To determine whether the risk of midazolam-induced ventilatory depression is related to the rate of midazolam administration, we compared the effect of rapid (over 15 s) and slow (over 5 min) administration of midazolam (0.1 mg/kg IV) on the hypercarbic ventilatory response of 10 healthy volunteers. During the first 5 min after the start of midazolam injection, the slope of the ventilatory response to CO2 was significantly lower when the subjects received midazolam rapidly (P less than 0.001). However, after completion of the infusion (between 5 and 20 min), depression of the CO2 response curve slope was independent of the rate of midazolam administration. Similarly, although minute ventilation and tidal volume measured at an end-tidal CO2 tension of approximately 46 mm Hg decreased more quickly after rapid administration of midazolam (P less than 0.001), these variables did not differ significantly between the two rates of administration once the slow infusion was complete. These results suggest that slow administration of midazolam provides no independent protection from respiratory depression.