Empiric Use of Flumazenil in Comatose Patients: Limited Applicability of Criteria to Define Low Risk

Abstract

Study objective: To develop clinical rules for the safe and effective use of flumazenil in suspected benzodiazepine overdose. Methods: We assembled a retrospective series of 35 consecutive comatose patients admitted between October 1992 and July 1993 to a toxicologic ICU with the presumptive diagnosis of drug overdose. These patients were divided into two groups. Group A (low-risk) patients had a clinical picture compatible with uncomplicated benzodiazepine intoxication (calm, without abnormalities in pulse or blood pressure, lateralizing signs, hypertonia, hyperreflexia, or myoclonus) in the absence of predefined electrocardiographic or clinical signs of tricyclic antidepressant or other proconvulsant overdose, and absence of an available history of long-term benzodiazepine treatment or an underlying seizure disorder. Group B ("non-low risk") comprised all other patients. Efficacy of flumazenil was categorized as complete awakening (with normal level of alertness), partial awakening, or no change in alertness level. The safety of flumazenil was defined on the basis of the absence of seizures or death. Results: In group A (n=4), flumazenil was associated with complete awakening in three patients and partial awakening in one. No seizures were observed. In group B (n=31), flumazenil was associated with complete awakening in 4 patients, partial awakening in 5, and no response in 22. In group B, five seizures occurred. Conclusion: Comatose patients with clinical or ECG criteria thought to contraindicate the use of flumazenil have a reasonably high risk of seizures after administration of this drug. Low-risk patients may be able to receive flumazenil safely, but they may be only a small portion of comatose patients with suspected overdose. [Gueye PN, Hoffman JR, Taboulet P, Vicaut E, Baud FJ: Empiric use of flumazenil in comatose patients: Limited applicability of criteria to define low risk. Ann Emerg Med June 1996;27:730-735.]