We evaluated the effects of a benzodiazepine antagonist, flumazenil (Ro 15-1788), 1 mg intravenous, on the hypoxic ventilatory response in 10 healthy patients who had received diazepam 0.97 +/- 0.34 mg/minute (mean +/- SD) for sedation during minor operative procedures. When assessed using a sedation analog scale, flumazenil significantly decreased diazepam-induced sedation. In only two of the five patients with evidence of diazepam-induced depression of the ventilatory response to hypoxia was there significant reversal of this depression after flumazenil administration. Finally, patients in whom the duration of sedation with diazepam was shorter (78 +/- 14 minutes) had a significantly greater decrease in the slope of their hypoxic ventilatory response curve (1.3 +/- 0.8 L.min-1.%sat-1 [-43% from baseline]) than did patients with longer sedation times (145 +/- 37 minutes duration; 2.0 +/- 0.8 L.min-1.%sat-1 [-5% from baseline]). These data suggest that flumazenil, 1 mg intravenous, is only partially effective in reversing diazepam-induced depression of hypoxic ventilatory drive and that tolerance may develop to the respiratory depressant effects of diazepam.