence, our findings suggest that bosentan PAH therapy was specifically associated with a stabilisation of the lung score despite other coexisting CTD respiratory manifestations, but, as expected, had less impact on other organ-related quality of life indicators.

VAS scores by 0.5–0.6 despite state of the art therapy.47 Hence, our findings suggest that bosentan PAH therapy was specifically associated with a stabilisation of the lung score despite other coexisting CTD respiratory manifestations, but, as expected, had less impact on other organ-related quality of life indicators. As for any instrument, the HAQ/VAS and SF-36 tools have well known limitations: the patients’ judgments about the extent of their disabilities may show marked individual variation,48 the HAQ/VAS does not capture the psychological distress felt by patients with PAH and/or CTDs,49 50 and neither instrument is specific for PAH. Consequently, they may have limited sensitivity in detecting changes in quality of life resulting from PAH treatment over time.51 New instruments, which were not available at the start of this study, such as the Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR)51 have been developed from qualitative, unstructured interviews with PAH patients and are expected to provide a more accurate assessment of the impact of PAH on quality of life, when validated translations permit their use in international studies.

, such as the Cambridge Pulmonary Hypertension Outcome Review (CAMPHOR)51 have been developed from qualitative, unstructured interviews with PAH patients and are expected to provide a more accurate assessment of the impact of PAH on quality of life, when validated translations permit their use in international studies. The minimum TLC for inclusion in our study was 50% of predicted. Low TLC may reflect interstitial lung disease or chest wall restriction or other cause for low lung volumes such as pleural disease. There was evidence of lung fibrosis in some cases but our experience and previously published data52 53 suggest that outcome is especially poor in CTD cases, particularly SSc, in which there is RHC proven pulmonary hypertension together with lung fibrosis. Hence, the inclusion of these patients would be more likely to have had a negative impact on the overall prognosis in the cohort.

ce and previously published data52 53 suggest that outcome is especially poor in CTD cases, particularly SSc, in which there is RHC proven pulmonary hypertension together with lung fibrosis. Hence, the inclusion of these patients would be more likely to have had a negative impact on the overall prognosis in the cohort. In conclusion, our study suggests that patients with PAH–CTD can be treated with bosentan in centres having PAH in addition to CTD expertise and that bosentan is effective for the treatment of this patient population. Our results suggest that bosentan improves or stabilises the clinical status in the majority of PAH–CTD patients and, most importantly, has a positive impact on survival. In contrast to the observed changes of SF-36 and HAQ scores not related to PAH, the HAQ lung VAS score stabilised and the SF-36 health transition item improved. Altogether, these results are consistent with sustained benefit in this patient population, which is usually characterised by a high morbidity and mortality.

. In contrast to the observed changes of SF-36 and HAQ scores not related to PAH, the HAQ lung VAS score stabilised and the SF-36 health transition item improved. Altogether, these results are consistent with sustained benefit in this patient population, which is usually characterised by a high morbidity and mortality. The authors would like to acknowledge the collaboration and commitment of all the local investigators and their staff: Belgium: Marion Delcroix, UZ Gasthuisberg, Leuven. France: Philippe Humbert, Hôpital Saint-Jacques, Besançon. Roland Jaussaud, Hôpital Robert Debré, Reims. Pascal Roblot, CHRU La Miletrie, Poitiers. Bernard Lorcerie, CHU Bocage, Dijon. Xavier Puechal, Le Mans. Germany: Michael Buslau, Sanitas Alpenklinik Inzell GmbH. Klaus Helmke, Krankenhaus München Bogenhausen, München. Ulf Müller-Ladner, Universitätsklinikum Regensburg, Regensburg. Italy: Francesco Trotta, Università degli Studi di Ferrara, Ferrara. Netherlands: Hendrika Bootsma, Academisch Ziekenhuis Groningen, Groningen. Pieter Van Paassen, Academisch Ziekenhuis Maastricht, Maastricht. Norway: Jan Tore Gran, Rikshospitalet, Oslo. Spain: Ana Prost, Hospital del Mar, Barcelona. Javier Orte, Hospital Ramon y Cajal, Madrid. José Roman, Hospital Dr Peset, Valencia.

. Netherlands: Hendrika Bootsma, Academisch Ziekenhuis Groningen, Groningen. Pieter Van Paassen, Academisch Ziekenhuis Maastricht, Maastricht. Norway: Jan Tore Gran, Rikshospitalet, Oslo. Spain: Ana Prost, Hospital del Mar, Barcelona. Javier Orte, Hospital Ramon y Cajal, Madrid. José Roman, Hospital Dr Peset, Valencia. Funding: The study was supported by an investigational grant from Actelion Pharmaceuticals Ltd. The database was retained by the sponsor, but the investigators had access to the complete database. The statistical analysis was performed by a statistician who is an employee of the sponsoring company and is listed among the authors (AM); the manuscript was reviewed and approved by the academic authors. The academic authors assume full responsibility for the completeness and accuracy of the content of the manuscript. Grant support: This research was supported by Actelion Pharmaceuticals Ltd.

f the sponsoring company and is listed among the authors (AM); the manuscript was reviewed and approved by the academic authors. The academic authors assume full responsibility for the completeness and accuracy of the content of the manuscript. Grant support: This research was supported by Actelion Pharmaceuticals Ltd. Competing interests: CPD has been a consultant to, or received research grants from the following companies: Genzyme Corporation, Actelion Pharmaceuticals, Aspreva Pharmaceuticals, Encysive Corporation, DigNa Pharmaceuticals. JJEP, H-HP, AG, GR, SDeV have no significant industry affiliation (all under 10 000 USD). AB has received speaking fees or consultancy fees from GSK, Actelion Pharmaceuticals and Pfizer less than 2K US$. AB’s employing Institution is involved in contract research for GSK, Actelion Pharmaceuticals, Pfizer, United Therapeutics, Encysive, Myogen, Merck. AB’s employing Institution’s research foundation has received unrestricted educational grants from GSK and Actelion Pharmaceuticals. FHJvdH is a consultant for Bristol Myers Squibb (BMS) and Abbott. LG is a consultant for Actelion Pharmaceuticals (advisory boards) and a member of the scientific council of Actelion France. AM, MD and OB are employees of Actelion Pharmaceuticals.

Ankylosing spondylitis (AS) typically strikes young adults, with the burden of disease attributable primarily to the resulting functional disability.1 The disease course varies widely. Some patients experience sacroiliitis alone, while others experience rapid progression to end-stage fusion of the spine, or total spinal ankylosis (TSA).2 Patients who develop TSA (ie, bamboo spine) experience significantly more functional impairment and are less likely to be employed compared with other patients with AS.3 In addition to substantial functional disability, patients with TSA may experience a more debilitating disease course. The fragility of the rigid spinal column increases the risk of spinal fractures and possible neurological sequelae, and spinal deformities may contribute to respiratory and other difficulties.1 In contrast to pre-existing concepts, patients with TSA may continue to have signs and symptoms of active AS, which are insufficiently responsive to non-steroidal anti-inflammatory drugs (NSAIDs).

fractures and possible neurological sequelae, and spinal deformities may contribute to respiratory and other difficulties.1 In contrast to pre-existing concepts, patients with TSA may continue to have signs and symptoms of active AS, which are insufficiently responsive to non-steroidal anti-inflammatory drugs (NSAIDs). Patients with TSA are typically excluded from participation in randomised controlled trials of therapeutic agents for AS. For example, the randomised controlled trials of the tumour necrosis factor (TNF) antagonists etanercept and infliximab have excluded AS patients with TSA.4 5 The Adalimumab Trial Evaluating Long-term Efficacy and Safety for AS (ATLAS) was the first large randomised controlled trial of a TNF antagonist in patients with active AS that permitted patients diagnosed with TSA.6 Our objective was to evaluate the long-term safety and efficacy of adalimumab in patients with TSA who had participated in ATLAS.

ating Long-term Efficacy and Safety for AS (ATLAS) was the first large randomised controlled trial of a TNF antagonist in patients with active AS that permitted patients diagnosed with TSA.6 Our objective was to evaluate the long-term safety and efficacy of adalimumab in patients with TSA who had participated in ATLAS. PATIENTS AND METHODS Patients ATLAS has been described in the published report of the 24-week, double-blind results.6 Adults with AS based on the modified New York criteria7 who had active disease were recruited for the study. ATLAS was designed with an a priori limit on enrolment of patients with TSA of 10%. A diagnosis of TSA was based on the investigators’ assessments of lateral radiographs of the cervical and lumbar spine and lateral views of chest radiographs. All enrolled patients had an inadequate response or intolerance of one or more NSAIDs, as defined by the investigators. Also, patients who had failed therapy with one or more disease-modifying antirheumatic drugs were allowed to participate. Each of the 43 study centres obtained independent ethics committee approval, and ATLAS was conducted in accordance with the Declaration of Helsinki. Compliance with local laws and customs was assured by investigators at the 43 centres in Europe (Belgium, France, Germany, Italy, The Netherlands, Spain, Sweden, and the United Kingdom) and the USA. Written informed consent was obtained from each patient before any study-related procedures were initiated.