The association between serum beta-carotene or retinol concentration and level of ventilatory function was investigated in a population of asbestos-exposed men with a high rate of current and former cigarette smoking. The study population consisted of 816 subjects enrolled in the pilot component of the Carotene and Retinol Efficacy Trial (CARET), a placebo-controlled trial of supplemental beta-carotene and retinyl palmitate for the chemoprevention of lung cancer. Data available for analysis included baseline questionnaire, spirometry, chest X-ray, food frequency questionnaire, and serum beta-carotene and retinol concentrations. Serum beta-carotene concentration was associated with FEV1 (p < 0.05) and FVC (p < 0.05), with an approximately 100-ml increase over predicted values associated with raising the serum concentration from the 25th to the 75th percentile of the distribution in the study population (absolute difference = 155 ng/ml), even after adjustment for the confounding effects of asbestos exposure and cigarette smoking. Raising the serum retinol concentration from the 25th to the 75th percentile (absolute difference = 211 ng/ml) was associated with an approximately 70 ml increase in FVC (p < 0.05) over the predicted value. These results provide support for the hypothesis that beta-carotene and retinol have a protective effect on loss of ventilatory function.