OBJECTIVE To confirm the hypothesis that the ventilatory response to hypoxia (VRH) may be abolished by hypocapnia. METHODS We studied four healthy subjects during intermittent positive-pressure ventilation delivered through a nasal mask (nIPPV). Delivered minute ventilation (Ed) was progressively increased to lower end-tidal carbon dioxide pressure (PETCO(2)) below the apneic threshold. Then, at different hypocapnic levels, nitrogen was added to induce falls in oxygen saturation, a hypoxic run (N(2) run). For each N(2) run, the reappearance of a diaphragmatic muscle activity and/or an increase in effective minute ventilation (E) and/or deformations in mask-pressure tracings were considered as a VRH, whereas unchanged tracings signified absence of a VRH. For the N(2) runs eliciting a VRH, the threshold response to hypoxia (TRh) was defined as the transcutaneous oxygen saturation level that corresponds to the beginning of the ventilatory changes. RESULTS Thirty-seven N(2) runs were performed (7 N(2) runs during wakefulness and 30 N(2) runs during sleep). For severe hypocapnia (PETCO(2) of 27.1 +/- 5.2 mm Hg), no VRH was noted, whereas a VRH was observed for N(2) runs performed at significantly higher PETCO(2) levels (PETCO(2) of 34.0 +/- 2.1 mm Hg, p < 0.001). Deep oxygen desaturation (up to 64%) never elicited a VRH when the PETCO(2) level was < 29.3 mm Hg, which was considered the carbon dioxide inhibition threshold. For the 16 N(2) runs inducing a VRH, no correlations were found between PETCO(2) and TRh and between TRh and both Ed and E. CONCLUSION During nIPPV, VRH is highly dependent on the carbon dioxide level and can be definitely abolished for severe hypocapnia.