Article Text

Download PDFPDF

Extended report
Calprotectin (a major S100 leucocyte protein) predicts 10-year radiographic progression in patients with rheumatoid arthritis
  1. H Berner Hammer1,
  2. S Ødegård1,
  3. S W Syversen1,
  4. R Landewé2,
  5. D van der Heijde1,3,
  6. T Uhlig1,
  7. P Mowinckel1,
  8. T K Kvien1,4
  1. 1
    Department of Rheumatology, Diakonhjemmet Hospital, Oslo, Norway
  2. 2
    Department of Rheumatology, University Hospital Maastricht, Maastricht, The Netherlands
  3. 3
    Leiden University Medical Centre, Leiden, The Netherlands
  4. 4
    Faculty of Medicine, University of Oslo, Oslo, Norway
  1. Correspondence to Dr H Berner Hammer, Department of Rheumatology, Diakonhjemmet Hospital, Box 23, Vinderen, N-0319 Oslo, Norway; hbham{at}online.no

Abstract

Background: Plasma levels of calprotectin, a major S100 leucocyte protein, are cross-sectionally associated with clinical and laboratory markers of inflammation and with radiographic damage in rheumatoid arthritis (RA). High amounts of calprotectin are found in synovial fluid from patients with RA.

Objective: To examine whether calprotectin might be an independent predictor of joint destruction over time.

Methods: 124 patients with RA were assessed at baseline and after 10 years with inflammatory markers (calprotectin, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR)), serological variables (antibodies to cyclic citrullinated peptide (anti-CCP), IgA rheumatoid factor (RF) and IgM RF) and radiographic and clinical assessments of joint damage (hand radiographs and Rheumatoid Arthritis Articular Damage (RAAD) score). Progression of radiographic damage was assessed according to the van der Heijde modified Sharp score.

Results: At both examinations the highest calprotectin levels were found in patients positive for anti-CCP, IgA and IgM RF. Calprotectin had moderate to good correlations with inflammatory and serological markers (r = 0.41–0.67). Patients with normal baseline calprotectin levels had a lower degree of joint damage. High univariate associations were found between baseline calprotectin levels and progression in the Sharp score as well as the RAAD score. Baseline calprotectin was independently associated with progression in the Sharp score and with the RAAD score in multiple linear regression analyses, including baseline levels of CRP, ESR, anti-CCP in addition to demographic variables.

Conclusion: Calprotectin was an independent predictor of clinical and radiographic joint damage after 10 years. These findings support the proposal that calprotectin may be a prognostic biomarker for erosive disease in patients with RA.

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Rheumatoid arthritis (RA) is an inflammatory joint disease in which longstanding synovitis causes radiographic changes that correlate with disability.1 Prognostic markers are important for the identification of patients with a potentially aggressive disease course in order to tailor treatment individually. Joint inflammation is the key feature in RA, and a biomarker reflecting the amount of inflammation as well as the prognosis of RA with respect to joint damage would be clinically helpful.

Calprotectin is a major leucocyte protein, constituting 40–60% of the cytosolic protein in neutrophil granulocytes as well as being a major monocyte/macrophage protein.2 3 4 It is released during activation and turnover of these leucocytes,5 6 7 and high plasma or serum concentrations of calprotectin have been found in patients with several inflammatory diseases.8 9 10 11 12 13 Calprotectin is one of the major leucocyte S100 proteins having both calcium binding and antimicrobial properties.5 7 It has also been called the L1 protein,6 MRP8/MRP14,14 S100A8/A9,7 calgranulin A and B15 or cystic fibrosis antigen.16 Calprotectin is specifically released during interaction of monocytes with inflammatory activated endothelium, probably at sites of local inflammation,17 and early recruited phagocytes expressing calprotectin have been found in the sublining layer of inflamed synovial tissue.18 The expression of calprotectin is strongest at the cartilage–pannus junction, which is the primary site of cartilage destruction and bone erosion in RA.18 In synovial fluid, high concentrations of calprotectin were identified in patients with RA and low concentrations in osteoarthritic patients,19 20 and a good correlation was found between the calprotectin levels in plasma and synovial fluid.11 19 Several studies have demonstrated associations between calprotectin and the level of inflammatory activity in different rheumatic joint diseases.8 9 10 11 12 13

Calprotectin was recently found to be cross sectionally associated with radiographic damage in patients with RA.21 The objective of this study was to examine the associations between calprotectin and factors known to be related to erosive disease in RA, and to analyse whether plasma concentrations of calprotectin from patients with RA of short disease duration could predict 10-year progression of joint damage.

Patients and methods

Patients

From a cohort of 238 patients with RA22 of short duration (the EURIDISS cohort,1 23 24 disease duration maximum 4 years) a total of 124 patients had calprotectin and radiographic assessments at baseline and 10-year follow-up and were included in these analyses. At baseline, these 124 patients had a mean (SD) age of 50.9 (12.6) years, a disease duration of 2.2 (1.2) years, and 94 (76%) were women. The patients underwent at baseline and at the 10-year follow-up a comprehensive 1-day examination, including laboratory, clinical and radiographic assessments.1 At baseline/10-year follow-up, 31 (25%)/44 (35%) of the patients used prednisolone with a median (interquartile range (IQR)) dose of 5 (5–8) mg/5 (5–7) mg, 65 (52%)/61 (49%) were receiving disease-modifying antirheumatic drugs (DMARDs) (methotrexate was used by 76 (61%)/41 (33%)), and 13 patients (10.5%) used anti-tumour necrosis factor (TNF) drugs at follow-up (maximal duration of treatment 1.5 years). The patients gave written consent according to the Declaration of Helsinki, and the study was approved by the local ethics committee.

Laboratory examinations

EDTA plasma from baseline and follow-up was frozen at −70°C and all samples were analysed for calprotectin at the same time by an ELISA, as described previously.21 25 Commercial ELISA kits are now available for calprotectin measurements. Assessments of calprotectin in plasma have shown coefficients of variation of 5% within assays and 13% between assays,25 and normal plasma levels are ⩽0.91 mg/l. The erythrocyte sedimentation rate (ESR) was assessed on the day of examination by the Westergren method, with normal levels ⩽20 mm/h. Frozen serum (−70°C) from baseline and follow-up was used for C-reactive protein (CRP) and serological assessments. CRP was measured by phyCardioPhase high-sensitive CRP nephelometry (Dade Behring, Milton Keynes, UK), with the lowest detectable limit 0.15 mg/l (normal levels <4 mg/l). Antibodies to cyclic citrullinated peptide (anti-CCP) were analysed by a second-generation ELISA (INOVA Diagnostics, San Diego, California, USA) and considered positive at a cut-off value of 25 U/ml (given concentrations 1–251 U/ml).26 IgM rheumatoid factor (RF) and IgA RF were measured by an ELISA, with a range of 2–300 U/ml, and cut-off point set at 25 U/ml.26

Clinical examinations

All patients were examined at follow-up by one rheumatologist (SØ) who performed a comprehensive clinical examination, including Health Assessment Questionnaire (HAQ) and the Rheumatoid Arthritis Articular Damage (RAAD) score. The RAAD is based on a clinical evaluation of joint damage in 35 large and small joints (score range 0–70),27 and has been shown to be a feasible method for assessment of clinical long-term articular damage in large RA populations.1 The RAAD score was not performed at baseline in this study.

Radiographic examinations

Hand radiographs were obtained at baseline and follow-up and were scored blinded for clinical data according to the van der Heijde modified Sharp score28 29 with known time order under the responsibility of two of the coauthors (DvdH and RL).1 The total score (sum of erosions and joint space narrowing) for hands (range 0–280) was used in the analyses. The radiographic progression was calculated as the difference between the total Sharp scores at the 10-year follow-up and baseline, and this Sharp progression score (SharpProgScore) was used in the further analyses. A SharpProgScore >2 was used as a cut-off point for radiographic progression.

Statistical methods

All the statistical analyses were performed using SPSS 14 (SPSS, USA). The variables had a skewed distribution, but their differences and the residuals from the regression analysis were found to be normally distributed. Comparisons between baseline and follow-up were performed by the Wilcoxon signed rank test, and the Mann–Whitney test was used for group comparisons. Correlations were performed by Spearman correlation analysis. Univariate regression analyses examined for associations, and multiple, linear regression analyses were performed to analyse for significant and independent contributions to the measures of joint damage. Receiver operating characteristic (ROC) analyses were performed to find the optimal cut-off levels for calprotectin, CRP and ESR to discriminate between patients with (>2 units) or without an increased SharpProgScore. These cut-off levels were used in the analyses for assessing the sensitivity and specificity as well as likelihood ratios (LRs) for the inflammatory variable to predict joint damage. Cumulative probability plots were used to display radiographic progression and RAAD across patients with different baseline levels of the inflammatory markers, and statistical group comparisons were performed by Kruskal–Wallis tests. For all the analyses the level of significance was set at 0.05.

Results

Table 1 shows the values of laboratory and radiographic assessments at baseline and follow-up as well as the RAAD score at follow-up. Inflammatory markers decreased (p<0.05), whereas radiographic damage increased (p<0.001) during the 10-year follow-up period.

Table 1

Median (interquartile range) levels of inflammatory and serological variables, HAQ and radiographic and clinical joint damage at baseline and at 10-year follow-up in 124 patients with RA

The percentages of patients positive for anti-CCP, IgA RF or IgM RF were at baseline/follow-up 60/54, 35/30 or 44/53, respectively. Patients positive for anti-CCP, IgA RF or IgM RF had higher calprotectin levels both at baseline and at follow-up than patients who were negative for these serological markers (p<0.001). Similarly, the levels of CRP and ESR were highest in patients who were positive as compared with those who were negative for the serological markers (p<0.05) (data not shown). At baseline/follow-up the correlation coefficients between calprotectin and CRP was 0.59/0.56 (p<0.001) and between calprotectin and ESR 0.67/0.51 (p<0.001). In addition, moderate correlations were found at baseline/follow-up between calprotectin and anti-CCP (0.41/0.51), IgA RF (0.43/0.59) and IgM RF (0.44/0.65), all with p<0.001. Significant correlations were also shown between RAAD and both SharpProgScore (r = 0.67, p<0.001) and total Sharp score at follow-up (r = 0.70, p<0.001).

In univariate regression analyses, the baseline levels of calprotectin, CRP, ESR and anti-CCP all had significant associations with the SharpProgScore and RAAD score (table 2). Of the inflammatory variables, only baseline calprotectin level remained significantly associated with the SharpProgScore (p = 0.045) and RAAD score (p = 0.012) in multiple linear regression analyses with adjustments for baseline levels of CRP, ESR and anti-CCP as well as sex, age and disease duration (table 3). The magnitude of the regression coefficient for calprotectin was only marginally influenced by adding potentially confounding variables such as CRP, ESR and anti-CCP. In contrast, the regression coefficients for CRP and ESR were markedly decreased when adding calprotectin in the linear regression equation. However, the effect of anti-CCP was entirely independent. When the total Sharp score at baseline was added to the multiple linear regression analyses, calprotectin tended to be associated with SharpProgScore (p = 0.084), while a significant association was found with RAAD (p = 0.025). Neither CRP nor ESR had significant associations with the two outcome variables for joint damage in corresponding analyses.

Table 2

Univariate regression analyses of associations between baseline levels of laboratory markers, demographic variables, disease duration and Sharp Progression Score (SharpProgScore) and the Rheumatoid Arthritis Articular Damage (RAAD) score as the dependent variables

Table 3

Multiple, linear regression analyses with Sharp Progression Score (SharpProgScore) and the Rheumatoid Arthritis Articular Damage (RAAD) score as the dependent variables and the baseline levels of laboratory markers, demographic variables and disease duration as independent variables

The ROC analysis identified a calprotectin level of 1.86 mg/l as the optimal cut-off point to discriminate between those with or without progression of radiographic joint damage. The sensitivity and specificity of this calprotectin threshold for detecting progression of total Sharp score was 69% and 66%, respectively, with an LR of 2.03 with increased calprotectin levels and 0.47 for low calprotectin levels. The ability of CRP and ESR to discriminate between patients developing or not developing radiographic progression after 10 years was similar (CRP: optimal cut-off value 2.33 mg/l, sensitivity/specificity 62%/77%, positive/negative LR 2.8/0.73 and ESR: optimal cut-off value 20 mm/h, sensitivity/specificity 79%/57%, positive/negative LR 1.83/0.36).

Figure 1 shows the SharpProgScore and RAAD score across increasing baseline levels of calprotectin (normal levels (⩽0.910 mg/l, n = 24), elevated levels (0.911–2.544 mg/l, n = 50) and highly elevated levels (2.545–42.000 mg/l, n = 50)). The scores were significantly different between the groups (p<0.001). Comparison of groups categorised according to normal, elevated or highly elevated (the last two groups of equal size) baseline CRP or ESR levels also showed significant groups differences for both the SharpProgScore, p<0.05/p<0.01, and RAAD, p<0.01/p<0.05, respectively (data not shown).

Figure 1

Cumulative probability plots for the associations between patients with normal (⩽0.910 mg/l, n = 24), elevated (0.911–2.544 mg/l, n = 50) or highly elevated (2.545–42.000 mg/l, n = 50) baseline calprotectin concentrations and the corresponding Sharp Progression Score (A) and Rheumatoid Arthritis Articular Damage score (B).

Discussion

Prediction of joint damage is considered important to optimise the individualised medical treatment of patients with RA. To our knowledge, this longitudinal study is the first to show that calprotectin consistently predicted two measures of joint damage (radiographic and clinical). Calprotectin is released from activated leucocytes during inflammation, which makes this association biologically plausible. In patients with RA, high amounts of calprotectin have been found in the synovial fluid17 19 20 30 and in the synovial tissue, primarily close to the cartilage.18

Several studies on patients with arthritis have shown high associations between plasma or serum levels of calprotectin and clinical measures of joint inflammation,7 8 9 10 11 12 13 and we have previously reported cross-sectional associations between calprotectin and joint damage.21 However, a prospective study of a smaller number of patients with RA with a wide range of disease duration did not identify calprotectin as a predictor of radiographic damage.31 This negative finding might be explained by a ceiling effect of the radiographic damage, since the median disease duration in that study was 7.8 years (interquartile range 2.3–19.4 years) at baseline. Only patients with short disease duration were included in our cohort, and the total Sharp score was low at baseline. Thus, this study had a better design to analyse calprotectin as a prognostic marker for joint damage.

Previous studies have identified both anti-CCP and RF to be associated with radiographic damage,32 33 and calprotectin has been found to have significant associations with the levels of anti-CCP and IgA/IgM RF.34 In RA, the inflamed synovium and the synovial fluid with increased content of leucocytes are possibly the main sources of calprotectin detected in plasma. The serological markers may also primarily derive from the synovium, which may explain the high associations. Importantly, calprotectin remained as an independent predictor of joint damage in the multivariate analyses that adjusted for the levels of anti-CCP.

High correlations were found between calprotectin and both CRP and ESR, and this observation is consistent with results from several studies in patients with RA8 10 21 35 or other rheumatic diseases.9 11 12 13 36 Calprotectin has also been found to be associated with other acute phase proteins such as fibrinogen and α1 antitrypsin.35 36 Thus, the protein may behave like an acute phase protein. However, an improved face validity of calprotectin as a marker is that the protein is released from activated leucocytes which derive mainly from the inflamed synovium in patients with RA. In contrast, the acute phase proteins are primarily produced in hepatocytes after induction by interleukins released during inflammation. Thus, calprotectin is a marker that differs from acute phase proteins by directly reflecting the amount of activated leucocytes in the inflamed joints. Biological drugs against the cytokines TNFα, interleukin (IL)1 or IL6 will, in addition to a reduction of the inflammatory activity, reduce the circulating cytokines and thus the concentration of the acute phase proteins. Calprotectin, however, is not influenced by these drugs, and will only reflect the leucocyte activation during inflammation. In addition, in clinical practice some patients have normal or low levels of CRP and ESR despite extensive arthritis. Calprotectin could in these patients be a useful inflammatory marker.

Work is continuing within OMERACT to develop validation criteria for a biomarker which will reflect structural damage outcomes, assay reproducibility and reliability, and different sources of variation in calprotectin levels and biomarker stability have to be further investigated.37 Associations with structural damage end points have to be confirmed in other prospective studies. However, the consistent finding across two end points reflecting damage adds to the robustness of our results. These findings, and the plausible biological explanation for the association between a major granulocyte protein and joint damage, suggest that calprotectin is a promising prognostic marker in RA.

REFERENCES

Footnotes

  • Funding SØ was supported by Norwegian Women’s Public Health Association and Norwegian Foundation for Health and Rehabilitation.

  • Competing interests None.

  • Ethics approval Approval from the regional ethics committee, Helse Sør-Øst.

  • Patient consent Patient consent received.