Association of C-reactive protein with high disease activity in systemic sclerosis: Results from the Canadian Scleroderma Research Group




This study was performed to determine the prevalence of elevated C-reactive protein (CRP) levels and the significance of CRP in clinical parameters in systemic sclerosis (SSc; scleroderma) patients.


Canadian Scleroderma Research Group data were used. Statistical comparisons were made for CRP levels ≤8 mg/liter versus >8 mg/liter, early (≤3 years from first non–Raynaud's phenomenon symptom) versus late SSc, and diffuse cutaneous SSc (dcSSc) versus limited cutaneous SSc (lcSSc). A survival analysis was analyzed between patients with normal versus elevated CRP levels.


A total of 1,043 patients (mean ± SD age 55.4 ± 12.1 years, mean ± SD disease duration of 11.0 ± 9.5 years) were analyzed; elevation of CRP level and erythrocyte sedimentation rate (ESR; >20 mm/hour) occurred in 25.7% and 38.2%, respectively. Mean ± SD baseline CRP level in dcSSc (11.98 ± 25.41 mg/liter) was higher than in lcSSc (8.15 ± 16.09 mg/liter; P = 0.016). SSc patients with an early disease duration had a higher mean ± SD CRP level (12.89 ± 28.13 mg/liter) than those with a late disease duration (8.60 ± 17.06 mg/liter; P = 0.041). Although not consistent in all subsets, CRP was significantly associated (P < 0.01) with ESR, modified Rodnan skin score (MRSS), worse pulmonary function parameters, disease activity, damage, and Health Assessment Questionnaire. CRP level seemed to normalize in many SSc patients over time. Total lung capacity <80% predicted, MRSS, and serum creatinine were predictors of elevated CRP levels in SSc (odds ratio [OR] 2.76 [95% confidence interval (95% CI) 1.73–4.40], P = 0.0001; OR 1.03 [95% CI 1.01–1.05], P = 0.005; and OR 1.005 [95% CI 1.001–1.010], P = 0.02, respectively). Survival for patients with elevated CRP levels was less than for patients with normal CRP levels (P = 0.001).


CRP level is elevated in one-quarter of SSc patients, especially early disease. It is correlated with disease activity, severity, poor pulmonary function, and shorter survival.


Systemic sclerosis (SSc; scleroderma) is an autoimmune connective tissue disease that affects more women than men, often beginning between ages 40 and 55 years (1). There are 2 subtypes of SSc, i.e., diffuse cutaneous SSc (dcSSc) and limited cutaneous SSc (lcSSc), according to the extent of cutaneous involvement (2); dcSSc has more extensive skin involvement, including the trunk or proximal extremities, and is associated with more internal organ involvement such as significant interstitial lung disease (ILD), cardiomyopathy, and scleroderma renal crisis (SRC) and with increased mortality (3–18).

Both erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP) are acute-phase reactants. ESR is a biomarker for increased morbidity and mortality in SSc (7, 12, 15, 19–21). However, data are limited about the impact of CRP associations with SSc activity and severity. CRP is produced by hepatocytes upon stimulation by interleukin-6 (IL-6) (22), and has been utilized as a marker of infection and inflammation (23). CRP level was elevated in 35–54% of Japanese patients with SSc, but the sample size was small (40 patients in each study) (24, 25). It appears to be elevated more commonly in dcSSc (26). CRP is associated with worse survival at levels >20 mg/dl (27). CRP level elevations are also associated with renal vascular resistance as detected by Doppler ultrasound (28) and inflammatory arthritis (29).

The current study was done to determine the prevalence of elevated CRP levels in a large multicenter cohort; to assess the associations of CRP with respect to clinical factors, including organ involvement, dcSSc and lcSSc subsets, and early disease versus later disease; to analyze changes in CRP level over time; to determine factors contributing to elevated CRP levels; and to determine the predictive value of CRP on survival.

Significance & Innovations

  • Although most patients with systemic sclerosis (SSc; scleroderma) have normal and stable C-reactive protein (CRP) concentrations, CRP level is elevated in one-quarter of scleroderma patients.

  • Elevation of CRP level is associated with high disease activity, severity, poor pulmonary function, and shorter survival in scleroderma.

  • Elevation of CRP level is more common in the diffuse cutaneous SSc subset, especially with early disease duration (≤3 years). Total lung capacity <80% predicted, modified Rodnan skin score, and serum creatinine are predictors of elevated CRP levels in scleroderma.


Study population.

The Canadian Scleroderma Research Group (CSRG) is a multicenter prospective adult SSc national registry across Canada (investigators of the CSRG are shown in Appendix A). Clinical and laboratory data (including annual complete blood cell count, ESR, CRP, creatinine, and creatine kinase) from adult SSc patients are collected annually in a comprehensive database with more than 1,000 variables that has been described elsewhere (30). Adult SSc patients within the database and enrolled between August 18, 2004 and April 27, 2010 who fulfilled the preliminary American College of Rheumatology criteria for the classification of SSc (scleroderma) (31), or were diagnosed by their rheumatologist with SSc, were analyzed. Patients were followed for a maximum of 6 annual visits.

Clinical and laboratory parameters.

CRP level was measured in milligrams per liter, and was considered high when >8 mg/liter. This value was the 75th percentile of the distribution of CRP in the CSRG cohort and the 95th percentile of sex and age matched for CRP in the healthy population (32). ESR was measured by the Westergren method (mm/hour) and was considered high when >20 mm/hour. Mean CRP concentrations (mg/liter) and the prevalence of elevated CRP levels in each disease subset were studied and associations with clinical parameters determined, including skin score measured by the modified Rodnan skin score (MRSS), estimated systolic pulmonary artery pressure (PAP) measured by transthoracic echocardiograms, pulmonary function tests (PFTs), ILD determined by chest radiographs and high-resolution computerized tomography (HRCT) of the chest, incidence of SRC, existence of arthritis determined by either tender joint count (TJC) or swollen joint count (SJC), tendon friction rubs, Scleroderma Disease Activity Score (SDAS) adapted from the European Scleroderma Study Group (EScSG) whole series activity index (33–36), Scleroderma Disease Severity Score (SDSS) adapted from Medsger et al (37), and Health Assessment Questionnaire (HAQ) disability index (DI).

Subsets were analyzed as lcSSc versus dcSSc (2), early (≤3 years from first non–Raynaud's phenomenon [RP] symptom to baseline visit) versus late (>3 years) disease duration, and receiving corticosteroids versus no corticosteroids. Exploration was done for serial annual CRP level measurements to determine how often the results changed, and also studying patients who were receiving steroids at one year and were off at another to determine if this altered CRP values. Inflammatory arthritis was measured as the existence of either SJC ≥4 or TJC ≥8, since SJC 4 and TJC 8 were the 95th percentile of the number of swollen and tender joints, respectively, in our CRSG cohort. SRC was considered as “ever,” since very few patients had new-onset SRC within the database. The relationship between patients with normal (≤8 mg/liter) and elevated (>8 mg/liter) CRP levels was also studied comparing age, sex, body mass index (BMI), existence of moderate to large pericardial effusion measured by transthoracic echocardiograms, proximal extremity girdle muscle and/or neck weakness, presence of tendon friction rubs, active digital ulcers, digital gangrene/necrosis, number of fingers with ulcers, hemoglobin, serum creatinine (μmoles/liter), serum creatine kinase (units/liter), presence of proteinuria on urinalysis, antinuclear antibody, anticentromere antibody, anti–topoisomerase I (Scl-70), steroid use, and use of any immunomodulator (such as methotrexate, azathioprine, cyclophosphamide, and mycophenolate mofetil). Factors that had significant correlations were selected for use in a predictive regression model. Survival curves for patients with entry into CSRG levels of normal versus high CRP were calculated.

Statistical analysis.

Database management and statistical analyses were performed using IBM SPSS statistics software, version 19. Frequencies were expressed as the percentage and compared between groups by Fisher's exact test. Normally distributed continuous variables were expressed as the mean ± SD and if data were skewed, medians and percentiles were provided. Mean and median differences were analyzed using independent t-tests or analysis of variance, Mann-Whitney U test, or Kruskal-Wallis test, as appropriate. Relationships between CRP and continuous/categorical clinical parameters were assessed by Spearman's correlation coefficients (rho). Statistically significant univariate analyses comparing parameters with CRP were used in multivariate logistic regression with CRP level status (normal versus elevated) as the dependent variable. Survival curves for patients were constructed using Kaplan-Meier survival estimates and log rank (Mantel-Cox) tests.


There were 1,145 SSc patients included from August 18, 2004 to April 27, 2010, of whom 1,043 patients had available data to analyze according to their disease subsets (either lcSSc or dcSSc). Eighty-six percent (86.1%) were women, with a mean age of 55.4 years and a mean disease duration of 11.0 years from onset of first non-RP symptoms to baseline visit. Thirty-eight percent had dcSSc, of whom 117 (10.6% of total) had early dcSSc (≤3 years since onset). Baseline characteristics are shown in Table 1.

Table 1. Baseline characteristics of patients of the Canadian Scleroderma Research Group*
 All patientslcSScdcSScEarly dcSSc (≤3 years)Late dcSSc (>3 years)
  • *

    Values are the percentage unless otherwise indicated. Percentages given are out of the number of patients with data available. lcSSc = limited cutaneous systemic sclerosis (scleroderma); dcSSc = diffuse cutaneous SSc; RP = Raynaud's phenomenon; MRSS = modified Rodnan skin score; RVSP = right ventricular systolic pressure; HRCT = high-resolution computed tomography; TLC = total lung capacity; FVC = forced vital capacity; DLCO = diffusing capacity of carbon monoxide; SJC = swollen joint count; TJC = tender joint count; SDAS = Scleroderma Disease Activity Score; SDSS = Scleroderma Disease Severity Score; HAQ = Health Assessment Questionnaire; DI = disability index; CRP = C-reactive protein; ESR = erythrocyte sedimentation rate.

  • DLCO <70% predicted, FVC >80% predicted, and TLC >80% predicted.

Total group, no. (%)1,043 (100)647 (62)396 (38)117 (10.6)269 (25.6)
Age, mean ± SD years55.39 ± 12.1357.06 ± 11.7752.64 ± 12.3151.2 ± 13.2153.34 ± 11.77
Women, no. (%)986 (86.1)575 (88.9)319 (80.6)90 (76.9)221 (82.2)
Disease duration (onset of RP symptoms to baseline visit), mean ± SD years14.7 ± 12.3317.04 ± 12.7010.88 ± 10.693.05 ± 4.8814.22 ± 10.83
Disease duration (onset of non-RP symptoms to baseline visit), mean ± SD years10.99 ± 9.5012.37 ± 10.069.02 ± 8.451.46 ± 0.7412.30 ± 8.16
MRSS, mean ± SD10.04 ± 9.545.94 ± 5.0718.34 ± 10.2922.20 ± 10.1416.59 ± 9.85
MRSS >1034.614.473.486.267.8
Elevated pulmonary pressure on echocardiogram (RVSP >40 mm Hg)25.725.228.124.730.2
Lung fibrosis (chest radiograph)21.519.026.723.227.8
HRCT performed, no. (%)290 (27.2)156 (25.7)111 (29.6)39 (34.8)69 (27.2)
 Lung fibrosis in those with HRCT31.426.
 HRCT chest: ground-glass appearance36.030.346.737.149.3
 HRCT chest: fibrotic interstitial change52.750.759.648.664.2
 HRCT chest: honeycomb appearance21.819.927.014.333.3
TLC <80% predicted23.020.029.725.031.2
FVC <80% predicted27.924.534.028.836.0
DLCO <75% predicted5755.859.952.062.5
Isolated low DLCO49.451.744.346.343.2
Scleroderma renal crisis ever4.
Arthritis (SJC ≥4 or TJC ≥8)8.67.610.211.210.2
Tendon friction rub14.76.330.039.724.1
SDAS score (range 0–10), mean ± SD2.09 ± 1.61.76 ± 1.342.71 ± 1.773.28 ± 1.912.43 ± 1.62
SDSS score (range 0–36), mean ± SD8.60 ± 3.977.54 ± 3.0710.72 ± 4.3110.59 ± 4.1510.74 ± 4.33
HAQ DI score (range 0–3), mean ± SD0.81 ± 0.710.67 ± 0.651.07 ± 0.721.16 ± 0.731.02 ± 0.71
Laboratory studies     
 Hemoglobin, mean ± SD gm/dl130.03 ± 15.27131.04 ± 13.52128.30 ± 16.88124.38 ± 19.05130.15 ± 15.41
 Serum creatinine, mean ± SD μmoles/liter81.07 ± 51.3178.95 ± 32.8586.01 ± 73.9994.07 ± 82.3080.72 ± 63.07
 Creatine kinase, mean ± SD units/liter95.57 ± 86.3190.37 ± 69.77108.36 ± 112.21131.39 ± 160.3198.38 ± 82.60
 CRP level (normal value ≤8), mean ± SD mg/liter9.54 ± 19.988.15 ± 16.0911.98 ± 25.4117.18 ± 34.739.77 ± 20.42
 Elevated CRP level >8 mg/liter25.723.231.141.526.2
 ESR (normal value <20), mean ± SD mm/hour21.90 ± 21.0320.24 ± 19.4724.70 ± 23.5728.72 ± 27.9022.70 ± 20.92
 Elevated ESR >20 mm/hour38.235.243.644.242.6
Serology studies     
 Antinuclear antibody95.
 Anticentromere antibody34.646.211.34.913.9
 Anti–topoisomerase I (Scl-70)15.011.423.423.323.1
 Anti–RNA polymerase III14.76.828.940.424.4
Immunosuppressive drug use (ever)     
 Mycophenolate mofetil1.
Hormones (ever use)     
 Oral contraceptives1.
 Hormone replacement therapy6.

At baseline, the mean ± SD CRP level was 9.54 ± 19.98 mg/liter and the median level was 3.60 mg/liter. As expected, the distribution of CRP level in SSc was skewed toward normal; therefore, outliers inflated the mean CRP level. The minimum level was 0.00 mg/liter and the maximum level was up to 249.00 mg/liter. The 25th percentile was 1.60 mg/liter and the 75th percentile was 8.35 mg/liter. Elevation of CRP concentration >8 mg/liter and ESR occurred in 25.7% and 38.2% of the entire group, respectively; these frequencies were highest in the early dcSSc subset (41.5% and 44.2%, respectively). Baseline CRP levels in dcSSc (mean ± SD 11.98 ± 25.41 mg/liter, median 5.00 mg/liter) were higher than those with lcSSc (mean ± SD 8.15 ± 16.09 mg/liter, median 3.05 mg/liter; P = 0.016 and P < 0.0001, respectively). Patients with SSc with a disease duration ≤3 years had higher CRP levels (mean ± SD 12.89 ± 28.13 mg/liter, median 4.45 mg/liter) than those with a disease duration >3 years (mean ± SD 8.60 ± 17.06 mg/liter, median 3.40 mg/liter; P = 0.041 and P = 0.009, respectively). Early dcSSc had higher CRP levels (mean ± SD 17.18 ± 34.73 mg/liter, median 5.65 mg/liter) than late dcSSc (mean ± SD 9.77 ± 20.42 mg/liter, median 4.20 mg/liter; P = 0.056 and P = 0.018, respectively).

CRP was moderately and significantly correlated with ESR in almost all subsets despite prednisone treatment. CRP was correlated with skin score (MRSS) overall, in dcSSc, and in late dcSSc subsets, but correlations were low (ρ = 0.2–0.3). CRP was correlated with systolic PAP, but correlations were low (ρ = ≤0.2). CRP had negative correlations with total lung capacity (TLC), forced vital capacity (FVC), and percentage of predicted diffusing capacity of carbon monoxide (DLCO) and was more apt to be elevated in those with TLC <80%, FVC <80%, and DLCO <75%. However, the correlation of CRP versus the diagnostic method of ILD by either chest radiograph or HRCT of the chest was very low (ρ = 0–0.2). CRP was not correlated with SRC ever or with inflammatory arthritis. An elevated CRP level occurred more often in those with tendon friction rubs overall and in dcSSc and early dcSSc subsets, but the correlation was weak. Significant correlations between CRP and SDAS, SDSS, and HAQ DI were observed overall and in most subsets (Table 2). In each of the 9 organ systems of the SDSS, CRP was correlated with the lung system in almost all of the disease subsets (Table 3). Overall and in both lcSSc and dcSSc patients, CRP was correlated with most of the organ systems of the SDSS.

Table 2. Relationships between CRP concentration and continuous/categorical clinical parameters*
CorrelationAll patientslcSScdcSScEarly dcSScLate dcSSc
Visit 1 (n = 1,043)Visit 2 (n = 677)Visit 1 (n = 647)Visit 2 (n = 411)Visit 1 (n = 396)Visit 2 (n = 266)Visit 1 (n = 117)Visit 2 (n = 64)Visit 1 (n = 269)Visit 2 (n = 202)
  • *

    Values are the nonparametric correlation (Spearman's ρ). CRP = C-reactive protein; lcSSc = limited cutaneous systemic sclerosis (scleroderma); dcSSc = diffuse cutaneous SSc; ESR = erythrocyte sedimentation rate; MRSS = modified Rodnan skin score; sPAP = systolic pulmonary artery pressure; TLC = total lung capacity; FVC = forced vital capacity; DLCO = diffusing capacity of carbon monoxide; HRCT = high-resolution computed tomography; SRC = scleroderma renal crisis; SDAS = Scleroderma Disease Activity Score; SDSS = Scleroderma Disease Severity Score; HAQ = Health Assessment Questionnaire; DI = disability index.

  • Correlation is significant at the 0.01 level (2-tailed).

  • Correlation is significant at the 0.05 level (2-tailed).

  • §

    On echocardiogram.

  • Chest radiograph was compared as presence of interstitial lung disease/pulmonary fibrosis (yes vs. no).

  • #

    No vs. swollen joint count ≥4 or tender joint count ≥8.

  • **

    Yes vs. no.

CRP vs. ESR0.3740.3370.3330.3920.3890.2640.4610.1960.3370.287
CRP vs. MRSS0.2090.1200.1200.0860.3020.1750.2030.1140.3040.218
CRP vs. sPAP§0.1530.1830.1570.2940.099−0.009−0.121−0.1330.2150.041
CRP vs. TLC−0.247−0.371−0.226−0.421−0.245−0.294−0.167−0.173−0.275−0.339
CRP vs. FVC−0.255−0.390−0.232−0.429−0.269−0.318−0.353−0.369−0.233−0.295
CRP vs. DLCO−0.194−0.268−0.209−0.307−0.209−0.189−0.290−0.183−0.178−0.185
CRP vs. lung fibrosis by chest radiograph0.1060.1180.1300.1250.0500.0770.0090.0630.0580.078
CRP vs. lung fibrosis by HRCT0.1860.1580.1860.1670.1270.1280.1620.0660.1250.135
CRP vs. SRC ever−0.0100.0030.0200.00−0.0180.018−0.2050.0490.079−0.014
CRP vs. arthritis#0.0280.0380.0030.0420.031−0.0090.1480.069−0.019−0.039
CRP vs. tendon friction rubs**0.1400.0680.062−0.0310.1780.1350.2670.2750.0850.070
CRP vs. SDAS0.2290.1550.1450.1320.3430.2000.1960.1140.3530.225
CRP vs. SDSS0.2130.2060.1560.2180.2890.2020.4340.1440.2470.236
CRP vs. HAQ DI0.2800.2700.1990.2560.3410.2730.3360.2930.3200.245
Table 3. Correlation of CRP level versus 9 domains of the Systemic Sclerosis Disease Severity score (at annual visits 1 and 2)*
DomainAll patientslcSScdcSScEarly dcSScLate dcSSc
Visit 1 (n = 1,043)Visit 2 (n = 677)Visit 1 (n = 647)Visit 2 (n = 411)Visit 1 (n = 397)Visit 2 (n = 266)Visit 1 (n = 117)Visit 2 (n = 61)Visit 1 (n = 269)Visit 2 (n = 201)
  • *

    Values are the nonparametric correlation (Spearman's ρ). CRP = C-reactive protein; lcSSc = limited cutaneous systemic sclerosis (scleroderma); dcSSc = diffuse cutaneous SSc; GI = gastrointestinal.

  • Correlation is significant at the 0.01 level (2-tailed).

  • Correlation is significant at the 0.05 level (2-tailed).

CRP vs. general0.1200.1440.0790.1710.1800.1380.1970.0580.1360.153
CRP vs. vessel0.0080.037−0.0070.0150.0160.0970.045−0.0200.0490.137
CRP vs. skin0.2100.1320.0590.0660.3000.1970.1560.1510.3310.207
CRP vs. joint0.1290.0710.0140.0260.1800.0780.1850.0020.1590.092
CRP vs. muscle0.0790.102−0.0120.0960.1510.0870.093−0.2750.1730.202
CRP vs. GI0.0660.0570.0720.1020.038−0.0500.165−0.1220.038−0.038
CRP vs. lung0.1910.2370.1860.2710.2110.1920.2720.1990.2010.177
CRP vs. heart0.1170.1180.1400.0740.0740.1520.1520.2380.0820.114
CRP vs. kidney0.027−0.003−0.0170.043−0.0240.236−0.145−0.0580.029
CRP vs. total score0.2130.2070.1560.2180.2890.2050.434−0.0080.2470.264

Most patients had normal CRP levels and comparing the mean CRP levels among 6 annual visits (Figure 1), the mean CRP level was not significantly different (P = 0.123); however, the mean ± SD CRP level decreased over time (9.54 ± 19.98 mg/liter at the first visit and 5.64 ± 7.89 mg/liter at the sixth visit). The median and mode CRP levels decreased from the first visit (3.6 mg/liter and 3.0 mg/liter, respectively) to the sixth visit (2.9 mg/liter and 1.0 mg/liter, respectively), but were also not significantly different. Patients currently taking prednisone had higher mean ± SD CRP levels (13.87 ± 27.72 mg/liter) than former use of prednisone (9.05 ± 15.03 mg/liter; P = 0.038) and never use (8.53 ± 17.12 mg/liter; P = 0.002). For patients who were not receiving prednisone at one visit but were at the next, there was a nonsignificant decrease in CRP level, but our numbers were small (Table 4). Mean skin score changes (MRSS) in dcSSc overall, early dcSSc, and late dcSSc subsets from the first 2 annual visits showed no significant difference between those with elevated or normal CRP levels (Table 5). CRP and immunosuppressive drugs and steroids were compared (Table 6).

Figure 1.

Box plot of 6 annual followup visits of C-reactive protein (CRP) for patients available at each visit. The line shows the mean CRP level of all visits (9.35 mg/liter). Sample size is variable from year 1 (n = 1,145) to year 6 (n = 89). Note that the patients had a long disease duration on average at study entry. Mean and median CRP levels are not different over time (P = 0.123 and P = 0.275, respectively).

Table 4. Change in C-reactive protein (CRP) level between the first 2 visits vs. prednisone status
 NMean ± SD CRP level changeIndependent-sample t-test
Adding prednisone70−5.50 ± 28.850.078
Not adding prednisone4280.94 ± 22.26 
Table 5. Relationship between CRP level at the first visit and skin score changes from the first to second annual visit*
 MRSS change from visit 1 to visit 2, mean ± SDP (independent-sample t-test)
  • *

    CRP = C-reactive protein; MRSS = modified Rodnan skin score; dcSSc = diffuse cutaneous systemic sclerosis (scleroderma).

Early dcSSc  
 CRP level >8 mg/liter−0.82 ± 6.490.176
 CRP level ≤8 mg/liter−3.45 ± 7.53 
Late dcSSc  
 CRP level >8 mg/liter−2.24 ± 7.230.312
 CRP level ≤8 mg/liter−1.01 ± 5.90 
Total dcSSc  
 CRP level >8 mg/liter−1.54 ± 6.930.920
 CRP level ≤8 mg/liter−1.65 ± 6.40 
Table 6. Comparison between SSc and normal (≤8 mg/liter) and elevated (>8 mg/liter) CRP levels at baseline visit*
Demographic or disease characteristicCRP level ≤8 mg/literCRP level >8 mg/literPSpearman's ρBinary logistic regression for CRP level >8 mg/liter
OR (95% CI)P
  • *

    Values are the number/total (percentage) unless otherwise indicated. SSc = systemic sclerosis (scleroderma); CRP = C-reactive protein; OR = odds ratio; 95% CI = 95% confidence interval; N/A = not available; BMI = body mass index; RP = Raynaud's phenomenon; CSRG = Canadian Scleroderma Research Group; MRSS = modified Rodnan skin score; PAP = pulmonary artery pressure; HRCT = high-resolution computed tomography; TLC = total lung capacity; FVC = forced vital capacity; DLCO = diffusing capacity of carbon monoxide; SJC = swollen joint count; TJC = tender joint count; SDAS = Scleroderma Disease Activity Score; SDSS = Scleroderma Disease Severity Score; HAQ = Health Assessment Questionnaire; DI = disability index; ESR = erythrocyte sedimentation rate.

  • † Spearman's rho correlation with CRP level (mg/liter) and univariate/multivariate logistic regression for elevated CRP level.

  • Correlation is significant at the 0.01 level (2-tailed).

  • §

    Correlation is significant at the 0.05 level (2-tailed).

Total group678/913 (74.3)235/913 (25.7)N/AN/AN/AN/A
Age at baseline visit, mean ± SD years55.9 ± 11.855.8 ± 12.00.9370.0940.999 (0.987–1.012)0.937
Women, no. (%)594 (87.6)197 (83.8)0.149−0.061N/AN/A
BMI, mean ± SD kg/m225.3 ± 5.326.9 ± 6.4< 0.00010.2321.051 (1.025–1.079)< 0.0001
BMI >25 kg/m2287/669 (42.9)121/229 (52.8)0.0110.1800.671 (0.496–0.906)0.009
Disease duration (onset of RP symptoms to baseline CSRG visit), mean ± SD years15.2 ± 12.314.2 ± 12.60.295−0.060N/AN/A
Disease duration (onset of first non-RP symptom manifestation to baseline visit), mean ± SD years11.5 ± 9.69.8 ± 9.10.018−0.067§0.980 (0.963–0.997)0.018
Disease duration ≤3 years133/664 (20)67/228 (29.4)0.0040.0880.602 (0.427–0.848)0.004
Diffuse cutaneous subset, %36.345.90.0130.1470.672 (0.492–0.917)0.012
Early diffuse cutaneous subset55/658 (8.4)39/227 (17.2)< 0.00010.1300.440 (0.283–0.684)< 0.0001
Early limited cutaneous subset66/658 (10.0)25/229 (10.9)0.705−0.018N/AN/A
MRSS, mean ± SD9.1 ± 8.612.8 ± 11.3< 0.00010.2091.039 (1.023–1.054)< 0.0001
Systolic PAP, mean ± SD mm Hg35.7 ± 14.240.2 ± 19.70.0150.1531.016 (1.005–1.027)0.005
Elevated systolic PAP >40 mm Hg101/443 (22.8)46/135 (34.1)0.010.1390.571 (0.376–0.869)0.009
Moderate to large pericardial effusion8/612 (1.3)4/196 (2.0)0.4980.000N/AN/A
Lung fibrosis (chest radiograph)136/646 (21.1)67/224 (29.9)0.0080.1060.625 (0.443–0.880)0.007
Lung fibrosis in those with HRCT162/535 (30.3)83/193 (43.0)0.0020.1860.576 (0.410–0.808)0.001
HRCT chest: honeycomb appearance36/170 (21.2)15/75 (20)1.000−0.040N/AN/A
TLC % predicted, mean ± SD95.7 ± 18.386.2 ± 20.3< 0.0001−0.2470.974 (0.965–0.983)< 0.0001
FVC % predicted, mean ± SD93.4 ± 19.284.0 ± 19.5< 0.0001−0.2550.976 (0.968–0.984)< 0.0001
DLCO % predicted, mean ± SD71.9 ± 21.265.44 ± 22.4< 0.0001−0.1940.986 (0.978–0.994)< 0.0001
TLC <80% predicted103/581 (17.7)76/197 (38.6)< 0.0001−0.2252.915 (2.039–4.166)< 0.0001
FVC <80% predicted139/601 (23.1)84/202 (41.6)< 0.0001−0.2202.366 (1.687–3.318)< 0.0001
DLCO <75% predicted323/588 (54.9)126/194 (64.9)0.015−0.1571.520 (1.086–2.129)0.015
Scleroderma renal crisis ever26/673 (3.9)8/231 (3.5)1.000−0.010N/AN/A
Proximal muscle weakness91/668 (13.6)51/230 (22.2)0.0030.061N/AN/A
Tendon friction rub87/670 (13.0)51/229 (22.3)0.0010.1400.521 (0.355–0.765)0.001
SJC ≥427/659 (4.1)13/230 (5.7)0.3560.009N/AN/A
TJC ≥825/659 (3.8)12/230 (5.2)0.3430.060N/AN/A
Arthritis (SJC ≥4 or TJC ≥8)44/659 (6.7)20/230 (8.7)0.3030.028N/AN/A
Active digital ulcer48/676 (7.1)26/232 (11.2)0.0520.060N/AN/A
No. of fingers with active digital ulcer, mean ± SD2.4 ± 2.22.3 ± 1.80.9370.062N/AN/A
Digital gangrene/necrosis7/676 (1.0)4/232 (1.7)0.4860.010N/AN/A
SDAS score (range 0–10), mean ± SD1.95 ± 1.432.63 ± 1.76< 0.00010.229N/AN/A
SDSS score (range 0–36), mean ± SD8.12 ± 3.639.83 ± 4.37< 0.00010.213N/AN/A
HAQ DI score (range 0–3), mean ± SD0.71 ± 0.631.03 ± 0.78< 0.00010.280N/AN/A
Hemoglobin, mean ± SD gm/liter131.6 ± 13.7126.4 ± 16.1< 0.0001−0.1830.976 (0.966–0.986)< 0.0001
Serum creatinine, mean ± SD μmoles/liter78.0 ± 31.991.8 ± 89.90.023−0.1091.005 (1.001–1.008)0.004
Creatine kinase, mean ± SD units/liter93.1 ± 75.099.2 ± 95.30.392−0.048N/AN/A
ESR, mean ± SD mm/hour19.0 ± 18.031.4 ± 26.1< 0.00010.3741.026 (1.018–1.034)< 0.0001
Elevated ESR >20 mm/hour203/627 (32.4)120/209 (57.4)< 0.00010.3300.355 (0.258–0.490)< 0.0001
Antinuclear antibody596/621 (96)201/215 (93.5)0.137−0.033N/AN/A
Anticentromere antibody224/621 (36.1)68/215 (31.6)0.247−0.042N/AN/A
Anti–topoisomerase I (Scl-70)94/621 (15.1)36/215 (16.7)0.5860.049N/AN/A
Proteinuria39/636 (6.1)24/205 (11.7)0.0140.083§0.493 (0.289–0.841)0.01
Current use of corticosteroids69/674 (10.2)55/232 (23.7)< 0.00010.1360.367 (0.248–0.543)< 0.0001
Current use of immunomodulator155/674 (23.0)69/232 (29.7)0.0430.0900.706 (0.505–0.985)0.041

Comparison between patients with elevated CRP levels and those with normal CRP levels showed that patients with elevated CRP levels had a significantly higher BMI; had shorter disease duration; were more likely to have dcSSc, higher MRSS, higher systolic PAP, higher prevalence of lung fibrosis, worse pulmonary function (TLC, FVC, and DLCO % predicted), a higher prevalence of proximal extremity girdle muscle or neck weakness, more tendon friction rubs, higher disease activity and severity, higher HAQ DI scores, lower hemoglobin (gm/liter), higher creatinine (μmoles/liter), higher ESR (mm/hour), and higher prevalence of proteinuria; and due to higher disease activity, were more likely to use steroids or immunomodulators (Table 6).

Using significant data from the univariate correlations, MRSS, TLC <80% predicted, creatinine, and systolic PAP were used for a multivariate logistic regression model (Table 6). BMI was not selected because both patients with a high BMI (>25 kg/m2) had higher CRP levels and a below normal BMI was also associated with severe disease; therefore, the relationship was not linear, but a “J curve.” ESR was not selected because it is similar to CRP. FVC <80% and DLCO <75% predicted were redundant because a pulmonary function parameter was already included. Disease activity and severity scores were not used because they were composite indices and the HAQ DI was not used because it is likely a result of activity and damage and not directly predictive of CRP. After adjustment for immunomodulator drugs currently used, MRSS (odds ratio [OR] 1.03 [95% confidence interval (95% CI) 1.01–1.05], P = 0.005), TLC <80% predicted (OR 2.76 [95% CI 1.73–4.40], P = 0.0001), and creatinine (OR 1.005 [95% CI 1.001–1.010], P = 0.019) were predictors of elevated CRP level.

With respect to death, 9.2% (106 patients) died during the observation period (maximum followup from time of entry 5 years 8 months). Death occurred more frequently in those with an elevated CRP level (14%) compared to a normal CRP level (7.1%; P = 0.002). Survival curves for patients entering the CSRG separated by baseline CRP level showed worse survival in the group with elevated CRP level (P = 0.001) (Figure 2).

Figure 2.

Kaplan-Meier curves for survival between patients with elevated and normal C-reactive protein (CRP) levels. Survival is less in those with an elevated CRP level at baseline entry into the Canadian Scleroderma Research Group database (P = 0.001).


One-quarter of prevalent SSc patients had an elevated CRP level and more patients with dcSSc than lcSSc had elevated CRP levels, and especially early dcSSc. CRP level has been noted to be less likely to be elevated in longer disease duration, as was the case in this cohort. In 2 smaller cohorts from Japan (24, 25), one-third to one-half had elevated CRP levels, but those patients had a shorter disease duration (4.4 years). CRP level may be elevated early when cytokines are causing more inflammation, and later there are less CRP level elevations as more fibrosis occurs. We found higher CRP values in worse ILD defined by those with worse PFTs. Perhaps this represents underlying alveolitis. Imaging for pulmonary fibrosis did not correlate with CRP. This could be due to the fact that many SSc patients with abnormal chest radiographs that demonstrate ILD/pulmonary fibrosis do not have active or progressive lung involvement but are stable between visits. However, when comparing restrictive lung function, there was an important association with CRP. Remote SRC was not related to CRP, since we could not compare current SRC and CRP because the database had a surplus of SRC occurring around their annual visits. However, elevated creatinine was associated with elevated CRP levels.

A correlation between CRP and New York Heart Association functional class (r = 0.23), right atrial pressure (r = 0.25) by right-sided heart catheterization (RHC), and 6-minute walking distance (r = 0.19) in idiopathic pulmonary artery hypertension (PAH) has been described (38). We also found that systolic PAP on echocardiogram was associated with CRP. This is a surrogate for PAH and pulmonary hypertension. The database does not record the results of RHC and markers of vascular endothelial injury have not been collected (such as vascular cell adhesion molecule 1, vascular endothelial growth factor, and von Willebrand factor). The latter has been reported as being associated with SSc PAH in lcSSc (39).

In the CSRG database, we did not demonstrate a correlation between CRP and arthritis, although many studies, especially in rheumatoid arthritis, have shown that CRP correlates closely with changes in inflammation/disease activity, radiologic damage, progression, and functional disability (40). Ascertaining swollen joints in patients with SSc may be inaccurate due to tight skin and sclerodactyly, and tender joints can be caused by osteoarthritis. SJCs were uncommon in our SSc patients, but we did not perform ultrasound to detect tenosynovitis.

The distribution of CRP level in the SSc patients was similar to the normal population (41–43) because it is skewed toward lower values (as would also be seen in ESR and in other diseases such as rheumatoid arthritis) and trended toward stability over time with respect to the mean (32, 44). The median CRP level of patients with SSc was higher than that reported in the healthy population and tended to regress over time. Regression of CRP in our study might be caused by a decreasing number of early dcSSc patients who usually have high CRP levels, especially since few patients with early dcSSc were followed in later visits (i.e., visits 4–6). Regression to the mean with respect to CRP over time was observed despite the long disease duration.

We found small to moderate correlations of CRP with many aspects of SSc such as ESR, MRSS, pulmonary function, tendon friction rubs, SDAS, SDSS, and HAQ DI. A recent study from Europe found that SSc disease activity defined by the EScSG activity index (33–36) and a modified 12-point activity index (45) were associated with HAQ DI, digital ulcers, MRSS, contractures, and CRP. Some lack of correlations in our study, especially in the early dcSSc subset, might result from small numbers of patients.

CRP is mainly secreted from hepatocytes via stimulation of IL-6 (22) as an innate nonspecific immune response to systemic inflammation. Higher IL-6 levels were described in early active dcSSc patients (46–50), which correlated with CRP (25, 47, 50). One study showed a strong relationship of IL-6 with disease activity, but this was poorly correlated with an acute-phase response (46). Systemic glucocorticoids do decrease CRP concentration and IL-6, but it may be confounding by indication where only the few patients with high inflammatory features are prescribed steroids (51).

Our study found that some features of SSc were not associated with elevated CRP levels, such as SRC ever, pericardial effusions, muscle weakness, arthritis, and digital ulcers. Perhaps this can be partly explained by the low prevalence in this database. In the multivariate logistic regression analysis, we found that MRSS, TLC <80% predicted, and serum creatinine were predictors of elevated CRP levels.

CRP antagonizes endothelial cell nitric oxide synthase mediated by the coupling of Fcγ receptor I (FcγRI) to FcγRIIB (52–54). Patients with elevated CRP levels had more cardiopulmonary manifestations and these are among the most common SSc-related deaths (55). In addition to age and sex, cardiac, pulmonary hypertension, lung disease, and high skin scores are associated with reduced SSc survival (17).

Limitations of our study are the inability to rule out infection such as small bowel overgrowth or upper respiratory tract infections in our population at their study visits unless it was clinically relevant. Despite the size of our sample, we may have a survival cohort that could underestimate the CRP association with early disease due to the long mean disease duration at cohort entry for many of the patients. Although we did analyze the early dcSSc subset, a larger cohort of incident dcSSc may have helped clarify some issues, such as correlation of CRP with many aspects of SSc. This is a large cohort with annual followup of CRP levels and many clinical factors are recorded with very little missing data and good followup rates. CRP and associations with SSc were studied cross-sectionally and prospectively in a multifaceted way.

CRP level is not elevated in most patients with SSc in a prevalent population, but is more apt to be increased in early dcSSc and in dcSSc compared to lcSSc. Despite the finding that 25% of patients had values above normal, CRP is related to skin scores, tendon friction rubs, pulmonary fibrosis, elevated PAP, elevated creatinine, and proteinuria. In a mostly prevalent SSc database, CRP level seems to regress over time. Higher values are associated with activity and damage scales.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Pope had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Muangchan, Harding, Khimdas, Baron, Pope.

Acquisition of data. Harding, Baron, Pope.

Analysis and interpretation of data. Muangchan, Harding, Khimdas, Bonner, Baron, Pope.



Investigators of the Canadian Scleroderma Research Group are as follows: M. Baron, J.-P. Mathieu, M. Hudson, S. Ligier, T. Grodzicky: Montreal, Quebec; J. Pope: London, Ontario; J. Markland: Saskatoon, Saskatchewan; D. Robinson: Winnipeg, Manitoba; N. Jones: Edmonton, Alberta; N. Khalidi, E. Kaminska: Hamilton, Ontario; P. Docherty: Moncton, New Brunswick; A. Masetto: Sherbrooke, Quebec; E. Sutton: Halifax, Nova Scotia; S. LeClercq: Calgary, Alberta; C. Thorne: Newmarket, Ontario; M. Fritzler: Advanced Diagnostics Laboratory, Calgary, Alberta, Canada.