Background/Purpose: Smoking has been found to be a risk or protective factor in certain autoimmune diseases. Yet, its role in the idiopathic inflammatory myopathies (IIM) has not been fully explored.  Thus, we assessed the role of smoking in adult IIM and selected phenotypes among different racial groups. 

Methods: Patients with probable or definite dermatomyositis (DM, n=186) and polymyositis (PM, n=250) by Bohan and Peter criteria enrolled in NIH studies for whom smoking history could be obtained from their records were studied.  Patients were defined as smokers if they had smoked at least 100 cigarettes in their lifetime.  Racial background was self-reported, and included 307 Caucasians, 96 African Americans (AA), and 33 patients of other races.  The presence of interstitial lung disease (ILD) was based on physician review of chest-x-ray, chest CT, lung biopsy, and/or pulmonary function tests.  Autoantibodies (Abs) were assessed in three composite groups — anti-synthetase autoantibodies (ASA) (directed against Jo-1, PL-12, PL-7, EJ, KS, OJ antigens), myositis-specific autoantibodies (all ASA and those directed against Mi-2, MJ, TIF-1, SRP antigens), and myositis-associated autoantibodies (directed against Ku, MAS, PM-ScL, tRNA, and U1RNP antigens) — as well as each autoantibody individually.  Categorical comparisons were done using chi-squared analysis or Fisher’s exact test.  Adjustments for age, race, and gender were done using logistic regression. 

Results: There was a significant difference in the frequency of ILD across the 3 racial groups with 60% of AA having ILD compared to 75% of Caucasians and 55% of Other races (p=0.002).  In Caucasians and Other races, smoking was positively associated with ILD (OR 1.79, 95% CI 1.07-2.98, p=0.025 and OR 6.53, 95% CI 1.2-35.57, p=0.023 respectively).  In AA, no significant association between ILD and smoking was observed (OR 0.82, 95% CI 0.32-2.02). 

In Caucasian DM/PM patients, smoking was positively associated with any ASA (OR 2.24, 95% CI 1.28-3.92, p=0.004) and anti-Jo-1 Abs (OR 2.4, 95% CI 1.31-4.4, p=0.004).  In Caucasians there was a negative association between smoking and anti-TIF-1 (OR 0.14, 95% CI 0.02-0.93, p=0.018), as well as anti-Mi-2 (p=0.045) and anti-MJ (p=0.025) Abs.  In AA, there was no association between smoking status and ASA as a group (Chi-square=0.7, p=0.391), anti-Jo-1 (Chi-square=0.0006, p=0.980), or anti-Mi-2 Abs (Chi-square=0.0265, p=0.871), and there was a negative association between U1RNP Abs and smoking (Chi-square=3.9, p=0.049).  Although the number of patients in the Other race category was small, there was a positive association between smoking and ASA (p=0.013), as well as with anti-Jo-1 Abs (p=0.034). 

Conclusion: These data suggest that in Caucasians smoking is a risk factor for developing ILD, as well as anti-Jo-1 Abs and ASA, but is protective for anti-TIF-1, -Mi-2, and -MJ Abs.   In contrast, in AA, smoking appears to have no impact on the development of ILD, anti-Jo-1 Abs, or ASA.  Thus, smoking may differentially modulate the clinical and serologic expression of myositis across racial groups.

« Back to 2015 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/tobacco-smoking-in-different-racial-groups-is-differentially-associated-with-the-development-of-myositis-autoantibodies-and-interstitial-lung-disease-in-the-idiopathic-inflammatory-myopathies/