Association of anti–cyclic citrullinated peptide and anti-Sa/citrullinated vimentin autoantibodies in rheumatoid arthritis



Rheumatoid arthritis (RA) is a common systemic autoimmune disease mainly characterized by persistent joint inflammation that results in loss of joint function, morbidity, and premature mortality, affecting 1% of the population worldwide (1, 2). In the last few years, many autoantibodies associated with RA have been described. Many patients with RA produce autoantibodies directed toward numerous autoantigens, but most of these autoantibodies are not specific for RA. Rheumatoid factors (RFs), the immunologic hallmark of RA, have modest RA disease specificity (up to 66%) (3, 4). Therefore, a search for novel autoantibodies and their respective target molecules that could be useful in the diagnosis and prognosis of RA, and in therapeutic decision making is warranted. The very high specificity of anti-Sa autoantibodies (92–98%), their positive predictive value (84–99%) for RA diagnosis, and their strong association with extraarticular manifestations and more severe articular disease support their diagnostic and prognostic value. However, immunoblotting, which has low sensitivity (21–43%), particularly in early RA, is the currently available method for anti-Sa detection (5–9).

Recently, a novel and very specific autoantibody for RA has been described. Patients with RA develop antibodies to as yet undefined proteins containing modified (citrullinated) arginine residues. Such citrulline residues are essential parts of the antigenic determinants recognized by the RA autoantibodies, and are therefore so-called anti–cyclic citrullinated peptide (anti-CCP) antibodies. The specificity of anti-CCP antibodies for RA is high (96–99%), and the reported sensitivity of patients with RA for anti-CCP autoantibodies is 66–88%, dependent on the characteristics of the RA population. The diagnostic value of anti-CCP antibodies is further substantiated by their presence in sera from patients with early RA, although they seem to occur mainly in RF-positive sera (3, 7, 10–16).

The Sa antigen has been identified as citrullinated vimentin (17). Few studies comparing the coexistence of anti-Sa and anti-CCP antibodies in sera of patients with RA have been published (7, 17, 18). We sought to determine the prevalence of anti-CCP in an a priori, anti-Sa–positive sample of patients with RA, and to compare these patients with anti-Sa–negative patients with RA.

Patients and Methods

Study population

Fifty-six patients who fulfilled the American College of Rheumatology (formerly the American Rheumatism Association) criteria for RA (19) were recruited at the Rheumatology Service at the University General Hospital Gregorio Marañón in Madrid, Spain. Twenty-eight of 56 patients with RA were randomly selected from the anti-Sa–positive RA database; the remaining 28 patients with RA were selected from the anti-Sa–negative database. Because we selected patients according to laboratory data, the study was retrospective. Eleven patients with systemic lupus erythematosus, 11 patients with osteoarthritis, and 6 patients with fibromyalgia were also studied as control groups. The study fulfilled the guidelines required by our ethics committee. All patients gave informed consent for the study. Recent-onset RA was defined as 1-year evolution of disease.


The CCP IgG assay was performed using synthetic CCP enzyme immunoassay technology (CCP-2; Euro-Diagnostica Inmunoscan RA, Medeon, Sweden), according to the manufacturer's instructions. The best cut-off level of anti-CCP antibodies determined by receiver operating curves was 25 units/ml. A titer of ≥100 units/ml was considered high. RF in serum was measured by rate nephelometry (IMMAGE; Beckman Coulter, Fullerton, CA).

Anti-Sa antibodies were determined by in-house immunoblotting using human placenta antigen. Human placenta extracts were prepared using the method described by Clark et al (20), with some modifications as described elsewhere (21–23). Electrotransfer to nitrocellulose was performed as described elsewhere (22–26). Membranes were blocked with 3% bovine serum albumin in Tris saline and incubated for 1 hour at room temperature with 1:5 dilutions of sera from patients with RA or other rheumatic diseases and from normal controls. After washing, membranes were incubated with an anti-IgG–peroxidase conjugate for 2 hours. After 3 washing steps, reactions were visualized with H2O2/4-chloro-1-naphthol.

Statistical analysis

Results were analyzed using the nonparametric Mann-Whitney U test because of the small sample size. Correlations between the values were determined by Spearman's rank test. P values less than 0.05 were considered significant.


Global prevalence of positive anti-CCP antibodies in our 56 patients with RA was 91.1% (n = 51), whereas the prevalence of high titers of anti-CCP antibodies (>100 units/ml) was 73.2% (n = 41) (Table 1). There was a wide range of positive values of anti-CCP (median 283.4 units/ml, range 25–3,343.5). Sensitivity, specificity, and positive predictive value of anti-CCP antibodies for RA were 92.8%, 93.7%, and 96.5%, respectively, in the entire RA group with respect to control groups; these values increased to 100%, 93.7%, and 98.1%, respectively, when only considering the anti-Sa–positive group.

Table 1. Sensitivity, specificity, PPV, and NPV for rheumatoid arthritis according to different cut-off levels of anti-CCP*
 Anti-CCP antibody levels
>25 (n = 51)>50 (n = 44)>75 (n = 43)>100 (n = 41)
  • *

    Values are the percentage (95% confidence interval). PPV = positive predictive value; NPV = negative predictive value; anti-CCP = anti–cyclic citrullinated peptide.

Sensitivity91.1 (84.8–97.4)78.6 (69.6–87.6)76.8 (67.5–86.1)73.2 (63.5–82.9)
Specificity85.2 (77.4–93.0)96.3 (92.2–100)96.3 (92.2–100)96.3 (92.2–100)
PPV92.7 (87.0–98.4)97.8 (94.6–100)97.7 (94.4–100)97.6 (94.2–100)
NPV82.1 (73.7–90.5)68.4 (58.2–78.6)66.7 (56.4–77.0)63.4 (52.8–74.0)

Demographic, clinical, and serologic characteristics of the patients with RA are described in Table 2 according to the presence or absence of anti-Sa antibodies. The titer of anti-CCP antibodies was significantly higher in the RF-positive group compared with the RF-negative group (410 units/ml versus 29.7 units/ml, P < 0.01), whereas no significant differences were observed in the anti-Sa–positive group versus the anti-Sa–negative group. As shown in Table 2, the median titer of anti-CCP antibodies was significantly higher in the anti-Sa–positive group than in the anti-Sa–negative group (522 units/ml versus 150 units/ml, P = 0.015). Indeed, significant correlations were found between RF-positive patients with RA with high anti-CCP titers (P < 0.01) and anti-Sa–positive patients with high anti-CCP antibodies (P = 0.01). Ten percent of all patients with RA had negative RF (n = 6), 2 of whom had positive anti-CCP antibodies and 2 had positive anti-Sa antibodies.

Table 2. Characteristics of the patients with RA with anti-Sa–positive and anti-Sa–negative autoantibodies*
CharacteristicAnti-Sa positive (n = 28)Anti-Sa negative (n = 28)P
  • *

    Values are the frequency (percentage) unless otherwise indicated. RA = rheumatoid arthritis; IQR = interquartile range; NS = not significant; anti-CCP = anti–cyclic citrullinated peptide; RF = rheumatoid factor; ESR = erythrocyte sedimentation rate.

Age at RA onset, median (IQR)53.0 (18–71)50 (18–70)NS
Age, median (IQR) years69.5 (34–81)66.9 (31–81)NS
Sex: male/female6 (21.4)2 (7.1)NS
Disease duration, median (IQR)15.0 (1–33)12.0 (2–40)NS
Deaths4 (14.3)4 (14.3)NS
Anti-CCP antibody, median2724NS
Anti-CCP antibody, median (IQR) units/ml522 (21.55–3,233.051)150 (21.15–3,343.53)0.015
Vasculitis1 (3.6)2 (7.1)NS
Bony erosions21 (75.0)22 (78.6)NS
Joint ankylosis5 (17.9)4 (14.3)NS
Anemia12 (42.9)13 (46.4)NS
Positive RF26 (92.9)24 (85.7)NS
ESR >50 mm/hour18 (64.3)15 (53.6)NS
Steroids22 (78.6)20 (71.4)NS
Methotrexate16 (57.1)18 (64.3)NS
Gold19 (67.9)22 (78.6)NS
Biologic therapies5 (17.9)2 (7.1)NS
Synoviorthesis6 (21.4)5 (17.9)NS
Joint surgery7 (25.0)9 (32.1)NS
Radiograph degree   
 1–215 (53.6)21 (75.0)NS
 3–413 (46.4)7 (25.0)NS

Interestingly, in the anti-CCP–positive RA group, we observed more severe radiologic grade (P = 0.05) and higher frequency of erythrocyte sedimentation rate >50 mm/hour (P = 0.05) compared with the anti-CCP–negative group. Also, bony erosions were more common in the anti-CCP–positive group, although they were not statistically significant (P = 0.08). In this study, patients with RA with high titers of anti-CCP antibodies had up to 5.1 times higher incidence of worse radiographic degree compared with patients who were anti-CCP negative (90.0% versus 63.9%).

Patients with recent-onset RA were defined as having <1 year of evolution of disease. Incidence of anti-CCP antibodies in patients with recent-onset RA (26 [46.4%] of 56 patients with RA) was 84.6% (72.7% with anti-CCP autoantibodies >100 units/ml). The main clinical features in this group of patients with recent-onset RA are shown in Table 3. Of 22 patients with anti-CCP antibodies, 2 (both with titers >100 units/ml) presented subcutaneous nodules (9.1% versus 0% without anti-CCP antibodies) and 10 (62.5% versus 100.0%) presented bony erosions; anti-CCP–positive patients had a higher prevalence of anti-Sa and RF antibodies, as shown in Table 3.

Table 3. Main characteristics of recent-onset rheumatoid arthritis patients (<1 year of disease evolution) according to anti-CCP autoantibodies*
CharacteristicAnti-CCP positive (n = 22)Anti-CCP negative (n = 4)Total (n = 26)
  • *

    Values are the absolute numbers (percentage). Anti-CCP = anti–cyclic citrullinated peptide; RF = rheumatoid factor.

Subcutaneous nodules2 (9.1)0 (0)2 (7.7)
Erosions10 (62.5)4 (100)14 (53)
Radiograph degree   
 110 (45.5)1 (9.1)11 (42.3)
 211 (50)3 (75)14 (53.8)
 31 (4.5)0 (0)1 (3.8)
Anti-Sa antibodies10 (45.5)1 (25)11 (42.3)
RF positive20 (90.9)2 (50)22 (84.6)


In recent years, a few RA-specific autoantibodies, including anti-Sa/citrullinated vimentin and anti-CCP among others, have been reported (3, 4, 7, 17). Given that the response of patients with RA to therapy is highly unpredictable, markers for prognosis of RA are extremely important. Indeed, the clinical response can vary from self-limited disease with no radiographic destruction in some patients to progressive joint destruction in others. Up to now, RF has been the only immunologic tool to characterize patients with RA, although it is not disease specific and is often absent in early disease (27–29). The specificity of the anti-CCP and anti-Sa/citrullinated vimentin for established RA and early RA has been documented (3–16, 18), and these assays might be useful to predict poor outcomes in patients with RA (7, 16, 30–33).

In agreement with previous reports (13), our findings indicate that anti-CCP antibodies are very specific for the disease and can be detected early in the disease. In our series of 56 patients, the global prevalence of anti-CCP antibodies was 91%, which is much higher than the 70% prevalence described in the literature. This may be due to the fact that half of the patients in the RA group were positive for anti-Sa/citrullinated vimentin antibodies. When anti-CCP antibodies were present in conjunction with anti-Sa/citrullinated vimentin, sensitivity and positive predictive value for RA increased to 100% and 98.1%, respectively. We observed a significantly higher prevalence of anti-CCP antibodies when the RF was positive. Anti-CCP antibodies in our cohort of patients were more specific for the diagnosis of RA than RF, in agreement with previous reports (7). Potentially more important for the clinician is the diagnostic power of these antibodies when RF is not detectable. In our cohort, a high positivity for either anti-Sa/citrullinated vimentin or anti-CCP antibodies was also observed in RF-negative patients with RA.

This study confirms the significant correlation between RF or anti-Sa/citrullinated vimentin antibodies and high titers of anti-CCP antibodies (7, 17, 18). As reported by Vossenaar et al (17), the vast majority of anti-Sa–positive patients with RA tested positive for anti-CCP antibodies, as did a substantial proportion of the anti-Sa–negative RA patients. Although the number of patients was small, anti-Sa–positive sera had significantly higher anti-CCP titers than anti-Sa–negative sera. Recently, Boire et al (18) have found that anti-Sa/citrullinated vimentin antibodies are significantly associated with erosive disease outcome, whereas RF and anti-CCP antibodies were not significant predictors. Therefore, anti-Sa/citrullinated vimentin antibodies as high titers of anti-CCP antibodies may be useful as prognostic markers in patients with early RA. With respect to cross reactivity between the citrullinated peptide/proteins in RA, anti-CCP and anti-Sa are both specifically reactive with citrullinated epitopes (17). However, both autoantibodies are detected by different assays that have a different sensitivity for RA. The CCP-2 test uses a single citrullinated epitope, whereas the citrullinated vimentin/Sa antigen contains many different citrulline residues and thus many different citrullinated epitopes. The immunoblotting technique to detect anti-Sa/citrullinated vimentin antibodies is expensive, time consuming, and does not enable the quantification of the antibodies. However, anti-CCP antibodies are detected by a commercial, sensitive, automated, and reproducible technique that enables their quantification. The commercial anti–CCP-2 test is currently sold as a diagnostic test, not as a prognostic test. Raising its cut-off would make it closer to a prognostic test and may be equivalent to performing an anti-Sa test either as described here or, better, as described by Boire et al who used citrullinated protein–enriched cell extracts in Western blot or a citrullinated protein–based enzyme-linked immunosorbent assay (18). The anti–CCP-2 assay is likely to perform less well as a prognostic tool when, as in the real world, there is no selection bias. In an era of emphasis on early aggressive therapies for RA, the ability to selectively choose a particular RA subset for early aggressive therapy that may be potentially toxic and expensive is essential.


The authors wish to thank Angela Puche Torres for her excellent technical assistance.