Article Text

Download PDFPDF

Concise report
Smoking is a major preventable risk factor for rheumatoid arthritis: estimations of risks after various exposures to cigarette smoke
  1. Henrik Källberg1,
  2. Bo Ding1,
  3. Leonid Padyukov2,
  4. Camilla Bengtsson1,
  5. Johan Rönnelid3,
  6. Lars Klareskog2,
  7. Lars Alfredsson1,4,
  8. EIRA Study Group
  1. 1Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden
  2. 2Department of Medicine, Rheumatology Unit, Karolinska Institutet/Karolinska Hospital, Stockholm, Sweden
  3. 3Unit of Clinical Immunology, Uppsala University/Akademiska Sjukhuset, Uppsala, Sweden
  4. 4Karolinska Institutet School of Public Health, Stockholm, Sweden
  1. Correspondence to Dr Henrik Källberg, Institute of Environmental Medicine, Box 210, Karolinska Institutet, 171 77 Stockholm, Sweden; henrik.kallberg{at}


Background Earlier studies have demonstrated that smoking and genetic risk factors interact in providing an increased risk of rheumatoid arthritis (RA). Less is known on how smoking contributes to RA in the context of genetic variability, and what proportion of RA may be caused by smoking.

Objectives To determine the association between the amount of smoking and risk of RA in the context of different HLA-DRB1 shared epitope (SE) alleles, and to estimate proportions of RA cases attributed to smoking.

Design, Setting and Participants Data from the Swedish Epidemiological Investigation of Rheumatoid Arthritis (EIRA) case–control study encompassing 1204 cases and 871 controls were analysed.

Main Outcome Measure Estimated OR to develop RA and excess fraction of cases attributable to smoking according to the amount of smoking and genotype.

Results Smoking was estimated to be responsible for 35% of anticitrullinated protein/peptide antibody (ACPA)-positive cases. For each HLA-DRB1 SE genotype, smoking was dose-dependently associated with an increased risk of ACPA-positive RA (p trend <0.001). In individuals carrying two copies of the HLA-DRB1 SE, 55% of ACPA-positive RA was attributable to smoking.

Conclusions Smoking is a preventable risk factor for RA. The increased risk due to smoking is dependent on the amount of smoking and genotype.

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.

Smoking is the most established environmental risk factor for developing rheumatoid arthritis (RA).1 One hypothesis on the effect of smoking is that smoking causes citrullination of peptides and in the context of RA susceptibility genes contribute to the elicitation of immunity to these citrullinated proteins/peptides and eventually to the onset of RA.2,,5

It is now of interest not only to decipher the aetiology of RA in the light of this gene–environment interaction, but also to take a public health perspective in determining the number of cases of RA attributable to smoking in different genetic contexts. We have therefore used our population-based study, the Epidemiological Investigation of Rheumatoid Arthritis (EIRA), to estimate the relative risk of RA conferred by different amounts of smoking in the context of different HLA-DRB1 genotypes, and to estimate the excess fraction (EF) of RA cases attributed to smoking.


Set up

EIRA is a population based case–control study. Information was collected from incident RA cases and controls matched for age, gender and residential area recruited between May 1996 and December 2003. We enrolled cases aged 18–70 years from 19 clinics located in the south and middle parts of Sweden. Almost all of the rheumatology units in the study area participated in the study. Each case was diagnosed according to the American College of Rheumatology 1987 criteria for RA6 by rheumatologists at the corresponding clinic, and was invited to participate in the study. Controls were randomly selected from a population register with consideration taken as to sex, age and resdential area and then sent a questionnaire by post. We managed to collect genetic, antibody and questionnaire information from 1205 cases (85%) and 872 controls (52%). We asked five questions regarding smoking: present and previous smoking; time point for start and/or stop of smoking and amount of cigarettes smoked per day.

We classified the amount of smoking into three groups (0–9, 10–19 and 20 pack years). One pack year is equivalent to smoking 20 cigarettes per day for 1 year.

We genotyped shared epitope (SE) alleles for participants who contributed blood samples. SE was defined as DRB1*01, DRB1*04 or DRB1*10 in the HLA-DRB1 locus.7

Antibody levels were measured by using the Immunoscan-RA Mark 2 (Euro-Diagnostica, Malmo, Sweden) enzyme-linked immunosorbent assay and the cut-off limit for anti-citrullinated protein/peptide antibody (ACPA)-positive RA was set to 25 U/ml. Details on study design, data collection, exposure information, genotyping and serological analysis are given elsewhere.2 8,,10

Ethical approval was obtained from relevant ethical committees and all the participants consented to contribute to the study on a voluntary basis.

Statistical analysis

We calculated OR of developing RA associated with different categories of smoking and the presence of SE alleles together with 95% CI by using logistic regression models. The interaction between smoking and the presence of SE alleles was evaluated as a departure from additivity of effects,11 12 and was estimated by calculating the attributable proportion due to interaction.12

All analyses were adjusted for the matching variables, age, sex and residential area.

A trend test for a dose–response relationship regarding smoking and the risk of developing ACPA-positive RA was performed separately for each SE allele category as suggested by Armitage et al.13

We calculated the excess fraction (of cases) attributable to smoking in per cent (EF%)14 as an indictor of the relevance of smoking as a risk factor for RA in the population. EF was calculated in relation to RA overall as well as to ACPA-positive RA and different HLA-DRB1 genotypes.

SAS version 9.1 for Windows was used to analyse all data.


A total of 61% of RA cases was ACPA positive. We have chosen to focus on ACPA-positive RA because we have not found any association between smoking, SE alleles or their combination regarding increased risks of developing ACPA-negative RA (the OR associated with smoking was 0.6 (95% CI 0.4 to 1.0) and with SE 0.8 (95% CI 0.4 to 1.7), as reported previously.2

Smoking and risk of ACPA-positive RA

Different amounts of pack years smoked without considering SE allele status were associated with ACPA-positive RA in a dose–response manner(p trend <0.0001), with the highest OR for ever smokers who had smoked 20 or more pack years (table 1).

Table 1

Characteristics of ACPA-positive RA cases and controls

For ex-smokers the increased risk of ACPA-positive RA was observed to decrease with the duration of time since smoking cessation (p trend <0.0001).

For intermediate ever-smokers (pack years 10–19), the increased risk of ACPA-positive RA diminished almost to the level of never-smokers 20 years after smoking cessation. Among heavy smokers, a relatively high OR was still observed even 20 years after the cessation of smoking (table 1).

Public health impact of smoking in terms of EF of cases attributable to smoking

We calculated the EF of cases attributable to smoking as an indicator of the relevance of smoking as a public health risk factor. For ACPA-positive RA, the EF attributable to smoking was 35% (95% CI 25% to 45%; 31% for women and 42% for men). The EF attributable to smoking for RA overall (ACPA-positive and ACPA-negative RA combined) was 20% (95% CI 7% to 26%), which indicates that smoking plays an important role in the occurrence of RA overall because ACPA-positive RA is the most common form of RA.

As smoking interacts with SE alleles (table 2, figure 1) we also calculated the EF of cases attributable to smoking by HLA-DRB1 SE genotype (table 3). Among ACPA-positive RA cases with double SE alleles 55% (95% CI 39% to 67%) could be attributed to smoking.

Figure 1

OR for different amounts of smoking (pack years (p.y.)) in combination with none (no shared epitope (SE)), one (single SE) or two (double SE) copies of SE alleles. The reference group was non-smokers without SE alleles.

Table 2

OR and attributable proportions due to interaction for different doses of smoking and SE alleles, regarding the risk to develop ACPA-positive RA

Table 3

Proportion of cases attributable to smoking EF by SE alleles, regarding ACPA-positive RA, ACPA-negative RA and total RA


This report describes smoking and the risk of RA in a novel way regarding the impact of smoking both concerning its contribution to RA and concerning the dose–dependent addition of risk conferred by smoking in individuals with different numbers of HLA-DRB1 risk alleles. We also provide novel data on the effects of cessation of smoking concerning the future risk of RA.

The demonstration that approximately one third of cases of ACPA-positive RA, the most severe variant of RA, appear to be attributed to smoking illustrates the impact of smoking as a major cause of RA at the population level. Notable is that this attribution appears to be higher in men than in women. The smoking attribution to RA is, however, smaller than the smoking attribution to lung cancer, which is estimated to be as high as 90%,15 but similar to that seen for ischaemic heart disease.16

The data on interactions between smoking and HLA-DR alleles in providing a risk of RA in the present report are in accordance with those published from studies in several different countries,4 5 17 including studies from the Nurses Health Cohort Study.5 In one other US-based case-only study the effects of smoking was more limited but still statistically significant.18

Other environmental and genetic factors may modify the effects of smoking on the development of RA. There are indications that airborne exposures such as silica and other air pollutants may enhance or possibly dilute the effects of smoking.19 There are also other lifestyle factors such as alcohol consumption or hormonal factors influencing the risk of RA that may interact with smoking.2021 The precise effects of smoking on RA development may thus vary considerably in various populations, but so far smoking has been shown to be an essential risk factor for RA in a majority of published studies.

We used case–control methodology to generate data regarding the risk of RA as a whole as well for ACPA-positive and ACPA-negative RA in the context of various HLA-DR genes and smoking. The fact that more than 50% of RA cases can be attributed to smoking in individuals carrying two copies of the HLA-DRB1 SE genes illustrates drastically how smoking may affect disease risk differently in different individuals. Although these types of data should not be taken as an argument for the genotyping of healthy individuals, they may provide a rationale for specific counselling against smoking for individuals with a family history of RA.

The problem of selection bias always encountered in case–control studies was minimised in the current study by the population-based design and by the observation that smoking habits did not differ between controls with and without blood samples. This, and the fact that the identification of controls was made by matching for age, sex and residential area in the same population that generated the cases, makes us confident that the results are reliable from a methodological standpoint.

In conclusion, the data presented in this paper on the impact of smoking on the development of RA, warrant more active information on the association between smoking and RA to the general public as well as to relatives of patients with RA.

To what extent cessation of smoking will be able to alleviate an already ongoing RA is as yet incompletely known, although it has been demonstrated that smoking contributes to cardiovascular disease, which is the major cause of premature death in RA.22 There are many reasons for the medical community to communicate the known facts on smoking and RA, with the aim of reducing the incidence of smoking and preventing RA and its consequences.


The authors would like to thank the following people for contributing to this report: all participating patients and controls; in EIRA, for recruiting patients: Ingeli Andréasson, Landvetter; Eva Baecklund, Akademiska Hospital; Ann Bengtsson and Thomas Skogh, Linköping Hospital; Birgitta Nordmark, Johan Bratt and Ingiäld Hafström, Karolinska University Hospital; Kjell Huddénius, Rheumatology Clinic in Stockholm City; Shirani Jayawardene, Bollnäs Hospital; Ann Knight, Hudiksvall Hospital and Uppsala University Hospital; Ido Leden, Kristianstad Hospital; Göran Lindahl, Danderyd Hospital; Bengt Lindell, Kalmar Hospital; Christin Lindström and Gun Sandahl, Sophiahemmet; Björn Löfström, Katrineholm Hospital; Ingmar Petersson, Spenshult Hospital; Christoffer Schaufelberger, Sahlgrenska University Hospital; Patrik Stolt, Västerås Hospital; Berit Sverdrup, Eskilstuna Hospital; Olle Svernell, Västervik Hospital; Tomas Weitoft, Gävle Hospital; for excellent data collection: Marie-Louise Serra and Lena Nise, who provided invaluable contributions to the collection of data and maintenance of the database. The sponsors of the studies did not have any role in the study design, data collection, data analysis, data interpretation or in writing the report.


View Abstract


  • Funding The EIRA study was supported by grants from the Swedish Medical Research Council, from the Swedish Council for Working Life and Social Research, from King Gustaf V's 80-year Foundation, from the Swedish Rheumatism Foundation, from Stockholm County Council, from the insurance company AFA, from the EU-supported AutoCure project, FAMRI (Flight Attendant Medical Research Institute), NIH (P60 AR047782), and from the COMBINE (Controlling Chronic Inflammatory Diseases with Combined Efforts) project.

  • Competing interest None.

  • Patient consent Obtained.

  • Ethics approval This study was conducted with the approval of the relevant ethical committees and all the participants consented to contribute to the study on a voluntary basis.

  • Provenance and peer review Not commissioned; externally peer reviewed.