Advertisement

Smoking Intensity, Duration, and Cessation, and the Risk of Rheumatoid Arthritis in Women

  • Karen H. Costenbader
    Correspondence
    Requests for reprints should be addressed to Karen H. Costenbader, MD, MPH, Brigham and Women’s Hospital, Division of Rheumatology, Immunology and Allergy, 75 Francis St., Boston, MA 02115.
    Affiliations
    Division of Rheumatology, Immunology, and Allergy, Section of Clinical Sciences, Robert B. Brigham Arthritis and Musculoskeletal Diseases Clinical Research Center, Brigham and Women’s Hospital, Harvard Medical School, Boston, Mass

    Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Mass
    Search for articles by this author
  • Diane Feskanich
    Affiliations
    Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Mass
    Search for articles by this author
  • Lisa A. Mandl
    Affiliations
    Division of Rheumatology, Department of Medicine, Hospital For Special Surgery, Cornell Weill Medical College, New York.
    Search for articles by this author
  • Elizabeth W. Karlson
    Affiliations
    Division of Rheumatology, Immunology, and Allergy, Section of Clinical Sciences, Robert B. Brigham Arthritis and Musculoskeletal Diseases Clinical Research Center, Brigham and Women’s Hospital, Harvard Medical School, Boston, Mass

    Channing Laboratory, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical School, Boston, Mass
    Search for articles by this author

      Abstract

      Background

      Cigarette smoking has been associated with rheumatoid arthritis (RA), but the importance of smoking intensity, duration, and time since quitting, and whether the risk is primarily for rheumatoid factor (RF) seropositive versus seronegative RA are still unclear.

      Methods

      We conducted a prospective analysis of smoking and the risk of RA among 103,818 women in the Nurses’ Health Study. A total of 680 RA cases, diagnosed from 1976 and 2002, were confirmed using a questionnaire and medical record review. Sixty percent were RF positive. Cox proportional hazards models calculated the relative risks (RRs) of RA with smoking, adjusting for reproductive and lifestyle factors.

      Results

      The RR of RA was significantly elevated among current (RR 1.43 [95% confidence interval 1.16-1.75]) and past smokers (RR 1.47 [95% confidence interval 1.23-1.76]), compared with never smokers. The risk of RA was significantly elevated with 10 pack-years or more of smoking and increased linearly with increasing pack-years (P trend <.01). A greater number of daily cigarettes and longer duration of smoking were associated with increased risk. The effect of smoking was much stronger among RF-positive cases than among RF-negative cases. The risk remained elevated in past smokers until 20 years or more after cessation.

      Conclusions

      In this large cohort, past and current cigarette smoking were related to the development of RA, in particular seropositive RA. Both smoking intensity and duration were directly related to risk, with prolonged increased risk after cessation.

      Keywords

      Rheumatoid arthritis (RA) is an autoimmune disease characterized by chronic, destructive, and debilitating arthritis, occasionally with systemic involvement. RA is the most common inflammatory arthritis, affecting approximately 1% of the adult population.
      • Gabriel S.E.
      • Crowson C.S.
      • O’Fallon W.M.
      The epidemiology of rheumatoid arthritis in Rochester, Minnesota, 1955-1985.
      The cause of RA is unknown, but it is presumed that environmental factors contribute to its development in the genetically predisposed.
      Case-control studies over the past 20 years have investigated the relationship between cigarette smoking and the development of RA.
      • Vessey M.P.
      • Villard-Mackintosh L.
      • Yeates D.
      Oral contraceptives, cigarette smoking and other factors in relation to arthritis.
      • Hazes J.M.
      • Dijkmans B.A.
      • Vandenbroucke J.P.
      • de Vries R.R.
      • Cats A.
      Lifestyle and the risk of rheumatoid arthritis cigarette smoking and alcohol consumption.
      • Heliovaara M.
      • Aho K.
      • Aromaa A.
      • Knekt P.
      • Reunanen A.
      Smoking and risk of rheumatoid arthritis.
      • Voigt L.F.
      • Koepsell T.D.
      • Nelson J.L.
      • Dugowson C.E.
      • Daling J.R.
      Smoking, obesity, alcohol consumption, and the risk of rheumatoid arthritis.
      • Symmons D.P.
      • Bankhead C.R.
      • Harrison B.J.
      • et al.
      Blood transfusion, smoking, and obesity as risk factors for the development of rheumatoid arthritis results from a primary care-based incident case-control study in Norfolk, England.
      • Uhlig T.
      • Hagen K.B.
      • Kvien T.K.
      Current tobacco smoking, formal education, and the risk of rheumatoid arthritis.
      • Krishnan E.
      • Sokka T.
      • Hannonen P.
      Smoking-gender interaction and risk for rheumatoid arthritis.
      • Stolt P.
      • Bengtsson C.
      • Nordmark B.
      • et al.
      Quantification of the influence of cigarette smoking on rheumatoid arthritis results from a population based case-control study, using incident cases.
      • Padyukov L.
      • Silva C.
      • Stolt P.
      • Alfredsson L.
      • Klareskog L.
      A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis.
      In these studies, the association of cigarette smoking with the development of RA seems to be stronger for men than for women, and several studies have not found an elevated risk of RA associated with cigarette smoking in women.
      • Hazes J.M.
      • Dijkmans B.A.
      • Vandenbroucke J.P.
      • de Vries R.R.
      • Cats A.
      Lifestyle and the risk of rheumatoid arthritis cigarette smoking and alcohol consumption.
      • Heliovaara M.
      • Aho K.
      • Aromaa A.
      • Knekt P.
      • Reunanen A.
      Smoking and risk of rheumatoid arthritis.
      • Uhlig T.
      • Hagen K.B.
      • Kvien T.K.
      Current tobacco smoking, formal education, and the risk of rheumatoid arthritis.
      • Krishnan E.
      • Sokka T.
      • Hannonen P.
      Smoking-gender interaction and risk for rheumatoid arthritis.
      In addition, the risk of RA associated with smoking has been reported to be highest for rheumatoid factor (RF) seropositive RA
      • Heliovaara M.
      • Aho K.
      • Aromaa A.
      • Knekt P.
      • Reunanen A.
      Smoking and risk of rheumatoid arthritis.
      • Symmons D.P.
      • Bankhead C.R.
      • Harrison B.J.
      • et al.
      Blood transfusion, smoking, and obesity as risk factors for the development of rheumatoid arthritis results from a primary care-based incident case-control study in Norfolk, England.
      • Krishnan E.
      • Sokka T.
      • Hannonen P.
      Smoking-gender interaction and risk for rheumatoid arthritis.
      • Stolt P.
      • Bengtsson C.
      • Nordmark B.
      • et al.
      Quantification of the influence of cigarette smoking on rheumatoid arthritis results from a population based case-control study, using incident cases.
      • Padyukov L.
      • Silva C.
      • Stolt P.
      • Alfredsson L.
      • Klareskog L.
      A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis.
      and in those who carry a copy of the human leukocyte antigen, DR subregion (HLA-DR) shared epitope.
      • Padyukov L.
      • Silva C.
      • Stolt P.
      • Alfredsson L.
      • Klareskog L.
      A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis.
      • Both increasing duration and intensity of cigarette smoking increase the risk of rheumatoid arthritis in women.
      • With >10 pack years of smoking, the risk of rheumatoid arthritis rises and does so in a dose-dependent manner.
      • After smoking cessation, it takes 20 years for the risk of rheumatoid arthritis to return to that of never smokers.
      • Most of the risk associated with smoking is for rheumatoid factor positive rheumatoid arthritis.
      Cohort studies also have examined cigarette smoking as a risk factor for RA. In 1990, Hernandez Avila and colleagues
      • Hernandez Avila M.
      • Liang M.H.
      • Willett W.C.
      • et al.
      Reproductive factors, smoking, and the risk for rheumatoid arthritis.
      reported an increased risk of RA and undifferentiated polyarthritis in a subset of women who currently smoked 15 to 24 cigarettes per day in the Nurses’ Health Study, although they did not find any significant associations with having ever smoked or in other subgroups. In a retrospective analysis in the Women’s Health Study, a randomized controlled trial of low-dose aspirin and vitamin E for the prevention of cardiac disease, increased duration of cigarette smoking, more so than smoking intensity, was associated with an increased risk of RA.
      • Karlson E.W.
      • Lee I.M.
      • Cook N.R.
      • Manson J.E.
      • Buring J.E.
      • Hennekens C.H.
      A retrospective cohort study of cigarette smoking and risk of rheumatoid arthritis in female health professionals.
      In the Iowa Women’s Health Study, a cohort study in elderly women, both increasing intensity and duration of smoking were associated with RA. The effect of smoking on the risk of RA seemed to last up to 10 years after smoking cessation, after which the risk of RA returned to that of never smokers.
      • Criswell L.A.
      • Merlino L.A.
      • Cerhan J.R.
      • et al.
      Cigarette smoking and the risk of rheumatoid arthritis among postmenopausal women results from the Iowa Women’s Health Study.
      Exposure to passive cigarette smoke (environmental tobacco) has been associated with chronic diseases including asthma,
      • Eisner M.D.
      Environmental tobacco smoke and adult asthma.
      cancer,
      • Brennan P.
      • Buffler P.A.
      • Reynolds P.
      • et al.
      Secondhand smoke exposure in adulthood and risk of lung cancer among never smokers a pooled analysis of two large studies.
      • Tredaniel J.
      • Boffetta P.
      • Saracci R.
      • Hirsch A.
      Environmental tobacco smoke and the risk of cancer in adults.
      and heart disease,
      • Glantz S.A.
      • Parmley W.W.
      Passive smoking and heart disease. Epidemiology, physiology, and biochemistry.
      • Law M.R.
      • Morris J.K.
      • Wald N.J.
      Environmental tobacco smoke exposure and ischaemic heart disease an evaluation of the evidence.
      but its association with RA has never been investigated. Thus, we sought to further explore the relationship between cigarette smoking, the role of intensity, the duration, and the effect of smoking cessation, as well as that of exposure to passive smoke, and the development of RA in women aged 30 to 81 years in a large prospective cohort study.

      Methods

      Study Population

      The Nurses’ Health Study is a prospective cohort of 121,700 female nurses aged 30 to 55 years in 1976, when the study began. Information was collected from the subjects by biennial questionnaires regarding diseases, lifestyle, and health practices. Follow-up of the original cohort has been greater than 94% through 2002.
      • Colditz G.A.
      • Manson J.E.
      • Hankinson S.E.
      The Nurses’ Health Study 20-year contribution to the understanding of health among women.
      All aspects of this study were approved by the Partners’ HealthCare Institutional Review Board.

      Identification of Rheumatoid Arthritis

      From 1976 to 1982, participants self-reported a diagnosis of RA or other connective tissue diseases (CTDs) including systemic lupus erythematosus, mixed CTD, scleroderma, polymyositis, dermatomyositis, or Sjögren syndrome in a write-in section of the questionnaire. Beginning in 1982, participants have been specifically asked whether they have a physician diagnosis of RA. For this study, we contacted 11,966 women reporting RA and 1673 women reporting any other CTD on any of the biennial questionnaires from 1976 to 2002. We requested permission to review their medical records and that they complete the CTD screening questionnaire (CSQ), which includes 6 questions on symptoms of RA.
      • Karlson E.W.
      • Sanchez-Guerrero J.
      • Wright E.A.
      • et al.
      A connective tissue disease screening questionnaire for population studies.
      The CSQ was scored as positive if 4 of 6 responses to these questions were positive. It was scored as “possible” if 3 of 6 were positive, in which case follow-up CSQs were mailed to the participant in the following cycles. In total, after 5 mailings, 10,455 (77%) of the women who had self-reported RA or any other CTD responded. After excluding subjects who denied the diagnosis of RA (n = 2239), had RA diagnosed before 1976 (n = 475), denied permission for record review (n = 1066), or had a negative CSQ for RA symptoms (n = 3052), we requested medical records from 3623 women and obtained 2737 (76%) records with adequate information. Two rheumatologists trained in chart abstraction independently conducted a medical record review examining the charts for the American College of Rheumatology diagnostic criteria for RA.
      • Arnett F.C.
      • Edworthy S.M.
      • Bloch D.A.
      • et al.
      The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis.
      Subjects with 4 of the 7 diagnostic criteria documented in the medical record were considered to have definite RA. We confirmed 807 new cases of definite RA diagnosed between 1976 and 2002 for a case confirmation rate of 29% of the medical records reviewed and 7% of the original self-reports.

      Population for Analysis

      For all analyses, we excluded prevalent cases of RA diagnosed before June 1976, nonresponders, and women who reported any CTD that was not subsequently confirmed to be RA by medical record review. We excluded women who reported any cancer (except nonmelanoma skin cancer) at baseline or during follow-up because cancer and its treatment can affect smoking behavior. Women were censored when they failed to respond to any subsequent biennial questionnaire. When women failed to report smoking status during a questionnaire cycle, we did not include these person-years in the analysis. Thus, the final group studied included 103,818 women followed from 1976 to 2002 and 680 cases of incident RA among those women. The overall prevalence of RA cases diagnosed after 1976 (when the women were aged 35-50 years) was 0.8%, which is close to the expected 1% to 2% prevalence rate of RA expected in a mainly white adult female population.
      • Gabriel S.E.
      • Crowson C.S.
      • O’Fallon W.M.
      The epidemiology of rheumatoid arthritis in Rochester, Minnesota, 1955-1985.
      • Kellgren J.H.
      Epidemiology of rheumatoid arthritis.

      Information on Smoking Exposures and Potential Confounding Variables

      All exposure information was self-reported on the mailed questionnaires administered every 2 years since 1976. On the initial Nurses’ Health Study questionnaire, participants reported whether they were current smokers or had ever smoked in the past and the age at which they began to smoke. Current smokers were asked for the number of cigarettes typically smoked per day, and former smokers reported the age at which they stopped smoking and the number of cigarettes smoked per day before quitting. On each subsequent questionnaire, participants reported whether they currently smoked and the number of cigarettes smoked per day. From these reports, we calculated smoking duration, pack-years of smoking (product of years of smoking and packs of cigarettes per day), time since quitting, and age at quitting. No data from participants who did not report smoking information were included for the specific questionnaire cycles missing this information. However, when smoking information was later provided it was updated correctly to account for the missing years. All smoking variables were time-varying. The information was updated every 2 years, because smokers often stop and restart smoking. Questions concerning passive cigarette smoke exposure were asked once in 1982. Participants were asked whether neither of their parents, their mother only, their father only, or both parents had smoked at home. Subjects also were asked to report the number of years they had lived with a smoker (including as a child and as an adult) and whether they were never, occasionally, or regularly exposed to cigarette smoke at work.
      Important reproductive covariates were chosen on the basis of our past findings of associations between reproductive factors and the risk of developing RA in this cohort.
      • Karlson E.W.
      • Mandl L.A.
      • Hankinson S.E.
      • Grodstein F.
      Do breast-feeding and other reproductive factors influence future risk of rheumatoid arthritis? Results from the Nurses’ Health Study.
      Age at menarche, regularity of menses between the ages 20 and 35 years, parity, duration of breastfeeding, postmenopausal hormone use, and body mass index were included as potential confounders of the smoking and risk of RA relationship. Alcohol intake and socioeconomic status were included as covariates in the multivariate models. Alcohol intake was reported every 2-year cycle starting in 1980 and classified in grams per day. Father’s occupation, asked in 1976, was chosen as a proxy for childhood socioeconomic status. There are no data on household income in the Nurses’ Health Study, and because all the women are nurses, the range of adult socioeconomic status is limited. Father’s occupation was classified as professional (including professional and managerial jobs) or nonprofessional (including clerical, sales, craftsman, service, laborer, farmer, and at-home jobs).

      Statistical Analysis

      Person-years of follow-up accrued from the date of return of the 1976 baseline questionnaire until the date of diagnosis of RA, as defined in the medical record; death; the report of cancer (excluding nonmelanoma skin cancer); or loss to follow-up, defined as no further return of questionnaires. When women failed to report smoking status during a questionnaire cycle, we did not include these person years in the analysis. Age-adjusted and multivariate Cox proportional hazards models were used to study the association between RA (developing from ages 30-81 years) and cigarette smoking. Information from each 2-year questionnaire was used to analyze the risk of RA in the next 2-year cycle. Age was categorized as less than 50, 50-54, 55-59, 60-64, and 65 or more years in age-adjusted models and in months in multivariate models. Tests for linear trend were calculated excluding the reference category of never smokers and using continuous values for smoking exposure. In multivariate models that adjusted for multiple smoking variables, never smokers were excluded. The correlation between the time since quitting cigarettes and smoking duration was measured using a Spearman correlation test.
      Stratified analyses were used to examine the effect of age at which smoking ceased and the subsequent risk of RA. Stratified analyses also were used to investigate the association of smoking with RF seropositive and RF seronegative RA separately, to examine the effect of living with a smoker separately among never and ever smokers, and to examine the risks in premenopausal compared with postmenopausal women. In addition, to analyze the effect of excluding the women who had reported cancer, a subanalysis in which they were included was performed. SAS version 6 was used for all analyses.
      SAS

      Results

      We examined characteristics of the women participating in the Nurses’ Health Study according to smoking status in 1990, the approximate midpoint of the follow-up period in this analysis (Table 1). Body mass index was lower among current smokers than it was among never and past smokers. Current smokers were less likely to be premenopausal, and among the parous women, they were the least likely to have breastfed for more than 12 months during all pregnancies. In addition, there were no important differences in the characteristics (listed in Table 1) of those women who responded to our additional mailings (n = 10,455), compared with those who did not (n = 3184).
      Table 1Age-standardized Characteristics of the Nurses’ Health Study Women in 1990 Within Categories of Smoking Status (n = 88,520)
      Never n = 38,677Past n = 33,076Current n = 15,425
      Age, mean56.356.655.9
      Body mass index, mean25.925.924.7
      Alcohol intake, g/d, mean3.64.04.3
      Age at menarche years, mean12.412.412.5
      Pack-years of cigarettes smoked, mean017.238.3
      Average number of cigarettes per day016.418.9
      Duration of smoking, y020.736.1
      Father in professional/managerial occupation (%)242726
      Usually/very irregular menses at age 20-35 y (%)121212
      Nulliparous (%)667
      Breastfeeding > 12 mo (among parous women) (%)191511
      Premenopausal (%)252420
      Postmenopausal hormone use (among postmenopausal women) (%)333632
      Parent smoked at home
      Assessed in 1982.
      (%)
      505957
      Lived > 30 years with a smoker
      Assessed in 1982.
      (%)
      71118
      Regularly exposed to smoke at work
      Assessed in 1982.
      (%)
      202132
      low asterisk Assessed in 1982.
      Of 680 RA cases with a mean age at diagnosis of 56 (±9) years, 401 (59%) were seropositive and 207 (30%) had radiographic changes characteristic of RA. Most of the RA cases (83%) had been diagnosed by a member of the American College of Rheumatology.
      The age-adjusted relative risk (RR) of developing RA was 1.49 (95% confidence interval [CI] 1.28-1.75) among women who had ever smoked, and the risk was similarly elevated in both current and past smokers (Table 2). Adjusting for the multiple potentially important covariates included in the multivariate models did not substantially alter risk estimates (multivariate RR of ever smoking 1.46 [95% CI 1.24-1.71]). In age-adjusted and multivariate models, higher pack-years were associated with increasing risk of RA with a significant dose-response trend (P trend <.01). The highest risk was observed in past smokers of 40 pack-years or more (multivariate RR 2.29 [95% CI 1.62-3.24]).
      Table 2Relative Risk of Rheumatoid Arthritis by Pack-years of Smoking Among Women in the Nurses’ Health Study, 1976-2002 (N = 103,818)
      CasesPerson-yearsAge-adjusted RR (95% CI
      Ninety-five percent confidence interval.
      )
      Multivariate RR
      Adjusted for body mass index (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none, <5, 5 to <9, 9 to ≤15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12 y), regularity of menses (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      (95% CI
      Ninety-five percent confidence interval.
      )
      Smoking status
       Never2371,020,5871.01.0
       Ever4431,270,1721.49 (1.28-1.75)1.46 (1.24-1.71)
       Current163491,8011.46 (1.20-1.79)1.43 (1.16-1.75)
       Past280778,3711.50 (1.26-1.79)1.47 (1.23-1.76)
      Pack-years
      Never2371,020,5871.01.0
      1-10103423,6691.08 (0.86-1.36)1.07 (0.85-1.36)
      11-2079255,9801.37 (1.06-1.77)1.35 (1.04-1.74)
      21-3090198,9981.97 (1.55-2.52)1.93 (1.50-2.47)
      31-4056155,1451.45 (1.08-1.95)1.36 (1.01-1.83)
      >4098196,8601.99 (1.57-2.53)1.86 (1.46-2.38)
      P for trend
      P for trend excludes never smokers.3
      P<.001P<.01
      Pack-years current smokers
       Never2371,020,5871.01.0
       1-2030144,7971.04 (0.70-1.55)1.04 (0.70-1.53)
       21-4072196,3501.62 (1.24-2.11)1.54 (1.18-2.02)
       >4059133,0111.76 (1.32-2.35)1.65 (1.23-2.21)
      P for trend
      P for trend excludes never smokers.3
      P=.01P=.01
      Pack-years past smokers
       Never2371,020,5871.01.0
       1-1094365,7721.11 (0.88-1.41)1.11 (0.87-1.42)
       11-2058169,0801.44 (1.08-1.92)1.40 (1.05-1.87)
       21-4074157,7931.90 (1.46-2.47)1.79 (1.37-2.33)
       ≥303963,8482.46 (1.74-3.48)2.29 (1.62-3.24)
      P for trend
      P for trend excludes never smokers.3
      P<.001P<.001
      low asterisk Ninety-five percent confidence interval.
      Adjusted for body mass index (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none, <5, 5 to <9, 9 to ≤15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12 y), regularity of menses (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      P for trend excludes never smokers.
      • Hazes J.M.
      • Dijkmans B.A.
      • Vandenbroucke J.P.
      • de Vries R.R.
      • Cats A.
      Lifestyle and the risk of rheumatoid arthritis cigarette smoking and alcohol consumption.
      To determine whether the intensity or the duration of cigarette smoking contributed more to the risk of RA, we examined these 2 components comprising pack-years separately (Table 3). In age-adjusted models, the risk of RA was elevated among current and past smokers of 15 cigarettes or more per day, with a significant linear trend for increasing risk with increasing amount smoked (P trend <.001 in both models). The risk of developing RA trended upward with increasing duration of smoking among past smokers (P trend <.001) and current smokers (P trend .04). The results of multivariate-adjusted models were similar to the age-adjusted models (Table 3). In the multivariate models limited to ever smokers in which smoking amount and duration were mutually adjusted for one another, both factors contribute independently to the RR of RA (Table 3). The increased RR of RA was demonstrated up to 20 years after cessation of cigarette smoking in both age-adjusted and multivariate models limited to ever smokers (Table 3). Smoking duration and time since quitting were highly negatively correlated (Spearman correlation −0.62).
      Table 3Relative Risk of Developing Rheumatoid Arthritis by Smoking Amount and Duration Among Women in the Nurses’ Health Study, 1976-2002 (n = 103,818)
      Current SmokersPast Smokers
      CasesPerson-yearsAge-adjusted RR
      Cox proportional hazards models, adjusted for age in months
      (95% CI
      Ninety-five percent confidence interval.
      )
      Multivariate RR
      Multivariate model adjusted for BMI (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none, <5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12), menstrual regularity (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      (95% CI
      Ninety-five percent confidence interval.
      )
      Multivariate RR
      Includes smokers only. Adjusted for all of above. Amount smoked was adjusted for smoking duration (continuous), and smoking duration was adjusted for smoking intensity (continuous). Years since quitting model was also adjusted for smoking intensity (continuous).
      (95% CI
      Ninety-five percent confidence interval.
      )
      CasesPerson-yearsAge-adjusted RR
      Cox proportional hazards models, adjusted for age in months
      (95% CI
      Ninety-five percent confidence interval.
      )
      Multivariate RR
      Multivariate model adjusted for BMI (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none, <5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12), menstrual regularity (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      (95% CI
      Ninety-five percent confidence interval.
      )
      Multivariate RR
      Includes smokers only. Adjusted for all of above. Amount smoked was adjusted for smoking duration (continuous), and smoking duration was adjusted for smoking intensity (continuous). Years since quitting model was also adjusted for smoking intensity (continuous).
      (95% CI
      Ninety-five percent confidence interval.
      )
      Smoking intensity
      Amount smoked is the last amount reported for past smokers and the current amount for current smokers.
      (cigarettes/day)
       Never2371,020,5871.01.02371,020,5871.01.0
       1-1442154,6171.17 (0.84-1.63)1.15 (0.82-1.60)1.0108374,8441.21 (0.96-1.52)1.18 (0.94-1.49)1.0
       15-2468198,6171.50 (1.14-1.96)1.48 (1.12-1.95)1.28 (0.87-1.89)107260,6931.74 (1.38-2.19)1.70 (1.34-2.14)1.39 (1.06-1.82)
       ≥2552123,4531.85 (1.37-2.51)1.80 (1.32-2.44)1.56 (1.03-2.36)57123,8051.92 (1.44-2.57)1.83 (1.36-2.45)1.44 (1.04-2.01)
      P for trend
      P for trend excludes never smokers.
      P<.001P<.001P = .28P<.001P<.001P<.01
      Smoking duration (y)
       Never2371,020,5871.01.0
       <2528140,6641.29 (0.85-1.97)1.27 (0.83-1.94)1.0
       25 to <302686,7271.41 (0.92-2.15)1.37 (0.90-1.94)1.11 (0.62-1.99)
       30 to <353394,3241.20 (0.82-1.75)1.17 (0.80-1.71)0.96 (0.54-1.70)
       35 to <403978,2821.72 (1.21-2.45)1.70 (1.19-2.43)1.39 (0.79-2.44)
       ≥403587,5841.55 (1.07-2.24)1.60 (1.10-2.32)1.23 (0.69-2.20)
      P for trend
      P for trend excludes never smokers.
      P = .04P = .03P = .05
       Never2371,020,5871.01.0
       <1047210,8071.02 (0.75-1.40)1.02 (0.75-1.40)1.0
       10 to <2078233,1811.45 (1.2-1.88)1.43 (1.10-1.85)1.29 (0.90-1.88)
       20 to <3069172,8511.59 (1.21-2.08)1.51 (1.15-1.98)1.35 (0.92-1.98)
       ≥3076146,8552.01 (1.54-2.63)1.95 (1.48-2.55)1.76 (1.19-2.59)
      P for trend
      P for trend excludes never smokers.
      P<.001P<.001P<.01
      Years since quitting
       Never2371,020,5871.01.0
       ≥2081273,9661.16 (0.90-1.50)1.14 (0.88-1.48)1.0
       10 to <2091233,6561.70 (1.33-2.16)1.65 (1.29-2.10)1.40 (1.02-1.91)
       0 to <10104265,2651.71 (1.35-2.15)1.65 (1.30-2.08)1.48 (1.09-2.00)
       Current163491,8011.45 (1.1-1.77)1.44 (1.17-1.77)1.25 (0.94-2.00)
      P for trend
      P for trend excludes never smokers.
      P<.01P = .14P = .14
      low asterisk Cox proportional hazards models, adjusted for age in months
      Ninety-five percent confidence interval.
      Multivariate model adjusted for BMI (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none, <5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12), menstrual regularity (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      § Includes smokers only. Adjusted for all of above. Amount smoked was adjusted for smoking duration (continuous), and smoking duration was adjusted for smoking intensity (continuous). Years since quitting model was also adjusted for smoking intensity (continuous).
      Amount smoked is the last amount reported for past smokers and the current amount for current smokers.
      P for trend excludes never smokers.
      To rule out an effect of the onset of RA symptoms on smoking habits, we performed 2 sensitivity analyses: ending all follow-up data 2 years before the date of RA diagnosis; and ending all follow-up data at the first onset of RA symptoms (date of RA symptoms from medical record review available for 639/680 cases). In these sensitivity analyses, the results were essentially unchanged. When follow-up was stopped 2 years before the diagnosis of RA, the fully adjusted RR of RA was 1.47 (95% CI 1.19-1.82) among current smokers and 1.50 (95% CI 1.25-1.80) among past smokers. When follow-up ended at the date of first symptoms of RA, the fully adjusted RR of RA was 1.37 (95% CI 1.10-1.71) among current smokers and 1.43 (95% CI 1.19-1.73) among past smokers.
      In an analysis stratifying the cases of RA by the presence or absence of RF, the RRs of seropositive RA associated with past and current smoking and with increasing pack-years were greater than those observed for all cases of RA and similar strong trends were observed (Table 4). Strong positive associations also were observed for intensity and duration of smoking (data not shown). In comparison, the RRs of seronegative RA associated with cigarette smoking were not as elevated for past and current smokers and no clear dose response was observed for pack-years (Table 4).
      Table 4Relative Risk of Developing Rheumatoid Factor Positive or Negative Rheumatoid Arthritis by Smoking Status and Pack-years of Smoking Among Women in the Nurses’ Health Study, 1976-2002 (N = 103,818)
      RF positive RARF negative RA
      CasesPerson-yearsMultivariate RR
      Adjusted for BMI (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none,<5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12), menstrual regularity (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      (95% CI
      Ninety-five percent confidence interval.
      )
      CasesPerson-yearsMultivariate RR
      Adjusted for BMI (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none,<5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12), menstrual regularity (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      (95% CI
      Ninety-five percent confidence interval.
      )
      Smoking status
       Never1361,019,1081.01011,018,7511.0
       Ever2661,267,7611.59 (1.29-1.97)1771,266,6501.28 (1.00-1.65)
       Current100490,8171.58 (1.21-2.06)63490,0781.23 (0.88-1.70)
       Past166776,9441.60 (1.27-2.02)114776,5721.31 (1.00-1.73)
      Pack-years
       Never1361,019,1081.01011,018,7511.0
       1-1056423,0261.05 (0.76-1.43)47422,9001.10 (0.78-1.57)
       11-2050255,4821.57 (1.13-2.18)29255,1991.08 (0.71-1.64)
       21-3055198,5072.18 (1.59-2.01)35198,2441.61 (1.09-2.38)
       31-4035154,8741.55 (1.06-2.25)21154,5751.12 (0.69-1.80)
       >4063196,5092.22 (1.63-3.02)35196,3471.43 (0.96-2.12)
      P for trend
      P for trend excludes never smokers.
      P<.01P = .55
      low asterisk Adjusted for BMI (<22, 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none,<5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12, >12), menstrual regularity (very regular, usually regular, usually irregular, very irregular), duration of breastfeeding (nulliparous, parous-never, 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal, postmenopausal-never, postmenopausal-past, postmenopausal-current).
      Ninety-five percent confidence interval.
      P for trend excludes never smokers.
      To investigate the effect of menopausal status on the risk of developing RA associated with cigarette smoking, Cox proportional hazards models stratified by menopausal status were performed. The risk of developing RA during the premenopausal years for all women in the cohort for ever smokers (multivariate RR 1.48 [95% CI 1.07-2.06]) was not substantially different than the risk among postmenopausal women ever smokers (multivariate RR 1.53 [95% CI 1.25-1.88]). In a sensitivity analysis including the women with cancer who had been excluded (65 cases of verified RA had been excluded), the risk estimates were very similar (multivariate RR among ever smokers 1.47 [95% CI 1.26-1.72]).
      Analyses examining passive cigarette smoke exposure included 453 incident RA cases diagnosed since 1982, the year in which this question was included on the questionnaire. Living with a parent smoker as a child and being exposed to smoke regularly at work were not associated with significantly increased risks of RA. Smokers were almost twice as likely to have lived with a smoker than nonsmokers, reducing the power to examine the risk of passive exposure to cigarette smoke in nonsmokers. Among nonsmokers, there was a suggestion that living with a smoker for greater than 30 years was associated with an increased risk of RA, although not significantly so (multivariate RR 1.59 [95% CI 0.92-2.74]), and no dose effect was seen (Table 5).
      Table 5Relative Risk of Rheumatoid Arthritis According to Number of Years Lived with Smoker Among Women in the Nurses’ Health Study, 1982-2002 (n = 79,283)
      Years lived with smokerEver SmokersNever Smokers
      CasesPerson-yearsAge-adjusted RR (95% CI
      Ninety-five percent confidence interval.
      )
      Multivariate RR
      Adjusted for BMI (<22 [ref], 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none [ref], <5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12 [ref], >12), parity (nulliparous [ref], parous, missing), duration of breastfeeding (nulliparous, parous-never [ref], 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal [ref], postmenopausal-never, postmenopausal-past, postmenopausal-current), and pack-years of smoking, continuous (among ever smokers).
      (95% CI
      Ninety-five percent confidence interval.
      )
      CasesPerson-yearsAge-adjusted RR (95% CI
      Ninety-five percent confidence interval.
      )
      Multivariate RR
      Adjusted for BMI (<22 [ref], 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none [ref], <5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12 [ref], >12), parity (nulliparous [ref], parous, missing), duration of breastfeeding (nulliparous, parous-never [ref], 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal [ref], postmenopausal-never, postmenopausal-past, postmenopausal-current), and pack-years of smoking, continuous (among ever smokers).
      (95% CI
      Ninety-five percent confidence interval.
      )
      Never73277,5301.01.035130,6641.01.0
      1-928107,3900.99 (0.64-1.53)1.04 (0.67-1.63)54134,9641.50 (0.98-2.29)1.53 (0.98-2.40)
      10-192097,6750.79 (0.48-1.29)0.75 (0.46-1.25)62175,8921.35 (0.89-2.05)1.38 (0.90-2.12)
      20-291683,0790.70 (0.41-1.21)0.67 (0.39-1.17)79180,7231.46 (0.98-2.16)1.44 (0.95-2.20)
      >302051,5401.38 (0.80-2.38)1.59 (0.92-2.74)57127,4561.37 (0.89-2.12)1.46 (0.92-2.32)
      P for trend
      P for trend excludes those who never lived with smoker.
      .38.77.59.27
      low asterisk Ninety-five percent confidence interval.
      Adjusted for BMI (<22 [ref], 22 to <25, 25 to <30, 30 to <35, >35 kg/m2), alcohol intake (none [ref], <5, 5 to <9, 9 to <15, >15 g/d), father’s occupation (professional, nonprofessional), age at menarche (<12, 12 [ref], >12), parity (nulliparous [ref], parous, missing), duration of breastfeeding (nulliparous, parous-never [ref], 1-3 mo, 4-11 mo, 12-23 mo, >24 mo), postmenopausal hormone use (premenopausal [ref], postmenopausal-never, postmenopausal-past, postmenopausal-current), and pack-years of smoking, continuous (among ever smokers).
      P for trend excludes those who never lived with smoker.
      To quantify the influence of cigarette smoking exposure on the incidence of RA in this cohort of women, population attributable risk estimates were computed. Two hundred new cases of RA per 100,000 in the population could have been prevented during the 26 years of follow-up of the 103,818 women included in this study. Given the genotypic mixture of this particular cohort, this represents 25% of the new cases of RA (population attributable risk percentage) that may have been prevented had none of these women smoked.

      Discussion

      In this large cohort of female nurses, both past and current cigarette smoking were associated with similarly elevated risks of RA of more than 40%. A strong dose response between pack-years of smoking and the risk of RA was demonstrated. Both smoking intensity and smoking duration contributed to this effect, and smoking greater than 15 cigarettes a day and greater than 20 years were both associated with increased risk. The risk of incident RA remained substantially elevated until 20 years after smoking cessation. This suggests that a threshold quantity of cigarette smoke exposure may exist in those genetically at risk for RA and that, once attained, the risk remains elevated for years to come. In this study, the first to investigate the relationship between passive smoke exposure and RA, a possible association between long-term passive smoke exposure and elevated risk of RA was suggested, although this question should be addressed in further studies. Our assessment of the independent effect of passive smoke exposure, however, was limited by the small proportion of nonsmokers who had had passive smoke exposure and by the fact that questions concerning passive smoke exposure were asked only once in 1982.
      We also observed that the risk associated with smoking was much higher for seropositive RA than for seronegative RA, as has been observed by others.
      • Heliovaara M.
      • Aho K.
      • Aromaa A.
      • Knekt P.
      • Reunanen A.
      Smoking and risk of rheumatoid arthritis.
      • Symmons D.P.
      • Bankhead C.R.
      • Harrison B.J.
      • et al.
      Blood transfusion, smoking, and obesity as risk factors for the development of rheumatoid arthritis results from a primary care-based incident case-control study in Norfolk, England.
      • Krishnan E.
      • Sokka T.
      • Hannonen P.
      Smoking-gender interaction and risk for rheumatoid arthritis.
      • Stolt P.
      • Bengtsson C.
      • Nordmark B.
      • et al.
      Quantification of the influence of cigarette smoking on rheumatoid arthritis results from a population based case-control study, using incident cases.
      • Padyukov L.
      • Silva C.
      • Stolt P.
      • Alfredsson L.
      • Klareskog L.
      A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis.
      The relationship between cigarette smoke exposure and RF, antibodies directed against the Fc portion of immunoglobulin G, is interesting and complex. Cigarette smoking has been linked to the presence of RF in the blood of healthy individuals without RA,
      • Korpilahde T.
      • Heliovaara M.
      • Knekt P.
      • Marniemi J.
      • Aromaa A.
      • Aho K.
      Smoking history and serum cotinine and thiocyanate concentrations as determinants of rheumatoid factor in non-rheumatoid subjects.
      • Tuomi T.
      • Heliovaara M.
      • Palosuo T.
      • Aho K.
      Smoking, lung function, and rheumatoid factors.
      • Houssien D.A.
      • Scott D.L.
      • Jonsson T.
      Smoking, rheumatoid factors, and rheumatoid arthritis.
      and it is not clear whether this represents a pre-RA state. Although seropositive RA is a relatively homogeneous phenotype, more likely to be associated with the HLA-DR4 shared epitope and joint erosions, destruction, and extra-articular disease,
      • Scott D.L.
      Prognostic factors in early rheumatoid arthritis.
      seronegative RA is more challenging to diagnose and, in studies such as these, may represent a more heterogeneous collection of diseases. Viral, psoriatic, reactive, and crystal-associated arthritides may all mimic RA, and time is often necessary for diagnosis confirmation. The finding that cigarette smoking is associated with a greater elevation of the risk of seropositive than seronegative RA underscores the likelihood that seronegative RA is a heterogenous and distinct condition.
      Some of the case-control studies in the past have reported the risk of RA is elevated in male, but not in female, smokers,
      • Heliovaara M.
      • Aho K.
      • Aromaa A.
      • Knekt P.
      • Reunanen A.
      Smoking and risk of rheumatoid arthritis.
      • Uhlig T.
      • Hagen K.B.
      • Kvien T.K.
      Current tobacco smoking, formal education, and the risk of rheumatoid arthritis.
      • Krishnan E.
      • Sokka T.
      • Hannonen P.
      Smoking-gender interaction and risk for rheumatoid arthritis.
      but our prospective cohort study demonstrates that cigarette smoking elevates the risk of RA in women as well. In these past studies, smoking was categorized as never/past/current or never/ever, and associations between the development of RA and the intensity and/or duration of cigarette smoking were not investigated.
      • Hazes J.M.
      • Dijkmans B.A.
      • Vandenbroucke J.P.
      • de Vries R.R.
      • Cats A.
      Lifestyle and the risk of rheumatoid arthritis cigarette smoking and alcohol consumption.
      • Heliovaara M.
      • Aho K.
      • Aromaa A.
      • Knekt P.
      • Reunanen A.
      Smoking and risk of rheumatoid arthritis.
      • Uhlig T.
      • Hagen K.B.
      • Kvien T.K.
      Current tobacco smoking, formal education, and the risk of rheumatoid arthritis.
      • Krishnan E.
      • Sokka T.
      • Hannonen P.
      Smoking-gender interaction and risk for rheumatoid arthritis.
      Given that the most substantial risk of RA is posed by the heaviest smoking and that women, until recently, have been lighter smokers than men,
      • Patel J.D.
      • Bach P.B.
      • Kris M.G.
      Lung cancer in US women a contemporary epidemic.
      • Giovino G.A.
      Epidemiology of tobacco use in the United States.
      the association between smoking intensity and RA in women may have been overlooked.
      When Hernandez Avila and colleagues
      • Hernandez Avila M.
      • Liang M.H.
      • Willett W.C.
      • et al.
      Reproductive factors, smoking, and the risk for rheumatoid arthritis.
      investigated cigarette smoking, along with several other exposures, in the Nurses’ Health Study from 1976 to 1984, they reported suggestive but nonsignificant findings: age-adjusted RR for RA in current smokers was 1.3 (95% CI 0.9-2.1) and 1.5 among former smokers (95% CI 0.9-2.3).
      • Hernandez Avila M.
      • Liang M.H.
      • Willett W.C.
      • et al.
      Reproductive factors, smoking, and the risk for rheumatoid arthritis.
      Only in the subgroup of current smokers of 15 to 24 cigarettes per day did they find a significant association with the risk of RA and “undifferentiated polyarthritis.” They had far fewer RA cases at that time, and the inclusion of this latter group of nonspecific cases may have biased their results toward the null.
      The Iowa Women’s Health Study
      • Criswell L.A.
      • Merlino L.A.
      • Cerhan J.R.
      • et al.
      Cigarette smoking and the risk of rheumatoid arthritis among postmenopausal women results from the Iowa Women’s Health Study.
      and a Swedish case-control study using incident RA cases
      • Stolt P.
      • Bengtsson C.
      • Nordmark B.
      • et al.
      Quantification of the influence of cigarette smoking on rheumatoid arthritis results from a population based case-control study, using incident cases.
      reported that the risk of developing RA in those who had quit smoking more than 10 years earlier was not significantly elevated above that seen in those who had never smoked. In the Nurses’ Health Study, however, we did not observe the risk of RA to decline below that of current smokers until 20 or more years after smoking cessation. It seems that a certain threshold amount of cigarette smoke exposure is necessary and that once attained, this exposure poses a latent, rather than an instantaneous, hazard and that the risk does not decrease with time until long after cessation of exposure.
      The biochemical and cellular pathways by which cigarette smoke augments the risk of RA have not been delineated. Cigarette smoke contains hundreds of potentially toxic components, including tars, resins, and free radicals, and is known to activate endogenous sources of free radicals, as well as neutrophils, monocytes, and platelets.
      • van der Vaart H.
      • Postma D.S.
      • Timens W.
      • ten Hacken N.H.
      Acute effects of cigarette smoke on inflammation and oxidative stress a review.
      • Seagrave J.
      • Barr E.B.
      • March T.H.
      • Nikula K.J.
      Effects of cigarette smoke exposure and cessation on inflammatory cells and matrix metalloproteinase activity in mice.
      • Garey K.W.
      • Neuhauser M.M.
      • Robbins R.A.
      • Danziger L.H.
      • Rubinstein I.
      Markers of inflammation in exhaled breath condensate of young healthy smokers.
      • Churg A.
      • Zay K.
      • Shay S.
      • et al.
      Acute cigarette smoke-induced connective tissue breakdown requires both neutrophils and macrophage metalloelastase in mice.
      • Burghuber O.C.
      • Punzengruber C.
      • Sinzinger H.
      • Haber P.
      • Silberbauer K.
      Platelet sensitivity to prostacyclin in smokers and non-smokers.
      Abnormalities in T-cell function,
      • Hughes D.A.
      • Haslam P.L.
      • Townsend P.J.
      • Turner-Warwick M.
      Numerical and functional alterations in circulatory lymphocytes in cigarette smokers.
      • Robbins C.S.
      • Dawe D.E.
      • Goncharova S.I.
      • et al.
      Cigarette smoke decreases pulmonary dendritic cells and impacts antiviral immune responsiveness.
      reduction in natural killer cells,
      • Moszczynski P.
      • Zabinski Z.
      • Moszczynski Jr, P.
      • Rutowski J.
      • Slowinski S.
      • Tabarowski Z.
      Immunological findings in cigarette smokers.
      impairment of humoral immunity,
      • Moszczynski P.
      • Zabinski Z.
      • Moszczynski Jr, P.
      • Rutowski J.
      • Slowinski S.
      • Tabarowski Z.
      Immunological findings in cigarette smokers.
      • Burton R.C.
      Smoking, immunity, and cancer.
      • Hersey P.
      • Prendergast D.
      • Edwards A.
      Effects of cigarette smoking on the immune system. Follow-up studies in normal subjects after cessation of smoking.
      and elevated levels of inflammatory markers such as interleukin-6 and C-reactive protein
      • Bermudez E.A.
      • Rifai N.
      • Buring J.E.
      • Manson J.E.
      • Ridker P.M.
      Relation between markers of systemic vascular inflammation and smoking in women.
      • Tracy R.P.
      • Psaty B.M.
      • Macy E.
      • et al.
      Lifetime smoking exposure affects the association of C-reactive protein with cardiovascular disease risk factors and subclinical disease in healthy elderly subjects.
      have been observed in smokers. Heavy cigarette smoking may act as a trigger in the genetically predisposed to seropositive RA, a likely “second hit” in those who are at already elevated risk. In a Swedish study, cigarette smoking dramatically increased the RR of seropositive RA (but not seronegative RA) to 7.5 (95% CI 4.2-13.1) in those who carried the HLA-DR4 “shared epitope,” the strongest known genetic risk factor for RA.
      • Padyukov L.
      • Silva C.
      • Stolt P.
      • Alfredsson L.
      • Klareskog L.
      A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis.
      This study does have some important limitations. As an observational study, it cannot uncover the molecular or biologic pathways linking exposure to cigarette smoke and the development of RA but can only point to interesting associations that should lead to further research. Only 7% of self-reported incident cases of RA were confirmed in this cohort. We have very stringent case-validation procedures that may have led to the exclusion of some RA cases but ensured high specificity of case classification. All unconfirmed RA and other CTD cases were excluded from the analyses. Given that all prevalent cases diagnosed before the start of the cohort were excluded, the 0.8% prevalence rate for new RA cases diagnosed after 1976 when women were aged 30 to 55 years is in line with the expected 1% to 2% prevalence rate of RA expected in a mainly white adult female population.
      • Gabriel S.E.
      • Crowson C.S.
      • O’Fallon W.M.
      The epidemiology of rheumatoid arthritis in Rochester, Minnesota, 1955-1985.
      • Kellgren J.H.
      Epidemiology of rheumatoid arthritis.
      We have also analyzed the characteristics of nonresponders to our supplemental questionnaire (CSQ) and request for medical records, and they do not seem to be different from those of the responders. The risk of RA was equally elevated in past and former smokers and, in our lagged analyses censoring follow-up at the onset of symptoms and at 2 years before RA diagnosis, the results were unchanged, suggesting no effect of early RA symptoms on smoking behaviors.
      The presence of RF was not assessed in a standardized fashion. Rather, RF tests were performed in many different laboratories and reported in medical records that were then reviewed for the presence of ACR criteria. The development of RF seropositivity may have been missed if only checked once in early RA, and thus a few cases of seropositive RA may have been misclassified as RF negative. In addition, we excluded subjects at the first report of any cancer other than skin cancer, believing that the development of a malignancy was likely to influence smoking behavior. In doing so, however, we may have been excluding those subjects who were the heaviest smokers, potentially at the highest risk of developing RA, biasing our findings. In our sensitivity analysis, including women who had previously developed cancer, we found that the overall risk of RA was unchanged.
      As a prospective study with exposure information collected before the development of incident RA and smoking information updated every 2 years, our study has the advantage of eliminating the biases that can occur in case-control studies, and reemphasizes that heavy cigarette smoking is an important risk factor for RA in women and that this risk is proportional to both the daily amount and number of years of smoking. The risk of RA remains elevated long after cessation of smoking, and adjustments for age, socioeconomic status, body mass index, and known reproductive risk factors have only minor influences on the risk of RA, suggesting that cigarette smoking acts through an independent pathway. Unfortunately, the prevalence of smoking among young women in this country has grown in recent years.
      One quarter of the 680 new cases of RA diagnosed after the age of 35 years in this cohort could have been prevented if none of these women had ever smoked. Our findings add to the long list of known health hazards posed by cigarette smoking and to the reasons that young women should be dissuaded from starting and encouraged to quit smoking cigarettes.

      References

        • Gabriel S.E.
        • Crowson C.S.
        • O’Fallon W.M.
        The epidemiology of rheumatoid arthritis in Rochester, Minnesota, 1955-1985.
        Arthritis Rheum. 1999; 42: 415-420
        • Vessey M.P.
        • Villard-Mackintosh L.
        • Yeates D.
        Oral contraceptives, cigarette smoking and other factors in relation to arthritis.
        Contraception. 1987; 35: 457-464
        • Hazes J.M.
        • Dijkmans B.A.
        • Vandenbroucke J.P.
        • de Vries R.R.
        • Cats A.
        Lifestyle and the risk of rheumatoid arthritis.
        Ann Rheum Dis. 1990; 49: 980-982
        • Heliovaara M.
        • Aho K.
        • Aromaa A.
        • Knekt P.
        • Reunanen A.
        Smoking and risk of rheumatoid arthritis.
        J Rheumatol. 1993; 20: 1830-1835
        • Voigt L.F.
        • Koepsell T.D.
        • Nelson J.L.
        • Dugowson C.E.
        • Daling J.R.
        Smoking, obesity, alcohol consumption, and the risk of rheumatoid arthritis.
        Epidemiology. 1994; 5: 525-532
        • Symmons D.P.
        • Bankhead C.R.
        • Harrison B.J.
        • et al.
        Blood transfusion, smoking, and obesity as risk factors for the development of rheumatoid arthritis.
        Arthritis Rheum. 1997; 40: 1955-1961
        • Uhlig T.
        • Hagen K.B.
        • Kvien T.K.
        Current tobacco smoking, formal education, and the risk of rheumatoid arthritis.
        J Rheumatol. 1999; 26: 47-54
        • Krishnan E.
        • Sokka T.
        • Hannonen P.
        Smoking-gender interaction and risk for rheumatoid arthritis.
        Arthritis Res Ther. 2003; 5: R158-R162
        • Stolt P.
        • Bengtsson C.
        • Nordmark B.
        • et al.
        Quantification of the influence of cigarette smoking on rheumatoid arthritis.
        Ann Rheum Dis. 2003; 62: 835-841
        • Padyukov L.
        • Silva C.
        • Stolt P.
        • Alfredsson L.
        • Klareskog L.
        A gene-environment interaction between smoking and shared epitope genes in HLA-DR provides a high risk of seropositive rheumatoid arthritis.
        Arthritis Rheum. 2004; 50: 3085-3092
        • Hernandez Avila M.
        • Liang M.H.
        • Willett W.C.
        • et al.
        Reproductive factors, smoking, and the risk for rheumatoid arthritis.
        Epidemiology. 1990; 1: 285-291
        • Karlson E.W.
        • Lee I.M.
        • Cook N.R.
        • Manson J.E.
        • Buring J.E.
        • Hennekens C.H.
        A retrospective cohort study of cigarette smoking and risk of rheumatoid arthritis in female health professionals.
        Arthritis Rheum. 1999; 42: 910-917
        • Criswell L.A.
        • Merlino L.A.
        • Cerhan J.R.
        • et al.
        Cigarette smoking and the risk of rheumatoid arthritis among postmenopausal women.
        Am J Med. 2002; 112: 465-471
        • Eisner M.D.
        Environmental tobacco smoke and adult asthma.
        Clin Chest Med. 2002; 23: 749-761
        • Brennan P.
        • Buffler P.A.
        • Reynolds P.
        • et al.
        Secondhand smoke exposure in adulthood and risk of lung cancer among never smokers.
        Int J Cancer. 2004; 109: 125-131
        • Tredaniel J.
        • Boffetta P.
        • Saracci R.
        • Hirsch A.
        Environmental tobacco smoke and the risk of cancer in adults.
        Eur J Cancer. 1993; 29A: 2058-2068
        • Glantz S.A.
        • Parmley W.W.
        Passive smoking and heart disease. Epidemiology, physiology, and biochemistry.
        Circulation. 1991; 83: 1-12
        • Law M.R.
        • Morris J.K.
        • Wald N.J.
        Environmental tobacco smoke exposure and ischaemic heart disease.
        BMJ. 1997; 315: 973-980
        • Colditz G.A.
        • Manson J.E.
        • Hankinson S.E.
        The Nurses’ Health Study.
        J Womens Health. 1997; 6: 49-62
        • Karlson E.W.
        • Sanchez-Guerrero J.
        • Wright E.A.
        • et al.
        A connective tissue disease screening questionnaire for population studies.
        Ann Epidemiol. 1995; 5: 297-302
        • Arnett F.C.
        • Edworthy S.M.
        • Bloch D.A.
        • et al.
        The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis.
        Arthritis Rheum. 1988; 31: 315-324
        • Kellgren J.H.
        Epidemiology of rheumatoid arthritis.
        Arthritis Rheum. 1966; 9: 658-674
        • Karlson E.W.
        • Mandl L.A.
        • Hankinson S.E.
        • Grodstein F.
        Do breast-feeding and other reproductive factors influence future risk of rheumatoid arthritis? Results from the Nurses’ Health Study.
        Arthritis Rheum. 2004; 50: 3458-3467
        • SAS
        SAS Institute, Inc, Cary, NC1990
        • Korpilahde T.
        • Heliovaara M.
        • Knekt P.
        • Marniemi J.
        • Aromaa A.
        • Aho K.
        Smoking history and serum cotinine and thiocyanate concentrations as determinants of rheumatoid factor in non-rheumatoid subjects.
        Rheumatology (Oxford). 2004; 43: 1424-1428
        • Tuomi T.
        • Heliovaara M.
        • Palosuo T.
        • Aho K.
        Smoking, lung function, and rheumatoid factors.
        Ann Rheum Dis. 1990; 49: 753-756
        • Houssien D.A.
        • Scott D.L.
        • Jonsson T.
        Smoking, rheumatoid factors, and rheumatoid arthritis.
        Ann Rheum Dis. 1998; 57: 175-176
        • Scott D.L.
        Prognostic factors in early rheumatoid arthritis.
        Rheumatology (Oxford). 2000; 39: 24-29
        • Patel J.D.
        • Bach P.B.
        • Kris M.G.
        Lung cancer in US women.
        JAMA. 2004; 291: 1763-1768
        • Giovino G.A.
        Epidemiology of tobacco use in the United States.
        Oncogene. 2002; 21: 7326-7340
      1. Women and Smoking. US Dept of Health and Human Services, Public Health Service, Office of the Surgeon General, Washington, DC2001
        • van der Vaart H.
        • Postma D.S.
        • Timens W.
        • ten Hacken N.H.
        Acute effects of cigarette smoke on inflammation and oxidative stress.
        Thorax. 2004; 59: 713-721
        • Seagrave J.
        • Barr E.B.
        • March T.H.
        • Nikula K.J.
        Effects of cigarette smoke exposure and cessation on inflammatory cells and matrix metalloproteinase activity in mice.
        Exp Lung Res. 2004; 30: 1-15
        • Garey K.W.
        • Neuhauser M.M.
        • Robbins R.A.
        • Danziger L.H.
        • Rubinstein I.
        Markers of inflammation in exhaled breath condensate of young healthy smokers.
        Chest. 2004; 125: 22-26
        • Churg A.
        • Zay K.
        • Shay S.
        • et al.
        Acute cigarette smoke-induced connective tissue breakdown requires both neutrophils and macrophage metalloelastase in mice.
        Am J Respir Cell Mol Biol. 2002; 27: 368-374
        • Burghuber O.C.
        • Punzengruber C.
        • Sinzinger H.
        • Haber P.
        • Silberbauer K.
        Platelet sensitivity to prostacyclin in smokers and non-smokers.
        Chest. 1986; 90: 34-38
        • Hughes D.A.
        • Haslam P.L.
        • Townsend P.J.
        • Turner-Warwick M.
        Numerical and functional alterations in circulatory lymphocytes in cigarette smokers.
        Clin Exp Immunol. 1985; 61: 459-466
        • Robbins C.S.
        • Dawe D.E.
        • Goncharova S.I.
        • et al.
        Cigarette smoke decreases pulmonary dendritic cells and impacts antiviral immune responsiveness.
        Am J Respir Cell Mol Biol. 2004; 30: 202-211
        • Moszczynski P.
        • Zabinski Z.
        • Moszczynski Jr, P.
        • Rutowski J.
        • Slowinski S.
        • Tabarowski Z.
        Immunological findings in cigarette smokers.
        Toxicol Lett. 2001; 118: 121-127
        • Burton R.C.
        Smoking, immunity, and cancer.
        Med J Aust. 1983; 2: 411-412
        • Hersey P.
        • Prendergast D.
        • Edwards A.
        Effects of cigarette smoking on the immune system. Follow-up studies in normal subjects after cessation of smoking.
        Med J Aust. 1983; 2: 425-429
        • Bermudez E.A.
        • Rifai N.
        • Buring J.E.
        • Manson J.E.
        • Ridker P.M.
        Relation between markers of systemic vascular inflammation and smoking in women.
        Am J Cardiol. 2002; 89: 1117-1119
        • Tracy R.P.
        • Psaty B.M.
        • Macy E.
        • et al.
        Lifetime smoking exposure affects the association of C-reactive protein with cardiovascular disease risk factors and subclinical disease in healthy elderly subjects.
        Arterioscler Thromb Vasc Biol. 1997; 17: 2167-2176