The association between pulmonary impairment and all-cause mortality was investigated among white subjects in a follow-up study of a large national cohort. Pulmonary function was measured during the National Health and Nutrition Examination Survey (NHANESI) (1971 to 1975); subsequent mortality information was obtained from the 1987 NHANES I Epidemiologic Follow-up Study (1982 to 1987). Of 4,764 white sample persons, ages 25 to 74 years examined during NHANES I, 658 (13.8 percent) were identified as having pulmonary impairment defined as a FEV1/FVC < or = 69 percent. A total of 743 (15.6 percent) sample persons died during the follow-up period. The association between pulmonary impairment and all-cause mortality was examined for male and female subjects separately using the Cox proportional hazards model controlling for age, smoking, educational level, body mass index, and respiratory diseases. The analysis suggests that reduced FEV1 percent predicted was a significant risk factor for mortality among both sexes, and the FEV1/FVC ratio was significantly associated with all-cause mortality among male subjects only.