You are on page 1of 8

The use of second generation anti-CCP antibody (anti-CCP2)

testing in rheumatoid arthritis – A systematic review


J.P. Riedemann1, S. Muñoz2, A. Kavanaugh3

1
Department of Rheumatology and ABSTRACT eral population (1). It is generally pro-
Clinical Epidemiology, 2CIGES, Faculty Objective: To evaluate the diagnostic gressive and affects many joints. In
of Medicine, Universidad de la Frontera, properties and predictive value of the patients who do not respond to therapy,
Temuco, Chile; Center for Innovative
second generation of anti-CCP antibo - RAcan cause significant functional dis-
Therapy, UCSD, Division of Rheumatology,
Allergy and Immunology, La Jolla, Cali- dies (anti-CCP2) in rheumatoid arthri - ability and loss of quality of life (2).
fornia, USA. tis (RA) patients. This poor prognosis has led to an em-
J. Pablo Riedemann, MD, MSc, Associate Methods: A systematic review of the phasis on early diagnosis and aggres-
Professor of Rheumatology and Clinical published literature between January sive treatment (3). Unfortunately, early
Epidemiology; Sergio Muñoz, PhD, Profes- 2002 and June 2005 was performed. diagnosis is difficult in many patients.
sor of Biostatistics; Arthur Kavanaugh, Data were extracted regarding the sen - For example, in many cases of early
MD, Professor of Medicine. sitivity and specificity of anti-CCP2 an - RA, the ACR classification criteria may
Please address correspondence to tibodies in making an accurate diagno - not be met.
Prof. J. Pablo Riedemann, MD. sis of RA, predicting future develop - Over the past few years, several new
E-mail: riedeman@ufro.cl ment of RA, and predicting future radi - autoantibodies have been described in
Clin Exp Rheumatol 2005; 23 (Suppl. 39): ological damage in RA patients. In patients with RA, and their clinical val-
S69-S76. addition, the prevalence of CCP2 anti - ue has been assessed. Most, such as an-
© Copyright CLINICALAND EXPERIMENTAL bodies in patients with other rheumatic tiperinuclear factor antibodies (APF),
RHEUMATOLOGY 2005. diseases was examined. antikeratin antibodies (AKA), and anti-
Results: Among 38 studies initially RA33, have not been successfully in-
Key words: Rheumatoid arthritis,
identified, 27 provided information on corporated into routine clinical practice
diagnosis, anti-CCP2 antibodies,
the use of anti-CCP2 testing. Diagnos - (4). A new group of autoantibodies that
systematic review.
tic properties were assessed in 13 stud - has generated particular interest are the
ies; reported sensitivities ranged from anti-cyclic citrullinated peptide antibo-
14.4% to 96%, and specificities from dies (CCP). First described by Schel-
88.9% to 100%. Odds ratios (OR) for lekens (5), the anti-CCP test has gener-
the future development of RA varied ated great interest in the past years.
from 15.9 among previously healthy in - In recent years a second generation of
dividuals to 37.8 among a group of pa - anti-CCPantibody tests, known as anti-
tients with undifferentiated arthritis. CCP2, have been developed, which may
Several studies suggested that the pres - have better performance characteristics
ence of anti-CCP2 antibodies is highly than the first generation of tests. The
predictive of current radiographic number of publications on anti-CCP2
damage and further damage progres - antibodies is growing exponentially, and
sion. a balanced presentation of the charac-
Conclusions: Anti-CCP2 has a low sen - teristics, merits and drawbacks of this
sitivity to be used as a screening test. test would appear of value for the prac-
However, a positive test is highly spe - ticing clinical rheumatologist. This es-
cific for RA. In addition, anti-CCP2 ap - say presents a critical systematic re-
pears to be highly predictive of the fu - view of published studies concerning
ture development of RA in both normal the diagnostic usefulness of the anti-
individuals and patients with undif - CCP2 antibodies.
ferentiated arthritis. Finally, the pres -
ence of anti-CCP2 antibodies appears Methods
to predict radiographic damage and A systematic review of published litera-
progression among patients with RA. ture following the methods of evi-
dence-based medicine was performed.
Introduction
Rheumatoid arthritis (RA) is a chronic, Literature review
systemic inflammatory disease that af- A search was conducted using elec-
fects approximately 0.5–1% of the gen- tronic databases (MEDLINE and EM-

S-69
Anti-CCP2 antibody testing in RA/ J.P. Riedemann et al.

BASE), restricted to English and Span- - Comparison among different manu- anti-CCP tests and were published
ish language articles. Since the term facturers; since 2002 involved anti-CCP2 kits.
"CCP2 antibodies" has not yet been de- - Clinical use of the test (to whom, These 27 studies are the sources of data
fined as a MeSH term, free text search what other test might be needed). for this systematic review.
was conducted using the following 4) Future areas of research necessary to The 27 reports in which an anti-CCP2
combination: help define optimal use of the test. test was used addressed different as-
CCP or CCP2 or anti-citrullinated fila- pects of its potential clinical applica-
grin antibodies or anti-cyclic citrulli- Results tions including (some publications in-
nated peptide antibodies or autoantibo- Between January 2002 and June 2005, vestigated more than one topic):
dies to cyclic citrullinated peptide a total of 38 papers about the clinical 1) Diagnostic performance: 13 publica-
In addition, references of the papers use of anti-CCP antibodies have been tions (15, 17-20, 24, 26, 27, 33, 34,
initially detected were hand-searched published (8-45). In the description of 38, 39, 41),
to identify additional relevant reports. the methods, 24 of the investigators 2) Prevalence and use in other rheumat-
Finally, as a quality control, a manual specified that their study involved a ic diseases: 7 publications (14, 21,
hand search of all reports published second generation or anti-CCP2 test. In 23, 25, 28, 42, 44),
during 2004 in the Annals of Rheumat- 12 publications the type of anti-CCP 3) Use as predictor of future develop-
ic Diseases was performed (no addi- test (i.e. CCP or CCP2) was not speci- ment of RA: 3 publications (12, 35,
tional publications were detected). fied, and in 2 studies investigators used 40),
their own reagents to measure anti- 4) Association with x-ray damage: 4
Paper review CCP antibodies. The principal authors papers (19, 22, 33, 41),
The scope of the review was restricted of the 12 papers who did not specify 5) Association with changes due to
only to those studies in which CCP2 the type of test evaluated were contact- treatment: 3 publication (8, 13, 31),
antibodies have been used. According ed via e-mail; 11 responded, indicating 6) Association with RA clinical mani-
to information provided by the manu- that 3 had included anti-CCP2. There- festations: 2 papers (18, 19).
facturers (6), kits for the determination fore, 27 of the 36 studies that were per- Of these 27 studies, 19 were performed
of anti-CCP2 antibodies were introduc- formed using commercially available in Europe, 4 in the USA, 2 in Latin
ed in early 2002. For some time after-
wards, some overlap existed with both Table I. Reports in which an anti-CCP2 test was used, addressing different aspects of its
anti-CCP and anti-CCP2 tests being potential clinical applications
available. Thus, this systematic litera-
ture review was restricted to papers Author (ref. no.) Reference Country Topic
published between January 2002 and
Alessandri (8) Ann Rheum Dis 2004 Italy #4
June 2005, in which it was clearly spe-
Berglin (12) Arthritis Research 2004 Sweden #3
cified that a CCP2 test was used.
Bobbio-Pallavicini (13) Arthritis Research 2004 Italy #4
As general methodology, an adaptation Bogliolo (14) J Rheumatol 2005 Italy #2
of the methods proposed by an ACR Bombardieri (15) Arthritis Research 2004 Italy #1
subcommittee to evaluate the utility of Choi (17) J Korean Med Sci 2005 Korea #1
immunologic laboratory testing in Correa (18) Biomedica 2004 Colombia/Argentina #1, #6
rheumatic diseases was used, with spe- De Ricke (19) Ann Rheum Dis 2004 Belgium #1, #5, #6
cial emphasis on the diagnostic aspects Dubucquoi (20) Ann Rheum Dis 2004 France #1
(7). The key matters that were to be Ferucci (21) Arthritis Rheum 2005 USA #2
addressed are: Forslind (22) Ann Rheum Dis 2004 Sweden #5
1) Definition of the test; Gottenberg (23) Ann Rheum Dis 2005 France #2
Grootenboer-Mignot (24) Scand J Rheumatol 2004 France #1
2) Background (historic and method-
Kasapcopur (25) Ann Rheum Dis 2004 Turkey #2
ological considerations);
Lee (26) Ann Rheum Dis 2003 USA #1
3) Clinical use of the test as a diagnos- Lopez-Hoyos (27) Rheumatology 2004 Spain #1
tic tool: Low (28) J Rheumatol 2004 USA #2
- Prevalence of a positive test among Mikuls (31) Arthritis Rheum 2004 USA #4
different patient populations (dis- Nielen (33) Ann Rheum Dis 2005 The Netherlands #1, #5
eases and country of origin) and Pinheiro (34) Ann Intern Med 2003 Brasil #1
normals; Ratapaa-Dahlqvist (35) Arthritis Rheum 2003 Sweden #3
- Diagnostic test properties (sensitivi- Suzuki (38) Scand J Rheumatol 2003 Japan #1
ty, specificity, positive and negative Vallbracht (39) Ann Rheum Dis 2004 Germany #1
predictive values [PPV and NPV]); Van Gaalen (40) Arthritis Rheum 2004 The Netherlands #3
Van Gaalen (41) Ann Rheum Dis 2005 The Netherlands #1, #5
- Test titer in different populations;
Van Noord (42) Ann Rheum Dis 2005 The Netherlands #2
- Effect of the titer on the perform-
Vander Cruyssen (44) Ann Rheum Dis 2005 Belgium #2
ance characteristics of the test;

S-70
Anti-CCP2 antibody testing in RA/ J.P. Riedemann et al.

America, 1 in Japan and 1 in Korea. Further laboratory work has been per- 6 reports, Axis-Shield in 5 reports, and
The first author, journal and year of formed, and a new generation of highly Inova Diagnostics in 1 report; in 1 re-
publication, place of the study and purified synthetic peptides, containing port the manufacturer was not speci-
scope of the paper are summarized in cyclic citrullinated residues called fied.
Table I. The present review will ad- CCP2, was introduced at the beginning The cut-off point to define a positive
dress topics 1 to 4 listed above. of 2002. This new synthetic peptide is test varied from >3.8 IU to 50 IU; 2 stu-
used in the commercial test that is cur- dies did not specify the cut-off point
Definition of the test rently available. The anti-CCP2 test is used.
Anti-CCP corresponds to a microtitre now available primarily from 3 differ-
based enzyme immunoassay for the in ent manufacturers: Sensitivity and specificity
vitro detection of antibodies in human 1) Euro-Diagnostica – The Netherlands The diagnostic properties of the anti-
serum or plasma that react with synthe- (http://www.eurodiagnostica.com), CCP2 test evaluated in each report are
tic peptides containing citrullinate resi- 2) Axis-Shield – UK (http://www.axis- also presented in Table II. Results var-
dues. shield.com), and ied depending upon the specific char-
An antibody system directed against a 3) Inova Diagnostics – USA (http:// acteristics of the RApatients studied.
protein component of the keratohyaline www.inovadx.com). The sensitivity of the anti-CCP2 test
granules in the cytoplasm of buccal reported for established RA patients
mucosa cells was first described in 1. Clinical use of anti-CCP2 as a varied from 64.4% (39) to 96% (18). In
1964, referred to as anti-perinuclear diagnostic test early RApatients, the sensitivity varied
factor (APF) (46). Despite reasonable Thirteen publications have addressed from 14.4% (39) to 83.5% (38). In the
sensitivity and specificity, the test nev- the usefulness of CCP2 in the diagnosis report of elderly onset RA patients, the
er achieved widespread use perhaps of RA (15, 17-20, 24, 26, 27, 33, 34, reported sensitivity was 64.7% (27). It
owing to technical difficulties. 38, 39, 41). was not possible to make a summary
In 1979, a new group of RA-specific Patients estimate or weighted average of the
antibodies was described. They were Most studies included RA patients de- sensitivity of anti-CCP2 testing, due to
referred to as antikeratin antibodies fined according the ACR 1987 criteria. the substantial heterogeneity among
(AKA). These antibodies react and The characteristics of patients included individual studies.
stain keratin-like structures in the in study populations are described in Patients included in control groups to
cornified layer of esophagus cryostat Table II. Five of the 13 publications in- assess the specificity of the CCP2 test
section. Different studies have shown cluded in their sample both patients also had different characteristics. In 6
that both APF and AKA react with the with early and established RA (18, 20, studies, the control group included both
same antigen moiety, the protein filag- 24, 38, 39). Four publications studied patients with rheumatic diseases and
grin (47). only established RA, patients with me- normal individuals (17, 18, 2 0 ,2 7 ,3 8 ,
Filaggrin (filament-aggregatin protein) dian disease duration of 5 to 14.6 years 39). In 5 studies, patients with various
is produced during the late stages of (15, 17, 19, 34). One publication stud- rheumatic diseases, including inflam-
epithelial cell differentiation. It is first ied samples of blood donated before matory and non-inflammatory condi-
synthesized as a phosphorylated pre- clinical diagnosis of RA (33). In one tions, but no normal individuals, were
cursor protein (profilaggrin), which is publication, elderly-onset RA and clas- studied (19, 26, 38, 39, 41). In 3 stud-
partly dephosphorylated and then sical RA patients were studied (27). ies, the type of rheumatic diseases in-
cleaved in 10-12 filaggrin subunits dur- One recent report evaluated CCP2 in a cluded in the control group were only
ing differentiation of epithelial cells. At cohort of early arthritis patients (41), inflammatory conditions (18,20,24). In
the conclusion of the process, about and in one report no data were provided 3 studies, the control group consisted
20% of the arginine residues are con- about the type of RA patients studied only of patients with only one specific
verted into citrulline by action of the (26), only that the patients had a mean condition, including patients with hep-
enzyme peptidylarginine deaminase. age of 55.4 years. atitis C virus infection (HCV) (15), pa-
In 1998, Schellekens (5) demonstrated The female/male rate of the study pop- tients with polymyalgia rheumatica
that citrulline was a major constituent ulation was described in 9 reports; the (PMR) (27), and patients with undiffer-
of antigenic determinants recognized percentage of female patients in those entiated arthritis (33). Among healthy
by antibodies present in RA sera. In studies range from 55% to 84.9%. The individuals included as part of the con-
2000, the same authors published the average age was reported in 9 studies, trol groups, in most studies the preva-
results of the first ELISA anti-CCP test and ranged from 45.9 to 64 years. The lence of anti-CCP2 antibodies ranged
study, in which they evaluated sera of average disease duration was from 5 to from 0 to 0.6%. One study including
patients with RA, non RA rheumatic 14.6 years in established RA patients, only 10 healthy controls reported a
diseases (RD), and some infectious dis- and 0.4 years for the early RApatients. 10% prevalence (18).
eases, reporting excellent specificity Of the 13 papers in which anti-CCP2 The specificity values reported ranged
(98%) and reasonable sensitivity (68%) was evaluated as a diagnostic test, the from 88.9% (38) to 100% (27) (Table
(48). manufacturer was Euro-Diagnostica in II).

S-71
Anti-CCP2 antibody testing in RA/ J.P. Riedemann et al.

S-72
Anti-CCP2 antibody testing in RA/ J.P. Riedemann et al.

Table III.

Author Test * Patients Sensitivity Controls Specificity

Dubucquoi 1, 2, 3 46 RA 1 = 85% 22 CTDs 1 = 90.9%


2 = 82% 2 = 90.9%
3 = 85% 3 = 95.5%

Author Test* Patients Prevalence of anti-CCP2 Controls Prevalence


Low 2 and 3 66 JIA 2 = RF (+) polyarthritis = 75% 9 adult RA 2 = RA= 0%
RF (-) polyarthritis/oligoarthritis = 0% SLE = 0%
3 = RF (+) polyarthritis = 90% 34 adult and child SLE Healthy = 0%
RF (-) polyarthritis/oligoarthritis = 25% 25 Healthy 3 = RA= 56%
SLE = 24%
Healthy = 0%

*Test manufacturer: 1 = Eurodiagnostica; 2) Axis-Shield; 3) INOVA.

Prevalence and use of anti-CCP2 3 test manufacturers in 46 patients with selected controls matched for age, sex,
in other rheumatic diseases RAand 22 patients with connective tis- date of sampling and residential area.
Seven studies have evaluated the pre- sue damage (CTD) as control. The sen- The overall sensitivity of tests for anti-
valence of anti-CCP2 in other rheumat- sitivity and specificity results obtained CCP2 antibodies performed prior to the
ic disease. Two studies involved pa- from the 3 different providers were development of arthritis among the
tients with PsA (14, 44), with reported very similar (Table III). A recent study samples of RA patients was 33.7%;
prevalences of positive anti-CCP2 anti- compared the performance of kits from specificity was 98.2%.
bodies of 15.7% and 7.8% respectively. different providers in 66 patients with In a subanalysis of the same cohort,
Two studies involved patients with Sjö- juvenile inflammatory arthritis (JIA) Berglin et al. analyzed the presence of
gren's syndrome and the prevalence of and 68 control subjects (28); most con- the shared epitope, anti-CCP2 antibod-
positive results was 7.5% (23) and 1.2% trol subjects were CCP2 negative with ies' and tests for rheumatoid factor
(42). Two other publications reported the Axis-Shield test, but only healthy (RF) in the same group of individuals
prevalence between 2% (25) and 90% controls were negative with the INOVA who subsequently developed RA. Be-
in patients with juvenile idiopathic arth- test (Table III). cause they were interested in the value
ritis; the variability may have been ac- of the shared epitope, only 59 RA pa-
counted for by the type of arthritis (28). Anti-CCP2 as a predictor of future tients who had blood samples available
Finally, in one study patients with juve- development of RA for DNA analysis were considered. In
nile arthritis were assessed and the pre- Three studies have addressed the value this group, the sensitivity of anti-CCP2
valence was 5.65% (21). of anti-CCP2 antibodies as a potential antibodies as predictor of future devel-
predictor of the future development of opment of RA was 37%, with a speci-
Comparison of test results using RA(12, 35, 40) (Table IV). ficity of 98%. In a logistic regression
different manufacturers Ratapaa-Dahlqvist et al. (35) designed analysis, anti-CCP2 antibodies had the
Two studies tested the same serum a nested case control study using 2 Swe- highest predictive value, with an odds
samples using kits from different man- dish cohorts. They identified 83 RA pa- ratio (OR) of 15.9 (12).
ufacturers. One report concerning the tients, who had donated blood samples Van Gaalen et al. studied a cohort of
usefulness of anti-CCP2 as a diagnostic prior to the development of RA, and 936 consecutive patients with recent-
test (20), compared the results from the compared these cases with randomly onset arthritis. Of the original 936 pa-

Table IV.

OR
Author (ref.) Sample/Design Follow-up Sensitivity Specificity (95% CI)

Ratapaa-Dahlqvist Nested case-control study Retrospective analysis of blood samples 33.7% 98.2% No data
(35) 83 RApatients collected before disease onset

Berglin Same group as above; Same as above 37.0% 98.0% 15.9


(12) subset of 59 RApatients

van Gaalen 936 recent onset arthritis; 3 years 50.4% 98.4% 37.8
(40) 318 patients with UAstudied (13.8 – 111.9)

S-73
Anti-CCP2 antibody testing in RA/ J.P. Riedemann et al.

tients, 346 could not be classified with with an OR of 14.8 (95% CI 7.2 to tients. The most fair and accurate esti-
any specific diagnosis after 2 weeks of 30.2). mate of the sensitivity of a diagnostic
evaluation. These patients were consid- In a recent report, van Gaalen et al. test will emerge from studies of pa-
ered to have undifferentiated arthritis (41) compared the diagnostic accuracy tients who are truly representative of
(UA), and were followed for 3 years.. and prognostic value of anti-CCP1 vs the population to which the test will be
Among 318 of the original 346 avail- anti-CCP2 tests in 467 early arthritis applied in clinical practice.
able for analysis after 3 years, 40% patients with a median symptom dura- In this review, the sensitivity of anti-
(127 of 318) had developed RA as de- tion of 3 months. Of those, 153 had RA CCP2 antibodies appears reasonably
fined by ACR criteria. Among 69 UA when evaluated over 4 years. Radio- good in studies which evaluated estab-
patients who had anti-CCP2 (+) at graphs of hands and feet were taken at lished RA patients, and varied from
baseline, 64 (93%) developed RA (OR baseline, 6 months, and years 1, 2, 3 64.4% to 96%. However, the sensitivi-
37.8, 95% CI 13.8-111.9) (40). and 4, which were available in 91 of ty was as low as 14.4% in early RA or
the 153 patients. A high rate of joint UA, in which clinicians may require
Possible associations of anti-CCP2 damage over a period of 4 years was more capacity to make a definitive di-
with radiographic damage seen in patients who were CCP2 (+) agnosis. The specificities ranged from
The utility of anti-CCP2 antibodies to (mean 7.3 points, SD 4.6, p = 0.003), 88.9% to 100%. However, the highest
identify patients who had radiographic compared with those who were nega- specificity (100%) was seen in a study
progression has been analyzed in 4 tive for anti-CCP2 (mean 1.6 Sharp- in which the control group is rather
reports (19, 22, 33, 41). De Ricke et al. points per year; SD 3.1). In regression small and included only patients with
reported a higher progression rate am- analysis which included the shared epi- polymylagia rheumatica and healthy
ong 117 CCP2 (+) patients compared to tope, anti-CCP1, anti-CCP2, and IgM individuals.
63 CCP2 (-) patients (p = 0.001) (19), RF, anti-CCP2 antibodies were the Analysis of published studies suggests
according to a modified Larsen score most significant predictor of joint dam- that is at least debatable whether anti-
(from 0 to 160 points) divided by the age. CCP2 antibodies are a substantial ad-
disease duration in years. Forslin et al. vance as a diagnostic tool. Further stu-
(22), reported on 379 patients who had Discussion dies are required to determine the pos-
disease durations ≤ 1 year and anti- The diagnosis of RAmay be difficult in sible advantages of the test. However,
CCP2 analyses. At baseline, the medi- early patients who may not have devel- currently it is reasonable to suggest that
an Larsen score was 5 (0 to 11) in the oped typical manifestations of RA. An the anti-CCP2 antibody test should not
208 CCP2 + patients versus 2 (0 to 10 early definitive diagnosis is desirable be used as a screening test to detect
/25th to 75th centile) among the 171 for early aggressive treatment. Clini- RA. One possible exception is psoriatic
CCP2 – patients (p = 0.008). At the end cians have been particularly interested arthritis, where an important minority
of the follow up, the Larsen score was in this new group of anti-CCP2 anti- of patients have been shown to have
15 (5 to 27) in the CCP2 + patients ver- bodies, which appear to improve early anti-CCP2 antibodies in 2 studies.
sus 5 (0 to 14) in the CCP2 – patients (p diagnostic capacities. In the present Beyond diagnostic sensitivity and spe-
= 0.0005). There was also a higher review, we identified 13 published cificity, one must consider possible pos-
change score (radiological progression) studies in which anti-CCP2 antibodies itive and negative predictive values in a
from baseline to end point of 12 in CCP were evaluated as a diagnostic test in clinical decision. The positive predic-
+ patients (4 to 25) versus 4 (0 to 12) in RA. Important differences were seen in tive value is the likelihood that an indi-
CCP – patients, p = 0.0005). the characteristics of the patients evalu- vidual with a positive test result actual-
The best predictor of both radiological ated, as well as the cut points to define ly has the particular disease of interest,
joint damage and progression, in uni- a positive test in individual studies. whereas the negative predictive value
variate and multiple analyses, was the These differences may explain the wide is the likelihood that an individual with
Larsen score followed by anti-CCP2. range of sensitivity results reported. a negative test result actually does not
Nielen et al. in a study compared the The sensitivity of a diagnostic test (the have the disease. These characteristics
usefulness of antibodies to citrullinated proportion of true positives in a group are highly dependant on the clinical
human fibrinogen vs anti-CCP2 (33), of individuals with a certain condition scenario in which the test is used. In
in consecutively gathered 379 early of interest) is a specific property of the other words, even with known and
arthritis patients (258 RAand 121 UA) test, and should remain constant. How- fixed sensitivity and specificity for the
who were followed over 2 years. At the ever, the calculated sensitivity will vary test, the results will vary depending on
end of the period, they had complete depending on characteristics of the the pretest probability.
data in 296 (78.1%) patients. Radiolog- patients used to evaluate it. If the sam- To illustrate this point, we analyze an
ical progression was evaluated at 2 ple studied includes only more severely example with a sensitivity of 75% and
years using the Sharp/van der Heijde affected patients, the sensitivity will a specificity of 95% (Table V). Results
method. With logistic regression analy- probably be higher than in populations are presented using pre-test odds, posi-
sis, they identified the anti-CCP2 (+) as with patients who have milder disease, tive and negative likelihood ratios and
the best predictor of x-ray progression or a group of more heterogeneous pa- post test odds, as well as predictive val-

S-74
Anti-CCP2 antibody testing in RA/ J.P. Riedemann et al.

ues. If the test is used in a group of were also seen in the control groups; all It remains unknown whether titers of
patients with a high probability of hav- patients were CCP2 negative with the anti-CCP2 antibodies are associated
ing the disease (established and typical Axis-Shield test with only healthy indi- with higher risk of radiological pro-
clinical picture of RA - estimated pre- viduals negative and 56% of RA and gression.
test probability 80% - pre-test odds = 24% of SLE patients positive with the
References
4), the positive predictive value will be INOVA test, which all were negative 1. LEE DM, WEINBLATT ME: Rheumatoid arth-
98.4% and the negative predictive val- with the Axis-Shield tests. ritis. Lancet 2001; 358: 903-11.
ue 48.7%. However, if the probability Another issue is the cut off value used 2 WILES NJ, SYMMONS DPM, HARRISON B et
of having or not having the disease is to define a positive result. As presented al.: Estimating the incidence of rheumatoid
arthritis: trying to hit a moving target. Arthri -
even (pre-test probability 50% - pre- in the results, the cut off varied signifi- tis Rheum 1999; 42: 1339-46.
test odds = 1), a positive test result will cantly in various reports of studies con- 3. WEINBLATT ME: Rheumatoid arthritis: More
determine a positive predictive value of ducted with a CCP2 test, even when aggressive approach improves outlook. Cleve
88.2% and a negative predictive value provided by the same manufacturer. Clin J Med 2004; 71: 409-13.
4. VAN BOEKEL MAM, VOSSENAAR ER, VAN DEN
of 78.3%. A third situation could be if Some standardization would appear HOOGEN FHJ, VAN VENROOIJ WJ: Autoanti-
the same test is applied to a population desirable. body systems in rheumatoid arthritis: speci-
of undifferentiated arthritis (estimated In addition to diagnostic utility, anti- ficity, sensitivity and diagnostic value. Arthri -
pre-test probability of RA 20% - pre- CCP2 antibodies have been evaluated tis Res 2002; 4: 87-93.
5. SCHELLEKENS GA, DE JONG BAW, VAN DEN
test odds 0.25), the positive predictive as predictors of future development of HOOGEN FHJ, VAN DE PUTTE LBA: Citrulline
value will be 65.2% and the negative RA and as predictors of radiological is an essential constituent of antigenic deter-
predictive value 93.5%. These differen- damage, using 2 different approaches. minants recognized by rheumatoid arthritis-
specific autoantibodies. J Clin Invest 1998;
ces illustrate that predictive values and Two studies (12,35) evaluated the pres-
101: 273-81.
the post-test odds are as important as ence of anti-CCP2 antibodies in blood 6.E URO-DIAGNOSTICA B.V: Beijerinckweg 18,
sensitivity and specificity in evaluating samples donated prior to the onset of L-6827 BN Arnhem, The Netherlands. http://
performance of a laboratory test in clin- disease in normal individuals. The sen- www.eurodiagnostica.com
7. Guidelines for immunologic laboratory test-
ical care. sitivity of the anti-CCP2 to identify ing in rheumatic diseases: An introduction.
An additional concern may involve the future RA was estimated at 33.7% to Arthritis Rheum 2002; 47: 429-33
comparability of different available 37%, with a specificity of 98%, an odds 8. ALESSANDRI C, BOMBARDIERI M, PAPA N
anti-CCP2 tests evaluated in only two ratio of 15.9. In the other study (40), et al.: Decrease of anti-cyclic citrullinated
peptide antibodies and rheumatoid factor fol-
reports. In the study of Dubucquoi (20), anti-CCP2 identified an odds ratio of lowing anti-TNFa therapy (infliximab) in
reagents from 3 manufacturers yielded future development of RA of 37.8 in rheumatoid arthritis is associated with clinical
comparable results in terms of sensitiv- patients with undifferentiated arthritis. improvement. Ann Rheum Dis 2004; 63:
ity and specificity. By contrast, some Thus, a normal individual or a patient 1218-21
9. AVCIN T, CIMAZ R, FALCINI F et al.: Preva -
important differences were observed in with undifferentiated arthritis with a lence and clinical significance of anti-cyclic
the study of Low (28), which compared positive anti-CCP2 antibody has a sub - citrullinated peptide antibodies in juvenile
results of tests performed with the stantial risk of future development of idiopathic arthritis. Ann Rheum Dis 2002; 61;
Axis-Shield and the INOVA reagents. RA. 608-11.
10. BAS S, GENEVAY S, MEYER O, GABAY C :
In patients with juvenile arthritis, the The 4 individual studies that have Anti-cyclic citrullinated peptide antibodies,
prevalence of anti-CCP2 in patients addressed radiological progression all IgM and IgArheumatoid factors in the diag-
with rheumatoid factor positive pol- agree that the presence of anti-CCP2 nosis and prognosis of rheumatoid arthritis.
Rheumatology 2003; 42: 677-80.
yarthritis varied from 75% with the antibodies is associated with greater
11. BAS S, PERNEGER TV, SEITZ M, TIERCYJM,
Axis-Shield test to 90% with the INO- radiographic progression, with an odds ROUX-LOMBARD P, GUERNE PA: Diagnostic
VA test. In patients with rheumatoid ratio of 14.8 (95% CI 7.2 to 30.2). The tests for rheumatoid arthritis: comparison of
factor negative polyarthritis/oligoarth- presence of anti-CCP2 antibodies should anti-cyclic citrullinated peptide antibodies,
anti-keratin antibodies and IgM rheumatoid
ritis, the prevalence of anti-CCP was be considered as a predictor of future factor. Rheumatology 2002; 41: 809-14.
0% with the Axis-Shield tests, versus development of RA as well as marker 12. BERGLIN E, PADYUKOV L, SUNDIN U et al.:
25% with the INOVA test Differences of progressive radiologic progression. A combination of autoantibodies to cyclic cit-

Table V.

Pre-test Pre-test Post-test odds


probability odds Sensitivity Specificity LR+ (*) LR- PPV(**) NPV(**) of a (+) result

80% 4 75% 95% 15 0.263 98.4% 48.7% 60


50% 1 75% 95% 15 0.263 88.2% 78.3% 15
20% 0.25 75% 95% 15 0.263 65.2% 93.5% 3.75

*LR: Likelihood ratio; **PPV: positive predictive value; NPV: negative predictive value

S-75
Anti-CCP2 antibody testing in RA/ J.P. Riedemann et al.

rullinated peptide (CCP) and HLA-DRB1 LAND P, DELAUNAY C, MEYER O, CHOL- tibodies in palindromic rheumatism. An abor-
locus antigens is strongly associated with fu- LET-MARTIN S, LABARRE C: Second gener- tive form of rheumatoid arthritis ? Rheum -
ture onset of rheumatoid arthritis. Arthritis ation anti-cyclic citrullinated peptide (anti- atology 2003; 42: 972-75.
Res 2004; 6: R303-08. CCP2) antibodies can replace other filaggrin 37. SARAUX A, BERTHELOT JM, DEVAUCH-
13. BOBBIO-PALLAVICINI F, ALPINI C, CAPO- antibodies and improve rheumatoid arthritis ELLE V et al.: Value of antibodies to citrulli-
RALI R, AVALLE S, BUGATTI S, MONTECUC- diagnosis. Scand J Rheumatol 2004: 33: 218- nated-containing peptides for diagnosing ear-
CO C: Autoantibody profile in rheumatoid 20. ly rheumatoid arthritis. J Rheumatol 2003;
arthritis during long-term infliximab treat- 25. KASAPCOPUR O, ALTUN S, ASLAN A et al.: 30: 2535-9.
ment. Arthritis Res 2004; 6: R264-272. Diagnostic accuracy of anti-cyclic citrullinat- 38. SUZUKI K, SAWADA T, MURAKAMI A et al.:
14. BOGLIOLO L, ALPINI C, CAPORALI R et al.: ed peptide antibodies in juvenile idiopathic High diagnostic performance of ELISA de-
Antibodies to cyclic citrullinated peptides in arthritis. Ann Rheum Dis 2004; 63: 1687-9. tection of antibodies to citrullinated antigens
psoriatic arthritis. J Rheumatol 2005; 32: 511- 26. LEE DM, SCHUR PH: Clinical utillity of anti- in rheumatoid arthritis. Scand J Rheumatol
15. CCP assay in patients with rheumatic dis- 2003; 32: 197-204.
15. BOMBARDIERI M, ALESSANDRI C, LABBA- eases. Ann Rheum Dis 2003; 62: 870-4. 39. VALLBRACHT I, RIEBER J, OPPERMANN M,
DIA G et al.: Role of anti-cyclic citrullinated 27. LOPEZ-HOYOS M, RUIZ DE ALEGRIA, BLAN - FORGER F, SIEBERT U, HELMKE K : Diag-
peptide antibodies in discriminating patients CO R et al.: Clinical utility of anti-CCP anti- nostic and clinical value of anti-cyclic citrul-
with rheumatoid arthritis from patients with bodies in the differential diagnosis of elderly- linated peptide antibodies compared with
chronic hepatitis C infection-associated pol- onset reumatoid arthritis and polymyalgia rheumatoid factor isotypes in rheumatoid
yarticular involvement. Arthritis Res 2004; 6: rheumatica. Rheumatology 2004; 43: 655- arthritis. Ann Rheum Dis 2004; 63: 1079-
R137-141. 657. 1084.
16. CARAMACHI P, BIASI D, TONOLLI E, PIER- 28. LOW JM, CHAUHAN AK, KIETZ DA, DAUD 40. VAN GAALEN FA, LINN-RASKER SP, VEN-
OPAN S et al.: Antibodies against cyclic cit- U, PEPMUELLER PH, MOORE TL: Determi- ROOIJ WJ et al .: Autoantibodies to cycli cit-
rullinated peptides in patients affected by nation of anti-cyclic citrullinated peptide anti- rullinated peptides predict progression to
rheumatoid arthritis before and after inflix- bodies in the sera of patients with juvenile rheumatoid arthritis in patients with undiffer-
imab treatment. Rheumatol Int 2005 February idiopathic arthritis. J Rheumatol 2004; 31: enciated arthritis. A prospective cohort study.
On Line. 1829-33. Arthritis Rheum 2004; 50: 709-15.
17. CHOI SW, LIM MK, SHIN DH, PARK JJ, SHIN 29. MADDALI S, MANETTI R, MELCHIORRE D 41. VAN GAALEN FA, VISSER H, HUIZINGA TW:
SC : Diagnostic performances of anti-cyclic et al.: Anti-cyclic citrullinated Peptide anti- A Comparison of the diagnostic accuracy and
citrullinated peptides antibody and antifilag- bodies are highly associated with severe bone prognostic value of the first-and second anti-
grin antibody in Korean patients with rheu- lesions in rheumatoid arthritis anti-CCP and cyclic citrullinated peptides autoantibody
matoid arthritis. J Korean Med Sci 2005; 20: bone damage in RA. Autoimmunity 2004; 37: (CCP1 and CCP2) test for rheumatoid arthri-
473-8. 495-501. tis. Ann Rheum Dis published online 30 Mar
18. CORREA PA, TOBON GJ, CITERA G et al.: 30. MEYER O, LABARRE C, DOUGADOS M et 2005
Anticuerpos anti-CCPen artritis reumatoidea: al.: Anticitrullinated protein/peptide antibody 42. VAN NOORD C, HOOIJKAAS H, VAN DEN
relacion con c aracteristicas clinicas, citoci- assays in early rheumatoid arthritis for pre- GOORBERGH BCM et al.: Diagnostic value of
nas Th1/Th2 y HLADRB1. Biomedica 2004: dicting five years radiographic damage. Ann anti-cyclic citrullinated peptide antibodies to
24: 140-52. Rheum Dis 2003; 62: 120-6. detect rheumatoid arthritis in patients with
19. D E RYCKE L, PEENE I, HOFFMAN IEA et al.: 31. MIKULS TR, O'DELL JR, STONER JA et al.: Sjögren's syndrome. Ann Rheum Dis 2005;
Rheumatoid factor and anti citrullinated pro- Association of rheumatoid arthritis treatment 64: 160-2.
tein antibodies in rheumatoid arthritis: diag- response and disease duration with declines 43. VAN ROSSUM M, VAN SOESBERGEN, DE KORT
nostic value, associations with radiological in serum levels of IgM rheumatoid factor and et al .: Anti-cyclic citrullinated peptide (anti-
progression rate, and extra-articular manifes- anti-cyclic citrullinated peptide antibody. CCP) antibodies in children with juvenile
tations. Ann Rheum Dis 2004: 63: 1587-93. Arthritis Rheum 2004; 50: 3776-82. idiopathic arthritis. J Rheumatol 2003; 30:
20. DUBUCQUOI S, SOLAU-GERVAIS E, LE- 32. NIELEN MMJ, VAN SCHAARDENBURG D, 825-9.
FRANC D et al.: Evaluation of anti-citrulli- REESINK HW et al.: Specific autoantibodies 44. VANDER CRUYSSEN B, HOFFMAN IEA, ZMI-
nated filaggrin antibodies as hallmarks for the precede the symptoms of rheumatoid arthri- ERCZAK HG et al .: Anti-citrullinated peptide
diagnosis of rheumatic diseases. Ann Rheum tis. A study of serial measurements in blood antibodies may occur in patients with psoriat-
Dis 2004: 63: 415-19. donors. Arthritis Rheum 2004; 50: 380-6. (in- ic arthritis. Ann Rheum Dis Online 4 Feb
21. FERUCCI ED, MAJKA DS, PARRISH L et al.: house test). 2005.
Antibodies against cyclic citrullinated pep- 33. NIELEN MMJ, VAN DER HORST AR, VAN 45. ZENG X, AI M, TIAN X et al.: Diagnostic val-
tide are associated with HLA-DR4 in simplex SCHAANDENBURG D et al.: Antibodies to ue of anti-cyclic citrullinated peptide anti-
and multiplex polyarticular-onset juvenile citrullinated human fibrinogen (ACF) have body in patients with rheumatoid arthritis. J
rheumatoid arthritis. Arthritis Rheum 2005: diagnostic and prognostic value in early arth - Rheumatol 2003; 30: 1451-5.
52: 239-46. ritis. Ann Rheum Dis Published online Jan- 46. NIENHUIS RLF, MANDENA EA : A new ser-
22. FORSLIND K, AHLMENM, EBERHARDT K, uary2005. um factor in patients with rheumatoid arthri-
HAFSTROM I, SVENSSON B for the BARFOT 34. PINHEIRO GC, SCHEIBERG MA, APARECIDA tis. The antiperinuclear factor. Ann Rheum
STUDY GROUP: Prediction of radiological out- DA SILVAM : Anti-cyclic citrullinated peptide Dis 1964, 23: 302-5
come in early rheumatoid arthritis in clinical antibodies in advanced RA. Ann Intern Med 47. HOET RM, VAN VENROOIJ WJ : The antiperin-
practice: role of antibodies to citrullinated 2003; 139: 234-5. uclear factor and antikeratin antibodies in
peptides (anti-CCP). Ann Rheum Dis 2004: 35. RANTAPAA-DAHLQVIST S, DE JONG BAW, rheumatoid arthritis. In SMOLEN J, KALDEN
63: 1090-5. BERGLIN E et al.: Antibodies against cyclic J and MAINIRN (Eds.): Rheumatoid Arthritis.
23. GOTTENBERG J-E, MIGNOT S, NICAISSE- citrullinated peptide and IgArheumatoid fac- Berlin, Springer Verlag 1992: 299-318.
ROLLAND P et al .: Prevalence of anti-cyclic tor predict the development of rheumatoid 48. SCHELLEKENS GA, VISSER H, DE JONG BAW
citrullinated peptide and anti-keratin antibod- arthritis. Arthritis Rheum 2003; 48: 2741-9. et al.: The diagnostic properties of rheuma-
ies in patients with primary Sjögren's syn- 36. SALVADOR G, GOMEZ A, VIÑAS O et al.: toid arthritis antibodies recognizing a cyclic
drome. Ann Rheum Dis 2005: 64: 114-17. Prevalence and clinical significance of anti- citrullinated peptide. Arthritis Rheum 2000;
24. GROOTENBOER-MIGNOT S, NICAISE-RO- cyclic citrullinated peptide and antikeratin an- 43: 155-63.77

S-76