Interactions Between Genome‐Wide Genetic Factors and Smoking Influencing Risk of Systemic Lupus Erythematosus. Issue 11 (23rd September 2020)

• Record Type: Journal Article
• Authors: Cui, Jing; Raychaudhuri, Soumya; Karlson, Elizabeth W.; Speyer, Cameron; Malspeis, Susan; Guan, Hongshu; Sparks, Jeffrey A.; Ni, Hongru; Liu, Xinyi; Stevens, Emma; Williams, Jessica N.; Davenport, Emma E.; Knevel, Rachel; Costenbader, Karen H.
• Is Part Of: Arthritis & rheumatology. Volume 72:Issue 11(2020)
• Page Start: 1863
• Page End: 1871
• Publication Date: 2020-09-23
• Subjects: Arthritis -- Periodicals; Rheumatism -- Periodicals; 616.72
• DOI: 10.1002/art.41414
• Languages: English
• ISSNs: 2326-5191

Abstract

Abstract : Objective: To identify interactions between genetic factors and current or recent smoking in relation to risk of developing systemic lupus erythematosus (SLE). Methods: For the study, 673 patients with SLE (diagnosed according to the American College of Rheumatology 1997 updated classification criteria) were matched by age, sex, and race (first 3 genetic principal components) to 3, 272 control subjects without a history of connective tissue disease. Smoking status was classified as current smoking/having recently quit smoking within 4 years before diagnosis (or matched index date for controls) versus distant past/never smoking. In total, 86 single‐nucleotide polymorphisms and 10 classic HLA alleles previously associated with SLE were included in a weighted genetic risk score (wGRS), with scores dichotomized as either low or high based on the median value in control subjects (low wGRS being defined as less than or equal to the control median; high wGRS being defined as greater than the control median). Conditional logistic regression models were used to estimate both the risk of SLE and risk of anti–double‐stranded DNA autoantibody–positive (dsDNA+) SLE. Additive interactions were assessed using the attributable proportion (AP) due to interaction, and multiplicative interactions were assessed using a chi‐square test (with 1 degree of freedom) for the wGRS and for individual risk alleles. Separate repeated analyses were carried out among subjects of European ancestry only. Results: The mean ± SD age of the SLE patients at the time of diagnosis was 36.4 ± 15.3 years. Among the 673 SLE patients included, 92.3% were female and 59.3% were dsDNA+. Ethnic distributions were as follows: 75.6% of European ancestry, 4.5% of Asian ancestry, 11.7% of African ancestry, and 8.2% classified as other ancestry. A high wGRS (odds ratio [OR] 2.0, P = 1.0 × 10 −51 versus low wGRS) and a status of current/recent smoking (OR 1.5, P = 0.0003 versus distant past/never smoking) were strongly associated with SLE risk, with significant additive interaction (AP 0.33, P = 0.0012), and associations with the risk of anti‐dsDNA+ SLE were even stronger. No significant multiplicative interactions with the total wGRS ( P = 0.58) or with the HLA ‐only wGRS ( P = 0.06) were found. Findings were similar in analyses restricted to only subjects of European ancestry. Conclusion: The strong additive interaction between an updated SLE genetic risk score and current/recent smoking suggests that smoking may influence specific genes in the pathogenesis of SLE.