Article Text

THU0040 In Rheumatoid Arthritis, Smoking is not Associated with Anti-Citrullinated Protein Antibodies (ACPA) Per SE, but with the Concurrent Presence of Rheumatoid Factor, Acpa and Anti-Carbamylated Protein Antibodies
  1. T.J. van Wesemael1,
  2. A. Muhammad1,
  3. D.P. Symmons2,3,
  4. A.J. MacGregor4,
  5. A. Barton3,5,
  6. L.A. Trouw1,
  7. T.W.J. Huizinga1,
  8. R.E.M. Toes1,
  9. S.M.M. Verstappen2,
  10. D. van der Woude1
  1. 1Leiden University Medical Center, Leiden, Netherlands
  2. 2Arthritis Research UK Centre for Epidemiology
  3. 3Arthritis Research UK Centre for Genetics and Genomics, Manchester
  4. 4Norwich Medical School, Norwich
  5. 5NIHR Manchester Musculoskeletal Biomedical Research Unit, Manchester, United Kingdom


Background In rheumatoid arthritis (RA) a biological hypothesis has been proposed linking smoking with citrullination, anti-citrulline autoimmunity and ACPA-positive RA1. An interaction has been described between smoking and the HLA-DRB1 shared epitope (SE) alleles for ACPA-positive RA, but this interaction has not been investigated in relation to the concurrent presence of other RA-associated antibodies.

However, smoking is associated with several other autoantibodies in various autoimmune diseases such as polymoysitis2 and systemic lupus erythematosus3, which raised the question if in RA, smoking is solely associated with ACPA or with autoantibodies in general.

Objectives To investigate if smoking is specifically associated with ACPA-positive RA, or with the presence of other RA-specific autoantibodies as well.

Methods Patients fulfilling the 1987 RA criteria from two independent early arthritis cohorts, the Leiden Early Arthritis Clinic (EAC: n=767) and the Norfolk Arthritis Register (NOAR: n=761), were used. Information on smoking (ever versus never) and autoantibody status: rheumatoid factor (RF), ACPA and anti-carbamylated protein antibodies (CarP), was collected at baseline.

Logistic regression analysis (odds ratios (ORs) and 95% CI) was performed to assess the association between the number of autoantibodies and smoking status. ORs were calculated using 2x2-comparisons with seronegative RA patients as reference category. In the EAC, interaction between smoking and SE was assessed using the following measures of biological interaction: relative excess risk due to interaction (RERI), the attributable proportion due to interaction (AP) and the synergy index (S).

Results There was no significant association between smoking and seropositive RA in patients who were positive for one or two autoantibodies in both EAC and NOAR. But smoking was significantly associated with triple autoantibody positivity in EAC (OR 1.79, 95%CI 1.26–2.55) and in NOAR (OR 2.96 (1.79–4.90)) (table 1). Furthermore, interaction measures between smoking and SE were not significant for patients with one autoantibody and partly significant for patients with two autoantibodies (S: 2.39 (1.35–4.23); AP: 0.57 (0.34–0.81)), whereas all three measures were significant in the triple positive group (S: 19.37 (12.01–31.25); AP: 0.95 (0.92–0.97); RERI: 357.58 (143.52–571.63)).

Conclusions Smoking and the interaction between smoking and SE is not associated with ACPA-positive RA, but with the concurrent presence of all three RA-associated autoantibodies. These data indicate that current hypotheses on the role of smoking in the pathophysiology of RA may need to be refined by taking the presence of other auto-antibodies into account as well.


  1. Klareskog L. et al. 2006Arthritis Rheum. 54:38-46.

  2. Chinoy H. et al. 2012 Ann Rheum Dis. 71:961-5.

  3. Freemer M.M. et al. 2006 Ann Rheum Dis. 65:581-4.

Disclosure of Interest None declared

Statistics from

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.