Effect of Smoking on Spirometry of African American and White Subjects

BackgroundSmoking is the single most important risk factor for COPD, yet there is still disagreement about the differences in the effect of smoking between white and African-American people. We hypothesized that the results of spirometry between smokers of the two races are equivalent if reference equations and lower limits of normal appropriate for each race are used.MethodsWe retrospectively analyzed all spirometry results in smokers over a 1-year period from the G.V. (Sonny) Montgomery VA Medical Center and excluded those that did not meet American Thoracic Society standards, or those from patients with additional medical problems. The remaining patients were classified by race and then matched for age and smoking history; 108 patients in each group were included, which met the power analysis goal of 98. The two groups were similar in age (57.5 years vs 57.0 years), smoking history (46.1 pack-years vs 46.0 pack-years), and body mass index (27.0 kg/m2 vs 28.3 kg/m2) for African Americans and whites, respectively. Data were analyzed using the unpaired t test, and p values were adjusted for multiple comparisons using the Bonferroni factor.ResultsThere were statistically significant differences between African American and white smokers in FVC (3.67 ± 0.07 L vs 4.26 ± 0.08 L, p = 0.001) and FEV1 (2.33 ± 0.07 L vs 2.72 ± 0.08 L, p = 0.002), as expected from the normal populations; however, there were no differences in FVC as percentage of predicted (89.1 ± 1.3% vs 86.7 ± 1.5%, p = 0.71) and FEV1 as percentage of predicted (71.9 ± 2.1% vs 72.2 ± 1.8%, p = 1.00) when the reference equations appropriate for race were used (third National Health and Nutrition Examination Survey). There were also no differences between the number of subject with abnormal FEV1/FVC results (56 African Americans vs 58 whites, p = 1.00) when the appropriate lower limits of normal were used.ConclusionsThere are no differences in spirometry findings between African Americans and whites when abnormality is defined appropriately using reference equations and lower limits of normal for each race. By using either percentage cutoffs for abnormality, or by adjusting for African-American equations only appropriate for whites, we were able to mimic with our data conflicting results in the literature.