Background/Purpose:  Studies have shown associations between environmental factors and rheumatoid arthritis (RA) risk. Also, genome-wide association studies (GWAS) have identified genetic markers associated with RA risk. We examined additive and multiplicative gene-environment interactions between 44 RA genetic markers and smoking, parity/breast-feeding, body mass index (BMI), and BMI at age 18 in determining RA risk.

Methods:  We used a pooled nested case-control sample of 541 RA cases, including 303 seropositive (CCP+ and/or RF+) cases, and 549 matched controls from the Nurses’ Health Study and Nurses’ Health Study II. Controls were matched on age, menopausal status, and postmenopausal hormone use. All participants were Caucasian. We genotyped 36 single nucleotide polymorphisms identified in previous GWAS and meta-analysis, and 8 HLA-DBR1 shared epitope (SE) genotypes at 4 digit resolution. The information on cumulative pack-years of smoking, parity and months of breast-feeding prior and BMI variables determined prior to RA diagnosis, and BMI at age 18 was extracted from questionnaires. We examined the main effects of the genetic and environmental factors. We tested for additive interactions using the attributable proportion (AP) due to interaction and tested for multiplicative interactions using a 1 d.f. test. We used unconditional logistic regression models, adjusted for matching factors. Bonferroni correction was used to adjust for multiple comparisons. We further stratified the cases into seropositive RA and seronegative RA phenotypes and tested for interactions separately.

Results:  We found significant main effects of HLA-DBR1*0401 (OR=1.77, 95% CI=1.32-2.38), any HLA-DBR1 SE (OR=1.80, 95% CI=1.41-2.29), and smoking (OR=1.55, 95% CI=1.21-1.99) on RA risk. For the interaction analyses, combined HLA-DRB1 SE and HLA-DBR1*0401 showed significant interaction with smoking in all RA and seropositive RA. Also, PTPN22 showed significant interaction with smoking in all RA (Table). The effect of interaction between BMI prior to RA diagnosis and PXK on RA risk was significant in all RA (AP=0.74, p<0.005) and seropositive RA (AP=0.80, p<0.005). We also found significant interaction effects on RA risk between BMI at age 18 and HLA-DBR1*0401 in all RA (AP=0.66, p=0.013) and seropositive RA (AP=0.67, p=0.031). However, we did not find any significant interaction between genetic factors and parity/breast-feeding. Also, there was no significant interaction found in the seronegative RA phenotype. Several multiplicative interactions were significant, but none remained significant after multiple comparisons correction.

Conclusion:  We found significant gene-environment interaction effects on RA risk with stronger associations for the seropositive RA phenotype. We also found HLA-DBR1*0401 had different AP when interacting with different environmental factors, which suggests different mechanisms for these interactions.