You are here

Article Text

Article menu
Clinical and epidemiological research
Concise report
Recent infections are associated with decreased risk of rheumatoid arthritis: a population-based case-control study
  1. Maria E C Sandberg1,2,
  2. Camilla Bengtsson1,
  3. Lars Klareskog3,
  4. Lars Alfredsson1,
  5. Saedis Saevarsdottir1,3
  1. 1Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden
  2. 2Epidemiology and Register Centre South, Skåne University Hospital, Lund, Sweden
  3. 3Rheumatology Unit, Department of Medicine, Karolinska Institutet/Karolinska University Hospital, Stockholm, Sweden
  1. Correspondence to Dr Maria E C Sandberg, Institute of Environmental Medicine, Box 210, Stockholm 171 71, Sweden; maria.sandberg{at}ki.se

Abstract

Objectives Do recent infections affect the risk of rheumatoid arthritis (RA)?

Methods We used the population-based case-control study EIRA (N=6401) on incident RA and healthy controls, matched for sex, age, calendar period and area of residence. Gastroenteritis, urinary tract infection, genital infection, prostatitis, sinusitis, tonsillitis and pneumonia during the 2 years before inclusion in the study were investigated. Conditional logistic regression was used to calculate OR, adjusting for smoking and socioeconomic status.

Results Infections in the gastrointestinal and urogenital tract before clinical onset were associated with a lowered risk of RA: gastroenteritis (OR=0.71 (95% CI 0.63 to 0.80)), urinary tract infections (OR=0.78 (95% CI 0.68 to 0.90)) and genital infections (OR=0.80 (95% CI 0.64 to 1.00)), while a non-significant association of similar magnitude was observed for the less common prostatitis (OR=0.64 (95% CI 0.38 to 1.08)). In contrast, no associations were observed for sinusitis, tonsillitis or pneumonia.

Conclusions Gastrointestinal and urogenital infections, but not respiratory infections, are associated with a significantly lowered risk of RA. The results indicate that infections in general do not affect the risk for RA, but that certain infections, hypothetically associated with changes in the gut microbiome, could diminish the risk.

  • Epidemiology
  • Rheumatoid Arthritis
  • Infections

https://doi.org/10.1136/annrheumdis-2014-206493

Statistics from Altmetric.com

    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.

    Introduction

    Infections have for a long time been proposed to be potential triggers of rheumatoid arthritis (RA), preceding the clinical onset of disease, but despite decades of studies there is very limited epidemiological evidence to support this hypothesis, apart from the well-studied association with Porphyromonas gingivalis through periodontitis.1 Attempts to confirm observations from mechanistic and animal studies indicating that specific infections play a role in the pathogenesis of RA have mostly focused on examining serological evidence for past infection.

    More recently, it has been recognised that changes in the composition of the microbiome in the gut and other mucosal surfaces may have a major impact on immune homeostatis and inflammatory diseases2–4 in both humans and rodents, and might thus influence the risk for RA.

    The aim of the present study has been to investigate whether infections in different sites influence the risk of RA making use of one of the largest population-based case-control studies of RA, the EIRA study, which has power to also separately investigate the two major disease subsets, that is, the subsets that are characterised by the presence or absence of anti-citrullinated protein antibodies (ACPAs).

    Methods

    Study population

    During 1996–2009, 7701 people in the south and middle parts of Sweden were invited to the EIRA study (3062 cases and 4639 controls); of these, 6401 chose to participate, which corresponds to 83%, 92% of the cases and 77% of the controls. The cases were newly diagnosed with RA according to the 1987 ACR criteria,5 and the controls were all free from RA and randomly selected from the population using incidence density sampling. Cases and controls were matched on age at diagnosis (±2.5 years), calendar year, sex and area of residence. The EIRA study has been described in further detail elsewhere.6–8

    Definition of exposure

    The exposure was self-reported in a questionnaire in which the timing of the infection (the year before inclusion in the study vs 1–2 years before) was also included for a proportion of the study population (N=4004 (63%), collected 1996–2005). The exact (translated) question was as follows; ‘Have you, during the last two years had any of the following diseases? 1. Stomach infection with diarrhoea, 2. Urinary tract infection (eg, cystitis or pyelonephritis), 3. Genital infection, 4. Inflammation in the prostate (prostatitis), 5. Sinusitis treated with antibiotics, 6. Tonsillitis or other throat infection treated with antibiotics’.

    Statistical analysis

    Conditional logistic regression was used to estimate the OR of RA associated with the different infections by complete-case analysis. Through the matched design of the study, the estimated ORs were adjusted for age at diagnosis, calendar period, sex and area of residence. We also, in a separate model, additionally adjusted for smoking and educational level, which have been shown to be a good marker for socioeconomic status in this population.9 Effect modification by timing of infection, smoking (ever vs never smokers) and shared epitope (no shared epitope genes vs any shared epitope genes) was investigated by including interaction terms in the model. To be able to investigate the potential effect modification by sex (since the study was matched on sex), we stratified our dataset and performed separated analysis for the effect of infections on women and on men. Finally, we investigated whether the effect of infections was different for ACPA-positive and ACPA-negative RA, by restricting the sample to cases with ACPA-positive and ACPA-negative RA and their controls, respectively.

    All analyses were carried out using SAS Statistical Package V.9.3. All participants have given informed consent and the Ethical Review Board at Karolinska Institutet, Stockholm, Sweden, approved the study.

    Results

    The final dataset consisted of 6401 study participants (2831 cases and 3570 controls), the proportion of discordant strata ranged between 42% (gastroenteritis) and 8% (pneumonia), the mean age at inclusion was 52 years old, and 72% were women. The median time of symptom duration for the cases when filling in the questionnaire was 204 days. For further details on the participants, see table 1.

    View this table:
    Table 1

    Basic characteristics of the included cases and controls

    Gastroenteritis, urinary tract and genital infections during the previous 2 years were each associated with a statistically significantly decreased risk of RA, albeit the estimate for genital infections was only borderline significant (OR=0.71 (95% CI 0.63 to 0.80), OR=0.78 (95% CI 0.68 to 0.90), and OR=0.80 (95% CI 0.64 to 1.00), respectively) (table 2). For patients who reported both gastrointestinal, urinary tract and genital infections during the last 2 years, the OR was 0.50 (95% CI 0.32 to 0.79). Further; the risk estimate for prostatitis was of similar magnitude although non-significant (OR=0.64 95% CI 0.38 to 1.08). For upper respiratory tract infections, sinusitis and tonsillitis, and for pneumonia, no association was observed. Overall, adjustments for smoking and socioeconomic status affected the estimates very little.

    View this table:
    Table 2

    OR for rheumatoid arthritis associated with each of the infections investigated together with 95% CI

    We found no evidence of different effect of infections less than a year ago, compared with infection 1–2 years ago, nor by sex, shared epitope status or smoking. Further, we found indications of a stronger effect of infections on ACPA-positive RA than ACPA-negative RA, though only the risk estimates for sinusitis were statistically significantly different (table 3). The risk increase that can be seen for sinusitis exposure on the risk for ACPA-negative RA is only borderline statistically significant and could potentially be due to chance. Finally, we investigated the potential effect modification of infections by shared epitope status among the ACPA-positive subset but found none.

    View this table:
    Table 3

    OR for ACPA-positive and ACPA-negative rheumatoid arthritis (RA) associated with each of the investigated infections together with 95% CI

    Discussion

    In this first study on site-specific infections prior to RA diagnosis, based on a large population-based case-control study on incident RA, a significantly decreased risk of RA was observed for study participants who reported gastroenteritis, urinary tract or genital infections during the 2 years prior to RA diagnosis. In striking contrast, no overall effect was seen for respiratory infections: sinusitis, tonsillitis or pneumonia. This could be particularly interesting in light of emerging data implicating that the microbiome in the gut may play a role in RA pathogenesis since mucosal sites are exposed to a high load of bacterial antigens and may thus represent the site of initiation or modification of inflammation in RA.2 ,3 ,10

    Strengths of this study include its population-based setting and that the RA cases were incident. Another important strength is the possibility of investigating the timing of infections, and since this analysis showed no different effect by timing there is no indication of reversed causality (ie, no reason to believe that the investigated infections in fact happened after and as a consequence of the diagnosis of RA). We defined a 2-year limit of infection history, both for the participants’ ability to remember these relatively common infections and due to the underlying hypothesis for including these questions: infections as triggers for the clinical onset of disease. However, we do recognise that the triggering of autoimmunity to citrullinated and other proteins often happens much longer than 2 years before disease onset.11 Further, a limitation of this study is the lack objective confirmation of the infections. Unfortunately, this information is not possible to obtain even in the Swedish landscape of registers since these infections rarely lead to contact with the healthcare system. This fact has further implications for the potential hypothesis that antibiotics used to treat the infections would have a preventive effect on RA. The questionnaire specifies antibiotics use only in the case of tonsillitis and sinusitis; presumably most pneumonia cases were also treated, but these are the infections without any effect on the risk of RA in our material. For the other infections, we unfortunately have no information on antibiotics use and can therefore not investigate this. However, it would seem that the only mechanism consistent with the differential effect depending on site of infection would be a mechanism linked to a specific type of antibiotics given for infections at these particular sites. Finally, reactive, ACPA-negative, arthritis has a direct link to infections and has symptoms that may, at least initially, be misinterpreted as RA, which could bias our results; reassuringly, however, we find that our results are similar or somewhat stronger in the ACPA-positive patients compared with the ACPA-negative and also that the timing of infection does not affect the results.

    The overwhelming majority of the previous research regarding RA and infections have investigated the potentially increased risk for serious infections during treatment with biological agents, for example, antitumour necrosis factor treatment. However; in 2005 a small case-control study reported that RA patients were significantly less affected by common infections prior to diagnosis compared with study participants free from RA. Our present results are thus in accordance with these earlier results, though, unfortunately, the site and timing of infections were not specified in this earlier report.12

    We see several possible biological mechanisms underlying our results: a decreased prevalence of infections may indicate a strong capacity for immune activation in general at mucosal surfaces, which could, in some not yet understood way, be linked to protection against RA. A second, and partly related, potential explanation would be that infections in the gut and urinary tract may change the composition of the microbiome and that such a change would influence the susceptibility to RA. Hypotheses on connections between inflammation and microbial composition in the gut and risk for RA have been raised and discussed for many years13 ,14 and recently reports from basic science suggest that changes in the gut microbiome may play a role in the pathogenesis of RA.2 ,3 ,10 In this context, we note with interest that the sites of infections that conferred a decreased risk in this study are primarily infected with gram-negative bacteria, while the sites in which infections did not confer a decreased risk are primarily infected with gram-positive bacteria. Lending some support to this hypothesis is that both sulfa (sulfasalazine) and tetracycline drugs, used for gram-negative infections at these particular sites, have in double-blind randomised clinical trials published before the era of ‘biologic’ treatments in RA been shown to be effective against RA.15–17

    In summary, we view these results, showing a decreased risk of RA for people with recent infections in the gastrointestinal and urogenital tract as hypothesis generating and look forward to future studies on the possible relationships between infections and microbiomes in different sites and risk for RA and other inflammatory diseases.

    Acknowledgments

    We acknowledge the EIRA study group and EIRA data collectors.

    References

    1. ↵
      1. Lundberg K,
      2. Kinloch A,
      3. Fisher BA, et al
      . Antibodies to citrullinated alpha-enolase peptide 1 are specific for rheumatoid arthritis and cross-react with bacterial enolase. Arthritis Rheum 2008;58:3009–19. doi:10.1002/art.23936
      OpenUrlCrossRefPubMedWeb of Science
    2. ↵
      1. Bedaiwi MK,
      2. Inman RD
      . Microbiome and probiotics: link to arthritis. Curr Opin Rheumatol 2014;26:410–15. doi:10.1097/BOR.0000000000000075
      OpenUrlCrossRefPubMed
    3. ↵
      1. Brusca SB,
      2. Abramson SB,
      3. Scher JU
      . Microbiome and mucosal inflammation as extra-articular triggers for rheumatoid arthritis and autoimmunity. Curr Opin Rheumatol 2014;26:101–7. doi:10.1097/BOR.0000000000000008
      OpenUrlCrossRefPubMed
    4. ↵
      1. Gomez A,
      2. Luckey D,
      3. Yeoman CJ, et al
      . Loss of sex and age driven differences in the gut microbiome characterize arthritis-susceptible 0401 mice but not arthritis-resistant 0402 mice. PLoS ONE 2012;7:e36095. doi:10.1371/journal.pone.0036095
      OpenUrlCrossRefPubMed
    5. ↵
      1. Arnett FC,
      2. Edworthy SM,
      3. Bloch DA, et al
      . The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis. Arthritis Rheum 1988;31:315–24. doi:10.1002/art.1780310302
      OpenUrlCrossRefPubMedWeb of Science
    6. ↵
      1. Bengtsson C,
      2. Berglund A,
      3. Serra ML, et al
      . Non-participation in EIRA: a population-based case-control study of rheumatoid arthritis. Scand J Rheumatol 2010;39:344–6. doi:10.3109/03009740903501634
      OpenUrlCrossRefPubMed
    7. ↵
      1. Klareskog L,
      2. Stolt P,
      3. Lundberg K, et al
      . A new model for an etiology of rheumatoid arthritis: smoking may trigger HLA-DR (shared epitope)-restricted immune reactions to autoantigens modified by citrullination. Arthritis Rheum 2006;54:38–46. doi:10.1002/art.21575
      OpenUrlCrossRefPubMedWeb of Science
    8. ↵
      1. Stolt P,
      2. Bengtsson C,
      3. 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. Ann Rheum Dis 2003;62: 835–41. doi:10.1136/ard.62.9.835
      OpenUrlAbstract/FREE Full Text
    9. ↵
      1. Bengtsson C,
      2. Nordmark B,
      3. Klareskog L, et al
      . Socioeconomic status and the risk of developing rheumatoid arthritis: results from the Swedish EIRA study. Ann Rheum Dis 2005;64:1588–94. doi:10.1136/ard.2004.031666
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Scher JU,
      2. Abramson SB
      . The microbiome and rheumatoid arthritis. Nat Rev Rheumatol 2011;7:569–78. doi:10.1038/nrrheum.2011.121
      OpenUrlCrossRefPubMed
    11. ↵
      1. Rantapaa-Dahlqvist S,
      2. de Jong BA,
      3. Berglin E, et al
      . Antibodies against cyclic citrullinated peptide and IgA rheumatoid factor predict the development of rheumatoid arthritis. Arthritis Rheum 2003;48:2741–9. doi:10.1002/art.11223
      OpenUrlCrossRefPubMedWeb of Science
    12. ↵
      1. Laskin MD,
      2. Pope JE,
      3. Krizova A, et al
      . Are infections increased in rheumatoid arthritis (RA) prior to diagnosis? Results of a case control study of RA compared to non-inflammatory musculoskeletal disorders. Scand J Rheumatol 2005;34:74–5. doi:10.1080/03009740510017904
      OpenUrlCrossRefPubMed
    13. ↵
      1. Olhagen B
      . The intestine and rheumatism. Acta Rheumatol Scand 1970;16:177–83. doi:10.3109/rhe1.1970.16.issue-1-4.21
      OpenUrlCrossRefPubMed
    14. ↵
      1. Svartz N
      . The treatment of rheumatic polyarthritis with acid azo compounds. Rheumatism 1948;4:180–5.
      OpenUrlPubMed
    15. ↵
      1. O'Dell JR,
      2. Haire CE,
      3. Palmer W, et al
      . Treatment of early rheumatoid arthritis with minocycline or placebo: results of a randomized, double-blind, placebo-controlled trial. Arthritis Rheum 1997;40:842–8. doi:10.1002/art.1780400510
      OpenUrlCrossRefPubMed
    16. ↵
      1. Ogrendik M
      . Antibiotics for the treatment of rheumatoid arthritis. Int J Gen Med 2013;7:43–7. doi:10.2147/IJGM.S56957
      OpenUrlPubMed
    17. ↵
      1. Stone M,
      2. Fortin PR,
      3. Pacheco-Tena C, et al
      . Should tetracycline treatment be used more extensively for rheumatoid arthritis? Metaanalysis demonstrates clinical benefit with reduction in disease activity. J Rheumatol 2003;30:2112–22.
      OpenUrlAbstract/FREE Full Text

    Supplementary materials

    Footnotes

    • Handling editor Tore K Kvien

    • LA and SS contributed equally.

    • Contributors All authors of this research paper have directly participated in the planning (LK, LA, SS, MECS), analysis (MECS, SS, LA), interpretation (MECS, SS, LA, LK, CB) and writing (MECS, SS, LK, LA, CB) of the study and all authors have read and approved the final submitted version.

    • Funding This study was financially supported by grants from the Swedish Medical Research Council, the Swedish Research Council for Health, Working Life and Welfare, the AFA foundation, Vinnova, King Gustaf V's 80-year foundation, the Swedish Rheumatic Foundation, Swedish Foundation for Strategic Research.

    • Competing interests None.

    • Patient consent Obtained.

    • Ethics approval Ethical Review Board at Karolinska Institutet, Stockholm, Sweden.

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