The effect of midazolam on the hypoxic ventilatory response of eight healthy volunteers was examined during isocapnic rebreathing. The magnitude of the slope of the ventilatory response to hypoxia (VE vs SaO2) decreased from 1.48 +/- 0.24 to 0.70 +/- 0.13 L.min-1.%SaO2(-1) (means +/- SE, P less than 0.005) after midazolam 0.1 mg/kg IV. The calculated ventilation at an arterial saturation of 90% also decreased from 28.6 +/- 4.4 to 19.9 +/- 2.7 L/min (P less than 0.05). Before midazolam, hypoxia to an SaO2 of 75 +/- 2% was associated with a 23 +/- 3 beats/min increase in heart rate; after midazolam, the increase in heart rate with hypoxia was only 4 +/- 2 beats/min (P less than 0.001). Additionally, a double-blind crossover study evaluated the effect of physostigmine on awareness and hypoxic ventilatory response after midazolam. The change in hypoxic response slope after physostigmine 2.0 mg IV (an increase of 0.28 +/- 0.34 L.min-1.%SaO2(-1] did not differ significantly from that after placebo (an increase of 0.03 +/- 0.22 L.min-1.%SaO2(-1], although physostigmine significantly increased awareness. It is concluded that a sedative dose of midazolam depresses hypoxic ventilatory response and attenuates the hyperpnea and tachycardia associated with hypoxemia. Furthermore, physostigmine-glycopyrrolate reversal of midazolam-induced sedation was associated with nausea (five subjects), vomiting (three subjects), and tachycardia without reversal of the depressed hypoxic ventilatory response.