Clinical evaluation of the short half-life injectable benzodiazepine midazolam, and of the specific antagonist flumazenil in upper digestive tract endoscopy.

In a first step, midazolam 0.1 mg/kg, midazolam 0.05 mg/kg and diazepam 0.15 mg/kg administered intravenously were blindly evaluated as a sedating preparation in 3 groups of each 30 patients undergoing gastroscopy. Although amnesia is better with midazolam 0.1 mg/kg, the induced sedation is protracted, which is not to be wished in ambulatory patients. On the other side, diazepam 0.15 mg/kg was locally less well tolerated. Taking into consideration the efficacy and the general and local tolerance, the dose of midazolam 0.05 mg/kg seems the best compromise. In a second step, the specific benzodiazepine antagonist flumazenil was blindly evaluated at two intravenous doses, 1 mg and 0.5 mg, against placebo for reversal of midazolam (0.1 mg/kg) induced sedation in gastroscopy patients. 1 mg of flumazenil (and to a lesser degree 0.5 mg) suppresses the sedation and amnesia induced by midazolam and normalizes attention and sensori-motor functions. However a certain degree of resedation can reappear 45-60 min. after the administration of flumazenil. Though this phenomenon did not alter the tests measuring attention and sensorimotor functions and the recovery of memory (contrary to placebo), it incites to be cautious when the drug is given to ambulatory patients.