To examine the usefulness of flow-volume curves as a screening test for the diagnosis of obstructive sleep apnea (OSA), we studied 405 consecutive patients referred for evaluation of possible OSA. All patients had full pulmonary function studies, which included measurements of maximal inspiratory and expiratory flow-volume curves, and nocturnal polysomnography, including continuous monitoring of snoring sounds. When the results were analyzed, of the 405 patients studied, 207 had OSA (apnea/hypopnea index [AHI] greater than 10) and 198 did not. Flow-volume curves were examined for the presence of upper airway obstruction defined as midvital capacity flow ratio (MVCFR = ratio of the maximal expiratory flow at 50% of vital capacity to maximal inspiratory flow at 50% of vital capacity) greater than 1.0. We found no significant difference in the values of MVCFR between the two groups: MVCFR was equal to 0.69 +/- 0.31 for nonapneic snorers, and 0.68 +/- 0.29 for the apneic snorers. Furthermore, we redefined the apneic and nonapneic groups using different cutoff values of AHI: 20, 30, 40, and 50. Independently of the AHI cutoff used, we found no significant difference in the MVCFR between the two groups. Linear regression analysis for the entire group of 405 patients revealed no significant correlation between MVCFR and the AHI or the snoring indices. We calculated the sensitivity, specificity, and predictive values of MVCFR for the diagnosis of OSA. We found that this test had 12% sensitivity, 86% specificity, 47% positive predictive value, and 46% negative predictive value.(ABSTRACT TRUNCATED AT 250 WORDS)