We examined flow-volume curves for their potential as screening tests for obstructive sleep apnea (OSA) in 401 patients referred for investigation of snoring. In all patients, we performed nocturnal polysomnography, maximum inspiratory and maximum expiratory flow-volume curves. The curves were examined for two features: 1) presence of flow oscillations (the "saw-tooth" sign), and 2) changes in their configuration that might suggest upper airway obstruction as documented by the expiratory/inspiratory flow ratios calculated at 50 and 75 percent of exhaled vital capacity (FR50 and FR25, respectively). Based on the results of nocturnal polysomnography, the patients were stratified according to severity into apnea groups, and the flow ratios and flow oscillations were compared among these groups. We found that neither the FR50 nor FR25 were significantly different among the groups. Inspiratory and expiratory flow oscillations were seen infrequently (32 patients) and tended to occur in patients with more severe sleep apnea. Neither the flow ratios nor the flow oscillations had good predictive values for snoring or sleep apnea. The sensitivity of the flow volume curve abnormalities ranged between 0 and 14 percent, but the specificity was high, ranging between 93 and 95 percent. We conclude that because of low sensitivity, flow-volume loops are not a useful screening test for the diagnosis of OSA in snoring patients.