The effects of constant-flow insufflation (CFI) of air in the trachea at the distal end of a modified endotracheal tube as the sole mode of ventilation during cardiopulmonary resuscitation (CPR) were studied in pigs. The ventilatory effect of CFI (15 +/- 2 L/min) generating a positive pressure of about 10 cm H2O with concomitant chest compression was studied first. In nine sedated, paralyzed animals disconnected from the ventilator, CFI alone did not significantly alter the decrease in PaO2 and the rise in PaCO2 observed during apnea. By contrast, the combination of precordial compression and CFI (CFI-CPR) maintained arterial blood gases over a 4-min period at the level obtained during mechanical ventilation. In the second part of the study, ventricular fibrillation was induced and CFI-CPR was compared with standard CPR using conventional mechanical ventilation during two successive 4-min periods, in random order. Ventilatory parameters were identical in the two situations, whereas hemodynamic parameters were similar or better with CFI-CPR than with standard CPR. Significant differences were observed between standard CPR and CFI-CPR for systolic aortic pressure (72 +/- 22 versus 82 +/- 27 mm Hg, respectively; p < 0.02) and for systolic (322 +/- 216 versus 431 +/- 237 ml/s; p < 0.01) and mean (116 +/- 106 versus 143 +/- 108 ml/s; p < 0.01) common carotid blood flows. The ease of use of CFI together with its beneficial hemodynamic effects suggests that CFI deserves to be investigated further as a mode of ventilation during CPR.