The diagnosis of sleep apnea syndrome (SAS) requires expensive and complex instrumentation. The purpose of the present study was to determine the value of end-tidal CO2 (EtCO2) in screening for sleep apneas. Thirty-nine patients referred to our sleep laboratory because of suspected SAS and ten normal subjects were studied. The EtCO2 was measured using an infrared spectrometer (POET) designed for simultaneous measurement of CO2 and pulse oximetry. In 29 subjects, expired gas was sampled with a nasobuccal mask (Respiron) with lateral orifices. In the other 20 subjects, sampling was done with nasobuccal prongs (Criticare) comprising a four-channel plastic tube to the mouth and the nostrils. Data from an 8-h night were transferred the following day to a microcomputer (Apple Macintosh) for processing. Apnea was defined as an absence of detection of CO2 for more than 10 s. Conventional polysomnography was performed (Respisomnographe). The number of apneas in 8 h and the apnea index (number of apneas in 1 h) were calculated after visual analysis on the screen of the polysomnograph and also with EtCO2 analysis. For recordings made with a nasobuccal mask, the regression curve between the apnea indices computed with EtCO2 and polysomnography was an order 2 polynomial curve (r = 0.76; p < 0.001), with an inflection point at 39 apneas per hour. For recordings with nasobuccal prongs, the correlation was very significant (r = 0.95; p < 0.0001), and the regression curve was linear. The EtCO2 with nasobuccal prongs appears to be a simple and reliable method for screening for SAS.