De Rooy et al.

Arthritis Research & Therapy 2011, 13:R180

http://arthritis-research.com/content/13/5/R180

RESEARCH ARTICLE Open Access

Can anti-cyclic citrullinated peptide antibody-

negative RA be subdivided into clinical
subphenotypes?
Diederik PC De Rooy1*, Annemiek Willemze2 , Bart Mertens1, Tom WJ Huizinga1 and
Annette HM Van der Helm-van Mil1

Abstract
Introduction: Studies investigating genetic risk factors for susceptibility to rheumatoid arthritis (RA) studied anti-
citrullinated peptide antibody (CCP)-positive RA more frequently than anti-CCP-negative RA. One of the reasons for
this is the perception that anti-CCP-negative RA may include patients that fulfilled criteria for RA but belong to a
wide range of diagnoses. We aimed to evaluate the validity of this notion and explored whether clinical
subphenotypes can be discerned within anti-CCP-negative RA.
Methods: The 318 patients with anti-CCP-negative RA (1987 ACR criteria), included in the Leiden Early Arthritis
Clinic between 1993 and 2006, were studied for baseline characteristics and radiologic progression data during a
mean follow-up of 5 years. Grouping was studied both at variable and patient levels. Principal components analysis
and partial least-squares regression were applied to study for clustering of variables. A cluster analysis was
performed to look for clustering of patients.
Results: The simultaneous presence of patient characteristics at disease presentation was observed for several
groups; however, the three largest groups of patients’ characteristics explained only 26.5% of the total variance.
Plotting the contribution of each patient to these three groups did not reveal clustering of patients. Comparable
observations were made when data on progression of joint destruction were studied in relation to baseline clinical
data. A cluster analysis, evaluating whether patients resemble each other, revealed no grouping of patients.
Altogether, no clinically distinguishable subphenotypes were observed.
Conclusions: The current data provide evidence that, for risk-factor studies, anti-CCP-negative RA patients can be
studied as one group.

Introduction negative RA. This subdivision was based on differences

Rheumatoid arthritis (RA) has been considered a rela- in genetic risk factors, histopathologic differences, and
tively homogeneous clinical syndrome for more than 50 differences in outcome of anti-CCP-positive and anti-
years. However, our current view of RA as a single dis- CCP-negative RA [4].
ease may become untenable, and the disease may be Several successful genome-wide studies for genetic
subdivided into a range of disorders based on improved risk factors for anti-CCP-positive RA have been per-
knowledge of its driving immunologic markers [1-3]. formed. Studies on genetic risk factors for anti-CCP-
During the last decade, a number of studies suggested negative RA are thus far lacking. One of the reasons for
that RA can be divided into two syndromes: anti-citrulli- this is the fear of phenotypic misclassification, as anti-
nated peptide antibody (CCP)-positive and anti-CCP- CCP-negative RA is often considered to be a heteroge-
neous disease [1,5]. For future risk-factor, translational,
and outcome studies on the subgroup of anti-CCP-nega-
* Correspondence: d.p.c.de_rooy@lumc.nl tive RA patients, it is essential to provide epidemiologic
1
Department of Rheumatology, Leiden University Medical Center, Leiden,
The Netherlands and clinical evidence on whether anti-CCP-negative RA
Full list of author information is available at the end of the article

© 2011 De Rooy et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons
Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
De Rooy et al. Arthritis Research & Therapy 2011, 13:R180 Page 2 of 7
http://arthritis-research.com/content/13/5/R180

can be considered one entity. In this study, we therefore experienced reader. The intraclass observer correlation
aimed to determine whether the group of anti-CCP- coefficient for the radiographic progression rate was
negative RA patients can be separated into clinically dis- 0.97. The radiographic SHS progression per year was
tinguishable subphenotypes. calculated for each patient by using a linear regression
analysis with the following formula: Y = a + bx. All
Materials and methods available radiographs per patient were used to estimate
Patients a patient’s progression rate (the b in the equation). At a
The 704 patients who were included between 1993 and group level, the median SHS values (± SD) at the subse-
2006 in the Leiden Early Arthritis Clinic and who were quent time points were 5.0 (9.8) at baseline; 7.0 (13.3) at
diagnosed with RA according to the 1987 ACR criteria 1-year follow-up; 9.0 (15.6) at 2 years; 9.0 (17.2) at 3
were selected; 318 patients had anti-CCP-negative RA years; and 10.5 (22.4) at 5 years of follow-up.
and were therefore selected for further analysis. The Lei-
den Early Arthritis Clinic previously has been described Statistical analysis
extensively [6]. In short, it is a population-based incep- To investigate whether clinical subphenotypes can be
tion cohort of patients presenting with arthritis to the discerned, two types of analyses were done. First, we
Department of Rheumatology of the Leiden University studied whether groups of patient characteristics fre-
Medical Center. This is the only referral center in a quently occur together; such clustering at the variable
health care region of approximately 400,000 inhabitants. level was studied by using the variable reduction techni-
Written informed consent was obtained from all ques Principal Components Analysis (PCA) and Partial
patients, and the cohort was approved by the local med- Least Squares regression (PLS). Second, we studied
ical ethical committee (Ethics Committee of the Leiden whether subgroups of patients can be discerned; such
University Medical Center). At first visit, the rheumatol- grouping at the patient level was studied by using a
ogist completed a questionnaire regarding the present- cluster analysis.
ing symptoms, as reported by the patient: type, PCA
localization and distribution of initial joint symptoms, Some overlap between clinical variables is extremely
duration and course of the initial symptoms, and the common (for example, a high swollen-joint count will
presence of inflammatory back pain and skin abnormal- often be accompanied by a high number of tender
ities. The patient’s smoking history and family history joints). PCA makes use of such overlap and combines
were assessed. Patients rated morning stiffness in min- variables that frequently occur together into compo-
utes (mean, 103; SD, 112). The Health Assessment nents. In this way, the number of variables explaining
Questionnaire (HAQ) was used to provide an index of data can be reduced, which makes datasets easier to
disability. A 44-joint count for swollen joints (SJC) was interpret. The components resulting from the PCA are
performed. Anti-CCP2 antibodies were measured in sera based on the observed variance and not on predefined
collected at baseline with enzyme-linked immunosor- hypotheses, making this technique suitable for exploring
bent assay (ELISA) (Immunoscan RA Mark 2; Eurodiag- unknown relations between variables. For each compo-
nostica, Arnhem, The Netherlands). Samples with a nent, the loading of each variable to the component is
value less than 25 units/ml were considered negative, provided. Loadings > 0.4 are generally considered rele-
according to the manufacturer’s instructions. IgM-Rheu- vant. For each component, an observed variance is pre-
matoid Factor (RF) was determined with ELISA. RF sented, indicating the percentage of the total variance in
titers ranged from 0 to 200 IU/ml. For the analyses, RF clinical variables that is explained by this component.
levels titers were divided into three groups: RF normal, Here, a PCA was performed, and the contribution of
RF moderately increased (1 to 3 times the reference each patient to the most important components was
value), and RF highly increased (> 3 times reference plotted to look for clustering of patients. The PCA was
value) [7]. Anti-modified citrullinated vimentin (MCV) performed by using the following baseline variables: age
antibodies were also measured with ELISA (Orgentec at inclusion, gender, symptom duration, acuteness of the
Diagnostika, Mainz, Germany); here, the cutoff level was onset of symptoms (subacute, within 1 week, or insi-
20 arbitrary units, according to the manufacturer’s dious, over more than 1 week), morning stiffness, fati-
instruction. All the mentioned baseline variables were gue, fever, smoking, family history of RA, three variables
studied here, as they may, on their own or in combina- on the distributions of involved joints (upper extremi-
tion with other characteristics, point to different disease ties, lower extremities, or both; symmetric or asym-
subsets. Annual radiographs of the hands and feet were metric; large joints, small joints, or both), C-reactive
taken during a mean follow-up period of 5 years (mini- protein, RF, the presence of baseline erosions, the num-
mum, 0; maximum, 14 years) and scored according to ber of swollen joints, anti-MCV positivity, inflammatory
the Sharp-van der Heijde method (SHS) by an back pain, and skin abnormalities.
De Rooy et al. Arthritis Research & Therapy 2011, 13:R180 Page 3 of 7
http://arthritis-research.com/content/13/5/R180

PLS patient characteristics, but grouping of patients. In other

A limitation of PCA is that it cannot take outcome mea- words, given all characteristics that were available of the
sures into account. Because patient response obtained patients, this evaluated which patients resemble each
during the disease course is at least of equal importance other. Accordingly, not the variables denoting them, but
to baseline characteristics for the aim of the present the patients themselves are subject to combination into
study, we also performed a Partial Least Squares regres- clusters. Hierarchic clustering was performed by using
sion. PLS does an analysis that is comparable to PCA, Gene Cluster, as described by Eisen et al. [9]
but that has the advantage that it makes use also of out-
come measures [8]. Here, a PLS regression was applied Results
to the same variables as included in the PCA as inde- PCA
pendent variables, but with the addition of the radiolo- The baseline characteristics of the anti-CCP-negative RA
gic damage over time as a dependent variable. This patients are depicted in Table 1. Entering baseline vari-
analysis allowed us to assess whether variance between ables into a PCA resulted in nine components. The first
patients can be characterized by distinguishable sub- component explained 10.0% of the variance, the second
groups. Also here, identified factors were plotted to look component, 8.6%, and the third component, 8.0%. The
for clustering, which may represent clinical subpheno- relative importance of the variables contributing to the
types. The PCA and PLS analyses were done by using different components is depicted in Table 1. For exam-
SPSS version 17.0 (SPSS Inc., Chicago, IL, USA). To ple, in the first component, the variables age, gender,
perform PLS, the relevant SPSS extension packages were and the presence of baseline erosions were grouped. In
downloaded from the links provided by the official SPSS the second component, the involvement of small joints
website. versus the involvement of large joints or both SJC and
Cluster analysis CRP were grouped. The component scores for indivi-
Subsequently, a cluster analysis was performed, which, dual patients of factors 1 through 3 were plotted against
in contrast to PCA and PLS, evaluates not grouping of each other, and no evident clustering was observed

Table 1 Baseline characteristics of the study population and the factors found in PCA
Component
Variable Baseline frequency 1 2 3 4 5 6 7 8 9
Age at inclusion (years, mean ± SD) 59.2 (16.2) 0.761
Female gender (n, %) 219 (68.9) 0.448 0.394
Subacute onset of symptoms (versus insidious) 188 (59.1) 0.592 -0.306 0.316
Morning stiffness (min; mean ± SD) 103.18 (112.0) -0.496 0.484
Fatigue (VAS; mean ± SD) 45.1 (29.9) -0.389 0.522
Symptom duration (days; mean ± SD) 172.1 (180.2) 0.394 -0.349 0.455
Family history of RA (n, %) 62 (19.5) -0.326 0.594 0.380
Past or present smoking (n, %) 128 (40.3) 0.652 0.375
Fever (n, %) 22 (6.9) -0.319 -0.426
a
Involvement of small/large joints (n, %) 0.604
Symmetry of involved joints (n, %) 216 (67.9) -0.379 0.308 0.331 0.370
b
Involvement of upper/lower extremities 0.602 -0.348
Inflammatory back pain (n, %) 16 (5.0) 0.466 0.531 0,321
Skin abnormalities (N, %)c 56 (18.1) 0.374
Swollen joint count (mean ± SD)d 11.3 (8.5) 0.445 0.365
CRP (mg/L; mean ± SD) 30.10 (34.4) 0.315 0.434 0.512
RF (IU/ml) (mean ± SD)e 7.12 (19.5) 0.305 -0.345 0.315 -0.602
MCV positivity (n, %) 59 (18.6) 0.470 0.376 -0.321
Erosive disease at baseline (n, %) 202 (63.5) 0.642 0.332
Explained variance per component (%) 10.0 8.6 8.0 7.4 6.8 6.5 5.6 5.5 5.4
Involvement of upper extremities: shoulder, elbow, wrist, or hand joints; involvement of lower extremities: joints of hip, knee, ankle, feet or toe joints.
a
Involvement of small/large joints: in 163 patients (51.3%), only small joints of hands and feet were involved. In 53 patients (16.7%) only large joints were
involved. In 93 patients (29.2%), both small and large joints were involved.aInvolvement of upper/lower extremities: in 156 patients (49.1%) only joints in the
upper extremities were involved. In 28 patients (8.8%), only joints in the lower extremities were involved. In 96 (30.2%) patients, joints in both lower and upper
extremities were involved.cSkin abnormalities: absence or presence of dermatological abnormalities such as psoriatic lesions, ulcers, rheumatoid nodules and so
on.dSwollen Joint Count: 44-Swollen Joint Count.eAnalyses were performed on RF in groups (less than reference value, 1 to 3 times reference value, or more than
3 times reference value). CRP, C-reactive protein; MCV, mutated citrullinated vimentin; RF, rheumatoid factor; VAS, visual analogue scale.
De Rooy et al. Arthritis Research & Therapy 2011, 13:R180 Page 4 of 7
http://arthritis-research.com/content/13/5/R180

(Figure 1a-c). As the first three components explain observed variation. The major important variables in the
most variance, these components were plotted. In these first latent factor were gender, symmetry of involved
plots, each dot indicates one single patient. For example joints, rheumatoid factor positivity, anti-MCV positivity,
in Figure 1a, a dot indicates how much the variance in age at inclusion, symptom duration at inclusion, and the
an individual patient is described by factor 1 (age, gen- presence of baseline erosions. The major variables in the
der, and the presence of baseline erosions) in relation to second factor were the same; this suggests that little dif-
factor 2 (involvement of small joints versus the involve- ference exists. The individual patient scores on these
ment of large joints or both SJC and CRP). factors were plotted against each other. Also here, no
clustering was observed (Figure 1d).
PLS Because of the absence of clustering, we sought a
Apart from baseline characteristics, outcome data may positive control, to verify that the method and data used
be informative to identify differences between patient do allow finding clusters. To this end, the PLS with
populations, so we next performed a data-reduction baseline and radiologic progression data was repeated
method that allows assessing radiologic-outcome data in on the total group of 704 RA patients instead of on the
addition to baseline data. To identify subsets of patients, subgroup of anti-CCP-negative patients. Anti-CCP sta-
PLS regression was used. With PLS, two latent factors tus was not included in this analysis, so that the analysis
were found that together accounted for 30.1% of the was not influenced by this variable. Again, two factors

" "

!" " !" "

!" !"

" "

!" "
!" "
#

!"
!"

Figure 1 Plots of the most important component loadings from PCA and PLS on 318 anti-CCP-negative RA patients. In these plots,
each dot indicates one single patient. Component scores indicate how strongly each component is represented in each patient. For example, in
(a), a dot indicates how much the variance in an individual patient is being described by factor 1 on the x-axis (age, gender, and the presence
of baseline erosions) in relation to factor 2 on the y-axis (involvement of small joints versus the involvement of large joints or both SJC and
CRP). If a concurrence of components was found, clustering of patients would be visible. In the PCA, clinical variables at disease onset were
explored. The same applies for the factors in PLS regression. In the PLS regression, the clinical variables at disease onset were explored together
with radiologic data on progression of joint destruction during a mean of 5 years of disease. CRP, C-reactive protein; PCA, principal components
analysis; PLS, partial least squares regression; RA, rheumatoid arthritis; SJC, swollen joint count.
De Rooy et al. Arthritis Research & Therapy 2011, 13:R180 Page 5 of 7
http://arthritis-research.com/content/13/5/R180

were found, explaining together 10.1% of the observed used did distinguish CCP-positive and CCP-negative
variance. The main variables of these components were patients as identifiable subphenotypes. Therefore, these
gender, rheumatoid factor, age at inclusion, and CRP. data do not support the hypothesis that anti-CCP-nega-
Clustering the first two factors revealed two clusters; tive RA is composed of different subsets, but rather pro-
one for anti-CCP-positive patients and one for anti- vide evidence that anti-CCP-negative RA can be
CCP-negative patients (see Figure 2). regarded as one disease, and therefore, risk-factor stu-
dies in anti-CCP-negative RA are feasible.
Cluster analysis The data evaluated concerned a wide variety of clinical
Finally, we explored whether anti-CCP-negative patients characteristics, such as the acuteness of the onset of
can be grouped into subgroups of patients with similar symptoms, the distribution of involved joints, the sever-
characteristics. To this end, a heat map was made in ity of fatigue, fever, skin abnormalities, inflammatory
which the patients with the most similarity clustered back pain, acute-phase reactants, and radiologic baseline
together. Cluster analysis showed no clustering of and progression data. The variables assessed are, in our
patients in the heat map, and this finding is supported view, variables that might in combination form patterns
by the dendrogram (Figure 3). Therefore, also with these characteristic of different disease subsets. However,
analyses, no distinguishable groups of patients with despite the evaluation of a large range of characteristics
similar characteristics were recognized. at disease presentation and the evaluation of long-term
radiologic follow-up data, no clear subphenotypes were
Sensitivity analysis discerned. We cannot exclude that when other variables
Patients negative for anti-CCP antibodies can harbor are assessed, conclusions might be different.
other autoantibodies. Here 24.2% of the anti-CCP-nega- To test whether the currently used data and methods
tive patients were positive for RF, and 18.6% were posi- are able to find clinical subsets, we also studied the total
tive for anti-MCV. It can be argued that it is more RA population, including 704 patients, of whom 318
appropriate to perform the analyses on patients negative were anti-CCP-negative patients. We observed that PLS
for anti-CCP, RF, and anti-MCV. Therefore, all analyses regression is able to discriminate between anti-CCP-
were repeated in this subgroup (n = 171). Similar obser- positive and anti-CCP-negative disease. This is in line
vations were made (data not shown). with published data that anti-CCP-positive RA has more
progressive joint destruction during the disease course
Discussion than does anti-CCP-negative RA [10]
This study determined whether anti-CCP-negative The present study did not aim to find statistically sig-
patients fulfilling the 1987 ACR criteria for RA can be nificant associations of baseline variables with the out-
subdivided into clinical subphenotypes and explored come. The present study also does not give any
extensive phenotypic characteristics at baseline, as well indication on whether the pathogenesis of anti-CCP-
as data on progression in joint destruction during the negative RA is heterogeneous or homogeneous between
disease course. In addition, several methods were patients. We explored whether clinical data provide evi-
applied, intending to find subgroups of either variables dence that different groups of patients compose the
or patients. With any method used, no clearly distin- group of anti-CCP-negative RA patients. If subclinical
guishable clusters were observed, although the methods phenotypes had been identified, this is relevant for

$+%&,--(.$+%&/)0&"*
!
!"#$%&'"
()*&%&'"

!! !

!!
Figure 2 Plots of the two major component loadings from PLS on the whole Leiden Early Arthritis Clinic (n = 704). Each dot indicates
one patient. Component scores indicate how strongly each component is represented in each patient. Patients positive for anti-CCP antibodies
are blue, whereas negative patients are red. CCP, citrullinated peptide antibody; PLS, partial least squares regression.
De Rooy et al. Arthritis Research & Therapy 2011, 13:R180 Page 6 of 7
http://arthritis-research.com/content/13/5/R180

Inflammatory back pain

Fever
Skin abnormalities
Acute onset
Smoking
Morning Stiffness
Fatigue
Symptom duration
Erosions at baseline
Age at inclusion
Swollen Joint Count
CRP
Gender
Symmetric Inflammation
Upper extremities involved
Lower extremities involved
Positive family history of RA

Figure 3 Heat map of the cluster analysis of 318 anti-CCP-negative RA patients. Heat map representing the presence or absence of
disease characteristics in individual patients. To make variables comparable, all values were transformed into binary values. For those variables on
a continuous scale, the following cut-offs were made: Morning Stiffness, ≤60 minutes; Fatigue, fatigue rated more than mean (45.1) on Visual
Analogue Scale; symptom duration more than or ≤12 weeks age at inclusion, age greater than mean age (59.2 years); swollen joint count, more
than four swollen joints; CRP, CRP greater than reference value (10 mg/L). The dendrograms depict the relative strength of correlations between
the variables and the patients, respectively. CCP, citrullinated peptide antibody; CRP, C-reactive protein; RA, rheumatoid arthritis; SJC, swollen joint
count; VAS, visual analogue scale.

future pathophysiological studies. Then it could be sug- Abbreviations

CCP: citrullinated peptide antibody; CRP: C-reactive protein; HAQ: Health
gested that, to prevent phenotypic misclassification,
Assessment Questionnaire; MCV: modified citrullinated vimentin; PCA:
studies should be done on anti-CCP-negative subpheno- principal components analysis; PLS: partial least-squares regression; RA:
types. The present observation of a lack of phenotypic rheumatoid arthritis; RF: rheumatoid factor; SHS: Sharp-van der Heijde Score;
SJC: swollen joint count; VAS: visual analogue scale.
heterogeneity within anti-CCP-negative RA suggests that
future studies on pathogenic mechanisms underlying Acknowledgements
anti-CCP-negative RA can be done on the total popula- The work of AHM van der Helm-van Mil is supported by a grant from the
Dutch Organization of Health Scientific Research and Development. A
tion of anti-CCP-negative patients.
Willemze’s work is supported by the Dutch Organization for Scientific
Research (AGIKO grant).
Conclusions This work is further supported by grants from the Masterswitch-project and
the Dutch Arthritis Foundation.
Based on the present data, we suggest that risk factors
studied on the anti-CCP-negative RA patients can be Author details
1
performed on the total group of 1987 ACR criteria-posi- Department of Rheumatology, Leiden University Medical Center, Leiden,
The Netherlands. 2 Department of Statistics and Bioinformatics, Leiden
tive, anti-CCP-negative RA patients.
University Medical Center, Leiden, The Netherlands.
De Rooy et al. Arthritis Research & Therapy 2011, 13:R180 Page 7 of 7
http://arthritis-research.com/content/13/5/R180

Authors’ contributions
DR performed the statistical analyses and wrote the first version of the
manuscript. AW and BM assisted with the statistical analyses. AH and TH
were responsible for the selection of patients. All authors contributed to
revising and adjusting the manuscript. All authors have read and approved
the manuscript for publication.

Competing interests
The authors declare that they have no competing interests.

Received: 6 July 2011 Revised: 17 August 2011

Accepted: 27 October 2011 Published: 27 October 2011

References
1. van der Helm-van Mil AH, Huizinga TW: Advances in the genetics of
rheumatoid arthritis point to subclassification into distinct disease
subsets. Arthritis Res Ther 2008, 10:205.
2. Pratt AG, Charles PJ, Chowdhury M, Wilson G, Venables PJ, Isaacs JD:
Serotyping for an extended anti-citrullinated peptide autoantibody
panel does not add value to CCP2 testing for diagnosing RA in an early
undifferentiated arthritis cohort. Ann Rheum Dis 2011, 70:2056-2058.
3. Padyukov L, Seielstad M, Ong RT, Ding B, Ronnelid J, Seddighzadeh M,
Alfredsson L, Klareskog L: A genome-wide association study suggests
contrasting associations in ACPA-positive versus ACPA-negative
rheumatoid arthritis. Ann Rheum Dis 2011, 70:259-265.
4. van Oosterhout M, Bajema I, Levarht EW, Toes RE, Huizinga TW, van
Laar JM: Differences in synovial tissue infiltrates between anti-cyclic
citrullinated peptide-positive rheumatoid arthritis and anti-cyclic
citrullinated peptide-negative rheumatoid arthritis. Arthritis Rheum 2008,
58:53-60.
5. Klareskog L, Catrina AI, Paget S: Rheumatoid arthritis. Lancet 2009,
373:659-672.
6. de Rooy DP, van der Linden MP, Knevel R, Huizinga TW, van der Helm-van
Mil AH: Predicting arthritis outcomes: what can be learned from the
Leiden Early Arthritis Clinic? Rheumatology (Oxford) 2011, 50:93-100.
7. Nell VP, Machold KP, Stamm TA, Eberl G, Heinzl H, Uffmann M, Smolen JS,
Steiner G: Autoantibody profiling as early diagnostic and prognostic tool
for rheumatoid arthritis. Ann Rheum Dis 2005, 64:1731-1736.
8. Chun H, Ballard DH, Cho J, Zhao H: Identification of association between
disease and multiple markers via sparse partial least-squares regression.
Genet Epidemiol 2011, 35:479-486.
9. Eisen MB, Spellman PT, Brown PO, Botstein D: Cluster analysis and display
of genome-wide expression patterns. Proc Natl Acad Sci USA 1998,
95:14863-14868.
10. van der Helm-van Mil AH, Verpoort KN, Breedveld FC, Toes RE, Huizinga TW:
Antibodies to citrullinated proteins and differences in clinical
progression of rheumatoid arthritis. Arthritis Res Ther 2005, 7:R949-R958.

doi:10.1186/ar3505
Cite this article as: De Rooy et al.: Can anti-cyclic citrullinated peptide
antibody-negative RA be subdivided into clinical subphenotypes?
Arthritis Research & Therapy 2011 13:R180.

Submit your next manuscript to BioMed Central

and take full advantage of:

• Convenient online submission

• Thorough peer review
• No space constraints or color figure charges
• Immediate publication on acceptance
• Inclusion in PubMed, CAS, Scopus and Google Scholar
• Research which is freely available for redistribution

Submit your manuscript at

www.biomedcentral.com/submit