Research Article: Anti-citrullinated protein antibody positive rheumatoid arthritis is primarily determined by rheumatoid factor titre and the shared epitope rather than smoking per se

Date Published: July 14, 2017

Publisher: Public Library of Science

Author(s): Dan Murphy, Derek Mattey, David Hutchinson, Graham R. Wallace.


To analyse the relationship between rheumatoid factor (RF) titre, smoking and HLA-DRB1 alleles coding a “shared epitope” (SE) in relation to anti-citrullinated protein antibody (ACPA) positivity in rheumatoid arthritis (RA).

RA patients (n = 658) attending rheumatology clinics in Cornwall, UK (cohort 1) were stratified according to RF and ACPA titre, and smoking pack years at diagnosis. A further 409 RA patients from North Staffordshire, UK (cohort 2) were studied to confirm the relationship between RF levels, smoking and ACPA positivity in relation to SE status.

In cohort 1 there was a trend (p<0.01) of increasing ACPA positivity rates with increasing levels of RF without statistically significant differences between patients who had never smoked and smokers (never smoked: 15/71 (21%) RF -ve, vs. 43/64 (67%) RF weak +ve, vs 88/100 (88%) RF strong +ve, ever smoked: 18/70 (26%) RF -ve vs. 66/83 (80%) RF weak +ve vs. 196/210 (93%) RF strong +ve). No significant gender difference was observed. No significant difference between smoking and ACPA positivity was seen in RF negative patients. Smoking >20 pack years conferred an increased risk of anti-CCP positive RA (158/200 (79%)), compared to having never smoked (146/235 (62%), p = <0.01), but this increased risk correlated with smokers’ RF positivity as the principal determinant on subsequent regression analysis of cohort 2. In cohort 2, ACPA positivity rates significantly increased with RF positivity and carriage of 1 or 2 SE alleles (p<0.01). Little or no relationship was observed in patients lacking SE. ACPA positivity in RA strongly associates with increasing RF titre independent of smoking. This relationship is dependent on carriage of SE alleles. There is no relationship between ACPA and smoking in RF negative patients.

Partial Text

Over the last 2 decades, smoking, HLA-DRB1 alleles that code a “shared epitope” (SE), and anti-citrullinated protein antibodies (ACPA) have emerged as the trinity of RA pathogenesis [1]. However, in routine clinical practice the use of both RF and ACPA remains, and for good reason: the presence of RF and ACPA in healthy individuals increases the risk of RA development over and above ACPA alone [2], and a positive RF and ACPA confers a far poorer radiological prognosis in established RA compared to either of these autoantibodies alone [3]. The 2010 ACR/EULAR RA classification criteria [4] includes a criterion that scores highly for those individuals with a strongly positive RF or ACPA (scoring 3 points), acknowledging that a strongly positive RF is of equal weighting in RA diagnosis to strongly positive ACPA. RF and ACPA co-exist in RA more than would be expected by chance alone: a study of established RA (n = 784) in Sheffield, UK, observed that 93% of RF patients were also ACPA positive [3]. The clustering of RF and ACPA is more pronounced with high titre RF as a study of 102 RA patients observed that 61% of RA patients with a RF >50 U/ml were ACPA positive (the expected frequency is 18.5% if RF>50 U/ml and ACPA occurred independently of each other) as opposed to only 25% of RA patients with a RF <50 U/ml [5]. This raises the possibility that a common process generates high titre RF and ACPA and determines a poorer prognosis in RA. Two cohorts of RA patients were studied. Cohort 1 consisted of males and females attending routine rheumatology clinics at the Royal Cornwall Hospital, Cornwall, UK from February 2015 to August 2016. The data was anonymised at source, collected as part of project IRAS ID 194833, approved by South West Regional Ethical Committee. This cohort was selected to achieve an intended total cohort size of six hundred gender matched patients for analysis. Patients reviewed in clinic with a new or existing diagnosis of RA during the study period were included for analysis, with 60 patients subsequently excluded (see below). 298 male and 300 female RA patients were included to provide a suitable cohort size for serological subgroup analysis, with approximately equal numbers of each gender selected to detect potential differences in smoking rates, age of onset, and serological status rather than to represent the overall male to female prevalence ratio (Table 1). Recruitment ceased when the target cohort size was reached. Data was recorded by clinicians in a standard questionnaire via face-to-face interview. Missing data was obtained by follow up telephone conversation. Age, age of RA onset, disease duration, smoking history prior to disease development, RF levels at disease onset and subsequent ACPA levels were recorded as part of routine clinical practice. Testing for SE status does not form part of routine clinical practice and was therefore unavailable. All patients fulfilled the 2010 ACR/EULAR RA criteria at diagnosis [4]. We have noted several interesting observations on the determination of ACPA positive RA. Smoking was associated with ACPA positivity only in RF positive patients, and especially in patients who were strongly RF positive. Furthermore, there was a strong relationship between increasing RF levels and ACPA positivity in RA, irrespective of smoking (Table 2). Underpinning this relationship is carriage of the SE. In RA patients who did not carry the SE, we found only weak, or no relationship between RF levels and the frequency of ACPA positivity. This suggests that the co-existence of 2 RA-associated autoantibodies cannot be explained by a propensity for antibody production per se as the frequency of ACPA positivity in patients without the SE who were RF negative was not significantly different to those who were RF strongly positive. These results raise important questions about the relationship between RF and ACPA in RA and the potential mechanisms underlying this relationship. We feel that processes involving the simultaneous or sequential generation of both RF and ACPA need to be considered in further studies investigating RA pathogenesis irrespective of smoking.   Source:


0 0 vote
Article Rating
Notify of
Inline Feedbacks
View all comments