To test the hypothesis that during cardiopulmonary resuscitation, chest compression with an unobstructed trachea raises and maintains intrathoracic pressure by collapsing airways and trapping air in the lung, we studied 11 dogs (20-32 kg). An inflatable vest compressed the thorax after induction of ventricular fibrillation. First, tracheal airflow was measured by a pneumotachometer during vest inflation and deflation in nine of the dogs. As expected, during the initial phase of vest inflation of cycles after ventilation, air moved out of the lungs, but then airflow stopped. After vest deflation, however, more air moved out of the lungs in eight of the nine dogs; this occurrence indicated that a portion of the inspired tidal volume was trapped during vest inflation. During cycles without prior ventilation, the amount of air expired by chest compression decreased, paradoxically, at higher peak vest pressures (p<0.002); this occurrence indicated that air was trapped at the higher vest pressures. The change in right atrial pressure was higher on cycles after ventilation than on cycles without prior ventilation (79 ± 12 vs. 67 ± 12 mm Hg [mean ± SEM],p<0.005), and lung volume was higher on cycles after ventilation (p<0.001). Next, a 5-Fr micromanometer was advanced down the airway in eight of the dogs. With the tip of the micromanometer 5-8 cm distal to the carina, a zone of high pressure was noted in seven dogs; this high pressure suggested a zone of airway collapse distal to the carina. With the micromanometer in the high-pressure zone, there was a correlation between the change in airway pressure and the change in right atrial pressure (r=0.90,p<0.001). Finally, contrast was instilled in the airways of four dogs. In all four dogs, collapse of 3-6-mm airways could be visualized during cineradiography. We conclude that air trapping occurs during vest inflation and that this phenomenon explains how intrathoracic pressure can be generated and maintained despite an unobstructed trachea.