ObjectivesTo assess the efficacy, usefulness, safety, and dosages of flumazenil required when flumazenil is used in the diagnosis of benzodiazepine-induced coma (vs. other drug-induced coma), and to reverse or prevent the recurrence of unconsciousness.DesignA two-phase study: a controlled, randomized, doubleblind study followed by a prospective, open study.SettingAn 800-bed, teaching, university-affiliated hospital.PatientsUnconscious patients (n equals 110) suspected of benzodiazepine overdose, graded 2 to 4 on the Matthew and Lawson coma scale, were treated with flumazenil, the specific benzodiazepine receptor antagonist. The first 31 patients were studied in a double-blind fashion, while the rest of the patients were given flumazenil according to an open protocol.InterventionsAll patients received supplemental oxygen; endotracheal intubation was performed, and synchronized intermittent mandatory ventilation was initiated whenever it was deemed necessary. A peripheral intravenous cannula was inserted, as were indwelling arterial and urinary bladder catheters. Blood pressure, electrocardiogram, respiratory rate, end-tidal CO2, and core temperature were continuously monitored. The first 31 double-blind patients received either intravenous flumazenil (to a maximum of 1 mg) or saline, while the rest of the patients were given flumazenil until either regaining consciousness or a maximum of 2.5 mg was injected. Patients remaining unconscious among double-blind patients or those patients relapsing into coma after the first dose were later treated in the open phase of the study. Treatment continued by boluses or infusion as long as efficacious.Measurements and Main ResultsFourteen of 17 double-blind, flumazenil-treated patients woke after a mean of 0.8 plus minus 0.3 (SD) mg vs. one of 14 placebo patients (p less than .001). Seventy-five percent of the aggregated controlled and uncontrolled patients awoke from coma scores of 3.1 plus minus 0.6 to 0.4 plus minus 0.5 (p less than .01) after the injection of 0.7 plus minus 0.3 mg of flumazenil. These patients had high benzodiazepine serum blood concentrations. Twenty-five percent of the patients did not regain consciousness. These patients had very high serum concentrations of nonbenzodiazepine drugs. Sixty percent of the responders who had primarily ingested benzodiazepines remained awake for 72 plus minus 37 mins after flumazenil administration; 40% relapsed into coma after 18 plus minus 7 mins and various central nervous system depressant drugs were detected in their blood in addition to benzodiazepines. Seventy-one percent of the patients had ingested tricyclic antidepressants. Seventy-eight percent of the responders were continually and efficaciously treated for less than equals 8 days. Fourteen (25%) of the intubated patients were extubated safely while 12 patients, who had shown increased respiratory insufficiency, resumed satisfactory respiration after flumazenil injection. Five cases of transient increase in blood pressure and heart rate were encountered. There were 27 mildly unpleasant ``waking'' episodes, such as anxiety, restlessness, and aggression, but no patient had benzodiazepine withdrawal signs, convulsions, or dysrhythmia, most noticeably absent in tricyclic antidepressant-intoxicated patients.ConclusionsFlumazenil is a valid diagnostic tool for distinguishing pure benzodiazepine from mixed-drug intoxication or nondrug-induced coma. Flumazenil is effective in preventing recurrence of benzodiazepine-induced coma. Respiratory insufficiency is reversed after its administration. Flumazenil is safe when administered cautiously, even in patients with coma caused by a mixed overdose of benzodiazepine plus tricyclic antidepressants.(Crit Care Med 1996; 24:199-206)