Current issues of ACP Journal Club are published in Annals of Internal Medicine


Flumazenil restored consciousness and improved diagnostic and therapeutic decisions in coma with suspected benzodiazepine poisoning

ACP J Club. 1991 Mar-April;114:45. doi:10.7326/ACPJC-1991-114-2-045

Source Citation

Höjer J, Baehrendtz S, Matell G, Gustafsson LL. Diagnostic utility of flumazenil in coma with suspected poisoning: a double blind, randomised controlled study. BMJ. 1990;301:1308-11.



To assess the diagnostic value and safety of flumazenil as a first-line treatment for comatose patients suspected of benzodiazepine overdose.


Randomized, double-blind, placebo-controlled therapeutic trial lasting about 20 minutes for each patient.


Medical intensive care unit of a Swedish hospital.


362 patients were treated for poisoning over a period of 18 months. 105 of these, age between 17 and 83 years, who were unconscious (scoring < 11 points on a modified [4 to 20 points] Glasgow coma scale) were included in the trial. Reasons for exclusion were pregnancy, epilepsy, or that benzodiazepine poisoning had been ruled out. 53 patients received flumazenil and 52 patients received placebo.


Patients were evaluated for the need for acute interventions and for other than routine diagnostic procedures. An intravenous injection of 10 mL of flumazenil (0.1 mg/mL), or 10 mL of placebo was given over 3 minutes. Patients were monitored for 15 minutes; then the coma score was reassessed and the trial terminated.

Main outcome measures

Frequency of interventions rendered unnecessary after the intervention; change in coma score. Blood and urine tests confirmed what poison had been ingested.

Main results

36 patients {68%}* in the flumazenil group and 37 {71%}* in the placebo group had taken benzodiazepines. 27 {51%}* of the 53 patients injected with flumazenil did not require at least 1 of the therapeutic or diagnostic interventions initially indicated, compared with no changes in interventions in the placebo group. {P<0.001}.* Flumazenil reduced indications for interventions as follows: for gastric lavage, 36% (95% CI 21% to 51%); for intubation, 40% (CI 25% to 55%); for artificial ventilation, 6% (CI 0% to 12%). None of the patients was able to respond comprehensibly at admission; within 10 minutes after the injection, 21 patients (40%) in the flumazenil group could give information on ingestion of drugs compared with 1 patient (2%) in the placebo group (P < 0.001). 9 patients in the flumazenil group (compared with 2 assigned to placebo) had adverse reactions after regaining consciousness; 8 of the 9 reactions were rated mild; 1 patient had a severe reaction — a sudden fall in blood pressure — which was reversed. All but 2 patients {group not stated} were discharged awake and in stable condition.


Flumazenil was effective in restoring consciousness in patients with suspected benzodiazepine poisoning and in reducing indications for standard therapeutic and diagnostic procedures, with few adverse effects.

Source of funding: Not stated.

Address for article reprint: Dr. J. Höjer, Medical Intensive Care Unit, Södersjukhuset, S-100 64 Stockholm, Sweden.

*Numbers calculated from data in article.


This trial confirms the utility of the antagonist, flumazenil for patients with benzodiazepine overdoses, in reducing the need for usual therapeutic interventions. In other double-blind, placebo-controlled trials in medical intensive care units, similar results were obtained by Ritz and colleagues (1) and Knudsen and associates (2), and Weinbroum and colleagues (3). Clinical depression returned within 1 hour, however, suggesting the need for repeated doses. This is not surprising because the elimination half-life is about 1 hour (4). Höjer and colleagues only addressed the issue of response to a single injection. Perhaps their success rate would have been greater with repeated injections or continuous infusion. They confirmed the presence of benzodiazepines in 36 of 53 and 37 of 52 patients in the flumazenil and placebo groups, respectively. This suggests that a single injection might be a useful diagnostic tool in the emergency department for comatose patients with suspected drug overdose. A more detailed study in this setting would be advisable.

Sharon Peters, MD
The General HospitalSt. John's, Newfoundland, Canada


1. Ritz R, Zuber M, Elsasser S, Scollo-Lavizzari G. Use of flumazenil in intoxicated patients with coma. A double-blind placebo-controlled study in ICU. Intensive Care Med. 1990;16:242-7.

2. Knudsen L, Lonka L, Srensen BH, et al. Benzodiazepine intoxication treated with flumazenil. Anaesthesia. 1988;43:274-6.

3. Weinbroum A, Rudick V, Sorkine P, et al. Use of flumazenil in the treatment of drug overdose: A double-blind and open clinical study in 110 patients. Crit Care Med. 1996;24:199-206

4. Amrein R, Leishman B, Bentzinger C, Roncari G. Flumazenil in benzodiazepine antagonism. Actions and clinical use in intoxications and anaesthesiology. Med Toxicol Adverse Drug Exp. 1987;2:411-29.