Smoking inhibits the frequency of bronchovascular bundle thickening in sarcoidosis

RATIONALE/OBJECTIVES: Smoking has been associated with decreased incidence and prevalence of sarcoidosis, but few studies have evaluated effects of smoking on clinical parameters of the disease. The objectives were to determine the association of smoking with radiographic patterns and to evaluate the associations of these smoking-related radiographic patterns on airflow obstruction in sarcoidosis. MATERIALS AND METHODS: Clinical data and computed tomography (CT) scans of 124 patients with sarcoidosis were reviewed. CT scans were assessed for lymph nodes, nodules, bronchiectasis, bronchovascular bundle thickening, displaced hilum, fibrosis, ground glass, emphysema, pleural changes, and alveolar opacities. CT patterns were compared between patients with and without a history of smoking. The effect of smoking on the associations between radiographic patterns and airflow obstruction was assessed with multivariable analysis. RESULTS: Smokers had less frequency of bronchovascular bundle thickening than nonsmokers (11/38 subjects [29%] vs 50/86 subjects [58%], P = .003) and more emphysema (7/38 subjects [18%] vs 1/86 subjects [1%], P = .001). Patients who had bronchovascular bundle thickening were less likely to have ever smoked (11/61 subjects [18%] vs 27/63 subjects [43%], P = .003) or be current smokers (4/61 subjects [7%] vs 15/63 subjects [24%], P = .008). Age (P = .003) and bronchovascular bundle thickening (P = .02) were independent predictors of airflow obstruction. There were no differences in smoking history between patients with airflow obstruction versus those without (10/37 subjects [27%] vs 28/87 subjects [32%], P = .63). CONCLUSIONS: In patients with sarcoidosis, smoking is associated with decreased frequency of bronchovascular bundle thickening, an important clinical manifestation of the lung disease. Further, bronchovascular bundle thickening and age are the only independent predictors of airflow obstruction, and smoking does not confound these associations.