Abstract

Gene-gene and gene-environment interactions are key features in the development of rheumatoid arthritis (RA) and other complex diseases. The aim of this study was to use and compare three different definitions of interaction between the two major genetic risk factors of RA—the HLA-DRB1 shared epitope (SE) alleles and the PTPN22 R620W allele—in three large case-control studies: the Swedish Epidemiological Investigation of Rheumatoid Arthritis (EIRA) study, the North American RA Consortium (NARAC) study, and the Dutch Leiden Early Arthritis Clinic study (in total, 1,977 cases and 2,405 controls). The EIRA study was also used to analyze interactions between smoking and the two genes. “Interaction” was defined either as a departure from additivity, as interaction in a multiplicative model, or in terms of linkage disequilibrium—for example, deviation from independence of penetrance of two unlinked loci. Consistent interaction, defined as departure from additivity, between HLA-DRB1 SE alleles and the A allele of PTPN22 R620W was seen in all three studies regarding anti-CCP–positive RA. Testing for multiplicative interactions demonstrated an interaction between the two genes only when the three studies were pooled. The linkage disequilibrium approach indicated a gene-gene interaction in EIRA and NARAC, as well as in the pooled analysis. No interaction was seen between smoking and PTPN22 R620W. A new pattern of interactions is described between the two major known genetic risk factors and the major environmental risk factor concerning the risk of developing anti-CCP–positive RA. The data extend the basis for a pathogenetic hypothesis for RA involving genetic and environmental factors. The study also raises and illustrates principal questions concerning ways to define interactions in complex diseases. Gene-gene and gene-environment interactions are key features in the development of rheumatoid arthritis (RA) and other complex diseases. The aim of this study was to use and compare three different definitions of interaction between the two major genetic risk factors of RA—the HLA-DRB1 shared epitope (SE) alleles and the PTPN22 R620W allele—in three large case-control studies: the Swedish Epidemiological Investigation of Rheumatoid Arthritis (EIRA) study, the North American RA Consortium (NARAC) study, and the Dutch Leiden Early Arthritis Clinic study (in total, 1,977 cases and 2,405 controls). The EIRA study was also used to analyze interactions between smoking and the two genes. “Interaction” was defined either as a departure from additivity, as interaction in a multiplicative model, or in terms of linkage disequilibrium—for example, deviation from independence of penetrance of two unlinked loci. Consistent interaction, defined as departure from additivity, between HLA-DRB1 SE alleles and the A allele of PTPN22 R620W was seen in all three studies regarding anti-CCP–positive RA. Testing for multiplicative interactions demonstrated an interaction between the two genes only when the three studies were pooled. The linkage disequilibrium approach indicated a gene-gene interaction in EIRA and NARAC, as well as in the pooled analysis. No interaction was seen between smoking and PTPN22 R620W. A new pattern of interactions is described between the two major known genetic risk factors and the major environmental risk factor concerning the risk of developing anti-CCP–positive RA. The data extend the basis for a pathogenetic hypothesis for RA involving genetic and environmental factors. The study also raises and illustrates principal questions concerning ways to define interactions in complex diseases. Rheumatoid arthritis (RA [MIM 180300]) is a prototype of an autoimmune disease with complex etiology that is assumed to involve several genetic as well as environmental factors. The clinical hallmark of RA is symmetrical inflammatory arthritis. The disease is more common among women than men. The major risk factors that have so far been reproducibly identified are genetic variations in the major histocompatibility complex, class II, DR beta 1 (HLA-DRB1 [MIM 142857]) and protein tyrosine phosphatase (PTPN22 [MIM 600716])1Gregersen PK Silver J Winchester RJ The shared epitope hypothesis.Arthritis Rheum. 1987; 30: 1205-1213Crossref PubMed Scopus (1983) Google Scholar, 2Huizinga TWJ Amos CI van der Helm-van Mil AHM Chen W van Gaalen FA Jawaheer D Schreuder GMT Wener M Breedvald FC Ahmad N et al.Refining the complex rheumatoid arthritis phenotype based on specificity of the HLA-DRB1 shared epitope for antibodies to citrullinated proteins.Arthritis Rheum. 2005; 52: 3433-3438Crossref PubMed Scopus (438) Google Scholar, 3Plenge RM Padyukov L Remmers EF Purcell S Lee AT Karlson EW Wolfe F Kastner DL Alfredsson L Altshuler D et al.Replication of putative candidate-gene associations with rheumatoid arthritis in >4,000 samples from North America and Sweden: association of susceptibility with PTPN22, CTLA4, and PADI4.Am J Hum Genet. 2005; 77: 1044-1060Abstract Full Text Full Text PDF PubMed Scopus (427) Google Scholar, 4Begovich AB Carlton VE Honigberg LA Schrodi SJ Chokkalingam AP Alexander HC Ardlie KG Huang Q Smith AM Spoerke JM et al.A missense single-nucleotide polymorphism in a gene encoding a protein tyrosine phosphatase (PTPN22) is associated with rheumatoid arthritis.Am J Hum Genet. 2004; 75: 330-337Abstract Full Text Full Text PDF PubMed Scopus (1120) Google Scholar genes and one environmental risk factor, smoking.5Stolt P Bengtsson C Nordmark B Lindblad S Lundberg I Klareskog L Alfredsson L 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: 835-841Crossref PubMed Scopus (460) Google Scholar, 6Padyukov 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: 3085-3092Crossref PubMed Scopus (484) Google Scholar, 7Klareskog L Stolt P Lundberg K Källberg H Bengtsson C Grunewald J Ronnelid J Harris HE Ulfgren AK Rantapaa-Dahlqvist S et al.A new model for an etiology of RA: smoking may trigger HLA-DR (SE)-restricted immune reactions to autoantigens modified by citrullination.Arthritis Rheum. 2006; 54: 38-46Crossref PubMed Scopus (1030) Google Scholar, 8Linn-Rasker SP van der Helm-van Mil AH Van Gaalen FA Kloppenburg M de Vries R le Cessie S Breedveld FC Toes RE Huizinga TW Smoking is a risk factor for anti-CCP antibodies only in rheumatoid arthritis patients that carry HLA-DRB1 shared epitope alleles.Ann Rheum Dis. 2006; 65: 366-371Crossref PubMed Scopus (234) Google Scholar With regard to all three risk factors, the major effects have been seen in one subset of RA, characterized by the presence of antibodies to citrullinated proteins (anti-CCP), but not in the subset of RA in which these antibodies are not detected.2Huizinga TWJ Amos CI van der Helm-van Mil AHM Chen W van Gaalen FA Jawaheer D Schreuder GMT Wener M Breedvald FC Ahmad N et al.Refining the complex rheumatoid arthritis phenotype based on specificity of the HLA-DRB1 shared epitope for antibodies to citrullinated proteins.Arthritis Rheum. 2005; 52: 3433-3438Crossref PubMed Scopus (438) Google Scholar, 3Plenge RM Padyukov L Remmers EF Purcell S Lee AT Karlson EW Wolfe F Kastner DL Alfredsson L Altshuler D et al.Replication of putative candidate-gene associations with rheumatoid arthritis in >4,000 samples from North America and Sweden: association of susceptibility with PTPN22, CTLA4, and PADI4.Am J Hum Genet. 2005; 77: 1044-1060Abstract Full Text Full Text PDF PubMed Scopus (427) Google Scholar, 7Klareskog L Stolt P Lundberg K Källberg H Bengtsson C Grunewald J Ronnelid J Harris HE Ulfgren AK Rantapaa-Dahlqvist S et al.A new model for an etiology of RA: smoking may trigger HLA-DR (SE)-restricted immune reactions to autoantigens modified by citrullination.Arthritis Rheum. 2006; 54: 38-46Crossref PubMed Scopus (1030) Google Scholar, 8Linn-Rasker SP van der Helm-van Mil AH Van Gaalen FA Kloppenburg M de Vries R le Cessie S Breedveld FC Toes RE Huizinga TW Smoking is a risk factor for anti-CCP antibodies only in rheumatoid arthritis patients that carry HLA-DRB1 shared epitope alleles.Ann Rheum Dis. 2006; 65: 366-371Crossref PubMed Scopus (234) Google Scholar Recently, a pronounced gene-environment interaction was identified between smoking and HLA-DRB1 shared epitope (SE) alleles.6Padyukov 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: 3085-3092Crossref PubMed Scopus (484) Google Scholar, 7Klareskog L Stolt P Lundberg K Källberg H Bengtsson C Grunewald J Ronnelid J Harris HE Ulfgren AK Rantapaa-Dahlqvist S et al.A new model for an etiology of RA: smoking may trigger HLA-DR (SE)-restricted immune reactions to autoantigens modified by citrullination.Arthritis Rheum. 2006; 54: 38-46Crossref PubMed Scopus (1030) Google Scholar, 8Linn-Rasker SP van der Helm-van Mil AH Van Gaalen FA Kloppenburg M de Vries R le Cessie S Breedveld FC Toes RE Huizinga TW Smoking is a risk factor for anti-CCP antibodies only in rheumatoid arthritis patients that carry HLA-DRB1 shared epitope alleles.Ann Rheum Dis. 2006; 65: 366-371Crossref PubMed Scopus (234) Google Scholar The demonstration of this gene-environment interaction, together with immunological studies in animal models of arthritis, led us to form a new etiologic hypothesis suggesting that smoking contributes to citrullination and to triggering of anti-citrulline immunity, which is restricted by HLA-DRB1 SE alleles.7Klareskog L Stolt P Lundberg K Källberg H Bengtsson C Grunewald J Ronnelid J Harris HE Ulfgren AK Rantapaa-Dahlqvist S et al.A new model for an etiology of RA: smoking may trigger HLA-DR (SE)-restricted immune reactions to autoantigens modified by citrullination.Arthritis Rheum. 2006; 54: 38-46Crossref PubMed Scopus (1030) Google Scholar, 9Lundberg K Nijenhuis S Vossenaar E Palmblad K van Venrooij WJ Klareskog L Zendman AJ Harris HE Citrullinated proteins have increased immunogenicity and arthritogenicity and their presence in arthritic joints correlates with disease severity.Arthritis Res Ther. 2005; 7: R458-R467Crossref PubMed Google Scholar Since the PTPN22 gene codes for a tyrosine phosphatase, with a potential function in the regulation of T-cell and B-cell activation, it is of obvious interest for the study of the etiology of RA to know how this more recently described risk gene interacts with the classical HLA-DRB1 SE genes, as well as with smoking. Studies of interaction among risk factors in the epidemiological literature have classically been performed using a departure from the additivity model originally described by Rothman, where a term—the attributable proportion due to interaction (AP)—is used to quantify the contribution of interaction to a disease risk, as compared with the contribution of each of the two risk factors added to each other.10Ahlbom A Alfredsson L Interaction: a word with two meanings creates confusion.Eur J Epidemiol. 2005; 20: 563-564Crossref PubMed Scopus (160) Google Scholar, 11Rothman KJ Epidemiology: an introduction. Oxford University Press, New York2002Google Scholar, 12Andersson T Alfredsson L Kallberg H Zdravkovic S Ahlbom A Calculating measures of biological interaction.Eur J Epidemiol. 2005; 20: 575-579Crossref PubMed Scopus (846) Google Scholar This method can also be used to quantify gene-gene interactions for unlinked loci. An alternative common method for quantifying gene-gene interactions is based on the calculation of the two risk factors' product term in a logistic-regression model (multiplicative or statistical interaction). Recently, another method for detecting gene-gene interaction, based on deviation from independence of penetrance in two unlinked loci, was proposed.13Zhao J Jin L Xiong M Test for interaction between two unlinked loci.Am J Hum Genet. 2006; 79: 831-845Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar As reported in this article, we used these three methods to calculate interactions between the HLA-DRB1 SE alleles and PTPN22 risk allele (R620W) in three different major case-control studies of RA: the Swedish Epidemiological Investigation of Rheumatoid Arthritis (EIRA) study, the North American RA Consortium (NARAC) study, and a Dutch case-control study based on the Leiden early arthritis cohort.5Stolt P Bengtsson C Nordmark B Lindblad S Lundberg I Klareskog L Alfredsson L 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: 835-841Crossref PubMed Scopus (460) Google Scholar, 14Bengtsson C Nordmark B Klareskog L Lundberg I Alfredsson L EIRA Study Group Socioeconomic status and the risk of developing rheumatoid arthritis: results from the Swedish EIRA study.Ann Rheum Dis. 2005; 64: 1588-1594Crossref PubMed Scopus (127) Google Scholar, 15Jawaheer D Seldin MF Amos CI Chen WV Shigeta R Monteiro J Kern M Criswell LA Albani S Nelson JL et al.A genomewide screen in multiplex rheumatoid arthritis families suggests genetic overlap with other autoimmune diseases.Am J Hum Genet. 2001; 68: 927-936Abstract Full Text Full Text PDF PubMed Scopus (308) Google Scholar, 16Jawaheer D Lum RF Amos CI Gregersen PK Criswell LA Clustering of disease features within 512 multicase rheumatoid arthritis families.Arthritis Rheum. 2004; 50: 736-741Crossref PubMed Scopus (42) Google Scholar, 17Van Aken J van Bilsen JH Allaart CF Huizinga TW Breedveld FC The Leiden Early Arthritis Clinic.Clin Exp Rheumatol. 2003; 21: S100-S105PubMed Google Scholar We used the largest one of these, which is also the one in which smoking information is the most detailed (the EIRA study5Stolt P Bengtsson C Nordmark B Lindblad S Lundberg I Klareskog L Alfredsson L 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: 835-841Crossref PubMed Scopus (460) Google Scholar, 6Padyukov 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: 3085-3092Crossref PubMed Scopus (484) Google Scholar, 7Klareskog L Stolt P Lundberg K Källberg H Bengtsson C Grunewald J Ronnelid J Harris HE Ulfgren AK Rantapaa-Dahlqvist S et al.A new model for an etiology of RA: smoking may trigger HLA-DR (SE)-restricted immune reactions to autoantigens modified by citrullination.Arthritis Rheum. 2006; 54: 38-46Crossref PubMed Scopus (1030) Google Scholar, 14Bengtsson C Nordmark B Klareskog L Lundberg I Alfredsson L EIRA Study Group Socioeconomic status and the risk of developing rheumatoid arthritis: results from the Swedish EIRA study.Ann Rheum Dis. 2005; 64: 1588-1594Crossref PubMed Scopus (127) Google Scholar), to calculate interaction between the PTPN22 risk allele and smoking, with regard to the risk of anti-CCP–positive RA, using the departure-from-additivity model. A significant interaction between HLA-DRB1 SE alleles and the PTPN22 R620W allele was seen with all methods when the three studies were pooled, but the departure-from-additivity model was able to identify interaction in each one of the studies analyzed separately, whereas the multiplicative model identified this interaction only when the three studies were pooled. No interaction was seen between smoking and PTPN22. This new description of the interaction patterns among the three major known risk factors of RA, which all work in one but not the other major subset of the disease, provides a basis for further etiologic investigations of this disease, with an increased emphasis on T-cell activation and specificity. In addition, the data illustrate the need for consideration and comparison of different methods for calculating interactions in complex diseases. This report is based on data from three different studies: (1) the EIRA, (2) the NARAC, and (3) the Leiden Early Arthritis Clinic (Leiden EAC). In each study, ethical permission was obtained from the relevant ethical committees, and all the participants consented to contribute to the study. In the EIRA study, we recruited cases and controls aged 18–70 years during May 1996 to December 2003 from a geographically defined area in the south and central regions of Sweden. A case was defined as a person in the study base who was given a first-time diagnosis of RA, according to the American College of Rheumatology (ACR) criteria of 1987,18Arnett FC Eworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS et al.The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis.Arthritis Rheum. 1988; 31: 315-324Crossref PubMed Scopus (17997) Google Scholar by rheumatologists at private as well as general health care units in the participating area. Controls were randomly selected from a continuously updated national population register, with consideration given to age, sex, and living area. If the selected control was not traceable, reported having RA, or refused to participate, a new control was selected using the same procedure. More details about the study are described in other reports of the EIRA study.5Stolt P Bengtsson C Nordmark B Lindblad S Lundberg I Klareskog L Alfredsson L 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: 835-841Crossref PubMed Scopus (460) Google Scholar, 6Padyukov 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: 3085-3092Crossref PubMed Scopus (484) Google Scholar, 7Klareskog L Stolt P Lundberg K Källberg H Bengtsson C Grunewald J Ronnelid J Harris HE Ulfgren AK Rantapaa-Dahlqvist S et al.A new model for an etiology of RA: smoking may trigger HLA-DR (SE)-restricted immune reactions to autoantigens modified by citrullination.Arthritis Rheum. 2006; 54: 38-46Crossref PubMed Scopus (1030) Google Scholar, 14Bengtsson C Nordmark B Klareskog L Lundberg I Alfredsson L EIRA Study Group Socioeconomic status and the risk of developing rheumatoid arthritis: results from the Swedish EIRA study.Ann Rheum Dis. 2005; 64: 1588-1594Crossref PubMed Scopus (127) Google Scholar, 19Stolt P Kallberg H Lundberg I Sjogren B Klareskog L Alfredsson L Silica exposure is associated with increased risk of developing rheumatoid arthritis: results from the Swedish EIRA-study.Ann Rheum Dis. 2005; 64: 582-586Crossref PubMed Scopus (138) Google Scholar The second study used in this report, the NARAC, consists of affected sibling pairs who fulfilled the 1987 ACR criteria18Arnett FC Eworthy SM Bloch DA McShane DJ Fries JF Cooper NS Healey LA Kaplan SR Liang MH Luthra HS et al.The American Rheumatism Association 1987 revised criteria for the classification of rheumatoid arthritis.Arthritis Rheum. 1988; 31: 315-324Crossref PubMed Scopus (17997) Google Scholar and matched, unrelated controls recruited from the New York Cancer Project. Details regarding NARAC are provided elsewhere.15Jawaheer D Seldin MF Amos CI Chen WV Shigeta R Monteiro J Kern M Criswell LA Albani S Nelson JL et al.A genomewide screen in multiplex rheumatoid arthritis families suggests genetic overlap with other autoimmune diseases.Am J Hum Genet. 2001; 68: 927-936Abstract Full Text Full Text PDF PubMed Scopus (308) Google Scholar, 16Jawaheer D Lum RF Amos CI Gregersen PK Criswell LA Clustering of disease features within 512 multicase rheumatoid arthritis families.Arthritis Rheum. 2004; 50: 736-741Crossref PubMed Scopus (42) Google Scholar The third study used, the Leiden EAC, consists of patients with confirmed RA with early onset and hospital-based controls with and without deep venous thrombosis. Details on the Leiden EAC and the controls used are given elsewhere.17Van Aken J van Bilsen JH Allaart CF Huizinga TW Breedveld FC The Leiden Early Arthritis Clinic.Clin Exp Rheumatol. 2003; 21: S100-S105PubMed Google Scholar, 20van der Meer FJ Koster T Vandenbroucke JP Briet E The Leiden Thrombophilia Study (LETS).Thromb Haemost. 1997; 78: 631-635PubMed Google Scholar Analysis of anti-CCP antibodies for samples from the EIRA study, as well as for samples from the Leiden EAC, was made with the Immunoscan-RA Mark2 ELISA, as described elsewhere.2Huizinga TWJ Amos CI van der Helm-van Mil AHM Chen W van Gaalen FA Jawaheer D Schreuder GMT Wener M Breedvald FC Ahmad N et al.Refining the complex rheumatoid arthritis phenotype based on specificity of the HLA-DRB1 shared epitope for antibodies to citrullinated proteins.Arthritis Rheum. 2005; 52: 3433-3438Crossref PubMed Scopus (438) Google Scholar, 21Rönnelid J Wick MC Lampa J Lindblad S Nordmark B Klareskog L van Vollenhoven RF Longitudinal analysis of anti-citrullinated antibodies (anti-CP) during 5 years follow-up in early rheumatoid arthritis: anti-CP status is a stable phenotype that predicts worse disease activity and greater radiological progression.Ann Rheum Dis. 2005; 64: 1744-1749Crossref PubMed Scopus (260) Google Scholar Cases with serum antibody levels >25 U/ml were regarded as anti-CCP positive. Rheumatoid factor status for cases was determined using standard procedures. Antibodies toward CCP in NARAC cases were measured using a second-generation commercial anti-CCP ELISA (Inova Diagnostics). Cases with antibody levels >20 U/ml were regarded as anti-CCP positive. Genotyping procedure of the PTPN22 and HLA-DRB1 genes is described in previous articles.2Huizinga TWJ Amos CI van der Helm-van Mil AHM Chen W van Gaalen FA Jawaheer D Schreuder GMT Wener M Breedvald FC Ahmad N et al.Refining the complex rheumatoid arthritis phenotype based on specificity of the HLA-DRB1 shared epitope for antibodies to citrullinated proteins.Arthritis Rheum. 2005; 52: 3433-3438Crossref PubMed Scopus (438) Google Scholar, 4Begovich AB Carlton VE Honigberg LA Schrodi SJ Chokkalingam AP Alexander HC Ardlie KG Huang Q Smith AM Spoerke JM et al.A missense single-nucleotide polymorphism in a gene encoding a protein tyrosine phosphatase (PTPN22) is associated with rheumatoid arthritis.Am J Hum Genet. 2004; 75: 330-337Abstract Full Text Full Text PDF PubMed Scopus (1120) Google Scholar, 7Klareskog L Stolt P Lundberg K Källberg H Bengtsson C Grunewald J Ronnelid J Harris HE Ulfgren AK Rantapaa-Dahlqvist S et al.A new model for an etiology of RA: smoking may trigger HLA-DR (SE)-restricted immune reactions to autoantigens modified by citrullination.Arthritis Rheum. 2006; 54: 38-46Crossref PubMed Scopus (1030) Google Scholar, 8Linn-Rasker SP van der Helm-van Mil AH Van Gaalen FA Kloppenburg M de Vries R le Cessie S Breedveld FC Toes RE Huizinga TW Smoking is a risk factor for anti-CCP antibodies only in rheumatoid arthritis patients that carry HLA-DRB1 shared epitope alleles.Ann Rheum Dis. 2006; 65: 366-371Crossref PubMed Scopus (234) Google Scholar, 9Lundberg K Nijenhuis S Vossenaar E Palmblad K van Venrooij WJ Klareskog L Zendman AJ Harris HE Citrullinated proteins have increased immunogenicity and arthritogenicity and their presence in arthritic joints correlates with disease severity.Arthritis Res Ther. 2005; 7: R458-R467Crossref PubMed Google Scholar, 22Olerup O Zetterquist H HLA-DR typing by PCR amplification with sequence-specific primers (PCR-SSP) in 2 hours: an alternative to serological DR typing in clinical practice including donor-recipient matching in cadaveric transplantation.Tissue Antigens. 1992; 39: 225-235Crossref PubMed Scopus (1704) Google Scholar Allocation to SE was based on local guidelines as described elsewhere, with the HLA-DRB1 alleles DRB1*01, DRB1*04, and DRB1*10 defined as SE alleles. For simplicity of evaluation, we assumed a dominant A allele model for the PTPN22 R620W polymorphism and a dominant SE allele model. Subjects with SE alleles were classified as having single or double SE alleles, for a separate evaluation of a potential gene-dosage effect. Smoking was considered only for the EIRA study. On the basis of self-reported information from the EIRA questionnaire, we divided participants into “ever smokers” and “never smokers.” An ever smoker was defined as a person who had ever smoked cigarettes. A never smoker was defined as a person who had never smoked cigarettes, pipes, cigars, etc. We calculated odds ratios (ORs) and 95% CIs of developing anti-CCP–positive RA and anti-CCP–negative RA associated with each of the factors—PTPN22 R620W alleles, HLA-DRB1 SE alleles, and smoking—by means of logistic-regression models. We estimated their separate effects, as well as the effect of their simultaneous presence. All analyses based on the EIRA study were adjusted for the matching variables: age (10 categories), living area (11 categories), and sex, together with smoking and presence of HLA-DRB1 SE alleles when appropriate. Analyses based on the NARAC study or the Leiden EAC were adjusted for sex and age (using the same categories as in EIRA). Interaction effects between the presence of HLA-DRB1 SE and PTPN22 R620W alleles were evaluated using three different approaches. First, interaction between presence of HLA-DRB1 and PTPN22 R620W alleles was evaluated using departure from additivity of effects as the interaction criteria (biological interaction), as suggested by Rothman.10Ahlbom A Alfredsson L Interaction: a word with two meanings creates confusion.Eur J Epidemiol. 2005; 20: 563-564Crossref PubMed Scopus (160) Google Scholar, 11Rothman KJ Epidemiology: an introduction. Oxford University Press, New York2002Google Scholar To quantify the amount of interaction in terms of departure from additivity of effects, the AP was calculated together with the 95% CI.12Andersson T Alfredsson L Kallberg H Zdravkovic S Ahlbom A Calculating measures of biological interaction.Eur J Epidemiol. 2005; 20: 575-579Crossref PubMed Scopus (846) Google Scholar The AP is the proportion of the incidence among persons exposed to two interacting factors that is attributable to the interaction per se (i.e., reflecting their joint effect beyond the sum of their independent effects). A more detailed description regarding explanation, definition, and programs used to calculate AP with CIs is given elsewhere.10Ahlbom A Alfredsson L Interaction: a word with two meanings creates confusion.Eur J Epidemiol. 2005; 20: 563-564Crossref PubMed Scopus (160) Google Scholar, 12Andersson T Alfredsson L Kallberg H Zdravkovic S Ahlbom A Calculating measures of biological interaction.Eur J Epidemiol. 2005; 20: 575-579Crossref PubMed Scopus (846) Google Scholar, 23Hosmer DW Lemeshow S Confidence interval estimation of interaction.Epidemiology. 1992; 3: 452-456Crossref PubMed Scopus (767) Google Scholar We assessed the AP for each study as well as the pooled estimate when all studies were combined (with adjustment for study). To investigate potential mechanistic pathways further, we also calculated interaction effects between the presence of PTPN22 R620W alleles and smoking and between smoking and the presence of HLA-DRB1 SE alleles, using the departure-from-additivity–based method. Second, interaction between presence of HLA-DRB1 and PTPN22 R620W alleles was evaluated using the cross-product of the risk factors in a logistic-regression model as interaction criteria (multiplicative or statistical interaction).10Ahlbom A Alfredsson L Interaction: a word with two meanings creates confusion.Eur J Epidemiol. 2005; 20: 563-564Crossref PubMed Scopus (160) Google Scholar, 11Rothman KJ Epidemiology: an introduction. Oxford University Press, New York2002Google Scholar We also pooled the three studies, to increase power to detect multiplicative interaction. In the analysis of multiplicative interaction, we adjusted for age, living area (only EIRA), and sex, as well as for study in the pooled analysis. Third, we used a recently developed method to assess gene-gene interaction, which is based on deviation from independence of penetrance of two unlinked loci. This is referred to as the “linkage disequilibrium (LD) statistic.”13Zhao J Jin L Xiong M Test for interaction between two unlinked loci.Am J Hum Genet. 2006; 79: 831-845Abstract Full Text Full Text PDF PubMed Scopus (96) Google Scholar We also calculated χ2 statistics to compare allele distribution in the different studies, for cases and controls separately. We used SAS software for Windows, version 9.1 (SAS Institute), to analyze the data. This report is based on a total of 1,977 cases and 2,405 controls, distributed across the different studies as follows: EIRA, 1,183 cases and 793 controls; NARAC, 430 cases and 731 controls; and Leiden EAC, 364 cases and 881 controls. The EIRA study mainly involved individuals born in Sweden. When those born outside Sweden were also taken into consideration, 97% were of white origin. Of the cases, ∼61% had anti-CCP–positive RA. The corresponding figures for presence of anti-CCP in the NARAC and the Leiden EAC studies were 81% and 56%, respectively. As published elsewhere,2Huizinga TWJ Amos CI van der Helm-van Mil AHM Chen W van Gaalen FA Jawaheer D Schreuder GMT Wener M Breedvald FC Ahmad N et al.Refining the complex rheumatoid arthritis phenotype based on specificity of the HLA-DRB1 shared epitope for antibodies to citrullinated proteins.Arthritis Rheum. 2005; 52: 3433-3438Crossref PubMed Scopus (438) Google Scholar, 4Begovich AB Carlton VE Honigberg LA Schrodi SJ