Does Midazolam Cause Retrograde Amnesia, and Can Flumazenil Reverse That Amnesia?

Abstract

Midazolam is a benzodiazepine agonist that has a rapid onset and fast distribution and that causes few cardiovascular or respiratory effects [1]. Midazolam is water soluble and has a low incidence of venous irritation after intravenous (IV) injection [2]. These properties make midazolam a good choice for conscious sedation and for induction and maintenance of anesthesia. Although midazolam causes anterograde amnesia [1], retrograde amnesia has not been demonstrated [3]. We report an incidence of profound amnesia, anterograde and retrograde, in a patient after a laparoscopic cholecystectomy. The patient's memory was restored after IV injections of the benzodiazepine antagonist flumazenil. Case Report A 61-yr-old man (84 kg, 177 cm, ASA physical status III) was admitted for a laparoscopic cholecystectomy. His medical history included coronary artery bypass grafting 9 mo earlier. Preoperative laboratory tests were within normal limits. We applied routine monitoring and induced anesthesia with fentanyl 150 micro g, midazolam 2 mg, thiopental 375 mg, and vecuronium 8 mg IV. Anesthesia was maintained with nitrous oxide 65%, oxygen 34%, and isoflurane 1%. Intraoperatively, oxygen saturation remained between 98% and 100%. Blood pressure ranged from 120/70 to 140/90 mm Hg, with a heart rate of 60-70 bpm. End-tidal carbon dioxide was 27 mm Hg, which increased to 35 mm Hg during the procedure. The surgery and anesthetic course were uneventful. The vecuronium was reversed with neostigmine 4 mg and glycopyrrolate 0.6 mg. The patient demonstrated spontaneous breathing with an adequate minute volume and was tracheally extubated. Upon admission to the postanesthesia care unit (PACU), blood pressure was 142/92 mm Hg, heart rate was 68 bpm, respirations were 20 breaths/min, and oxygen saturation was 98%. It was 1 h and 40 min after induction of anesthesia. The patient appeared awake and could follow simple commands, but he was restless and disoriented. He was unaware of his surroundings or that he had undergone surgery. He denied knowing his name, his wife's name, or his children's names. The patient was closely monitored in the PACU during the next 2 h with no improvement in memory. Flumazenil 0.2 mg IV was administered and resulted in immediate improvement in mental status. One minute later, a second dose of 0.2 mg was administered, after which the patient demonstrated complete return of both long- and short-term memory. When the patient was informed that 5 min earlier he had not known his wife's name, he replied, "How could I forget my wife of 37 years?" Thirty minutes later, the patient was transferred to a surgical floor. The remainder of his hospitalization was uneventful. Discussion Our patient exhibited profound amnesia, both anterograde and retrograde. Memory was assessed using direct (free recall) measures. Indirect (wordstem completion) measures were not used. Was our patient's memory loss pure amnesia, or was it a result of prolonged sedation or abnormal kinetics caused by interaction with narcotics, anesthesia, and the alteration in CO2 level? Anterograde amnesia has been demonstrated in adult patients after administration of midazolam IV [4], intramuscularly [5], and orally [6]. Retrograde amnesia has not been shown with midazolam use [7,8]. Twersky et al. [9] investigated the effect of midazolam sedation on retrograde and anterograde recall and recognition in children. They found that midazolam diminishes anterograde recognition with no effect on retrograde memory. Ghoneim and Mewaldt [3] report that benzodiazepines impair acquisition of new information without affecting retention or retrieval of previously stored information. Midazolam is known to produce sedation as well as amnesia [2]. It has been argued that impaired memory is a result of the sedation rather than a specific effect of the drug [3]. Our patient was not sedated according to a subjective rating scale. Instead of sedating a patient, midazolam can sometimes trigger hostility, rage, and physical violence [10]. This is known as a paradoxical reaction. Our patient was restless and disoriented but not hostile or violent, therefore ruling out a paradoxical reaction. The amnesia was not the result of a chemical impairment, since postoperative values for electrolytes, glucose, urea, and ammonia were normal. Narcotic effect was ruled out on the following clinical evidence: pupils were normal size and reactive, and respirations were 20 breaths/min with an adequate minute ventilation and normal chest expansion. Hypoxemia was ruled out because oxygen saturation had been adequate during the entire perioperative period. Selective cerebral infarction appeared unsupported since there had been no signs of hypoxemia and vital signs had remained stable. In addition, the patient had been positioned with the head in the neutral position during surgery; therefore, blood flow to the brain had not been compromised. Although it is rarely so specific, an embolic phenomenon can not be excluded without radiologic evaluation of the brain. Since the patient's memory was restored after the administration of flumazenil, a cerebral computed tomography scan was not performed. Flumazenil is a benzodiazepine receptor antagonist that inhibits the central effects of benzodiazepine agonists by reversible competitive binding at the benzodiazepine receptor. The onset of clinical effects occurs within one to five minutes [11]. Flumazenil reverses the amnesiac and sedative effects of midazolam [12,13]. A multicenter clinical study, however, found that flumazenil was less effective in reversing midazolam-induced amnesia than sedation, with only 60% of patients demonstrating partial recovery of memory [14]. Curran and Birch [15] found no difference between flumazenil and placebo on free recall, delayed recall, and recognition in patients who had received midazolam. In our patient, improvement of short- and long-term memory was demonstrated within one minute of receiving flumazenil. It is interesting, however, that flumazenil completely reversed the retrograde amnesia but not the anterograde amnesia, since the patient had no recall of his time spent in the operating room or in the PACU prior to receiving the flumazenil. While further research is needed to determine the exact role that midazolam and flumazenil play in anterograde and retrograde amnesia, health-care providers should consider the residual effects of midazolam when a patient exhibits signs of anterograde or retrograde amnesia. A trial reversal with flumazenil should be attempted for any patient who has received midazolam and who exhibits persistent and unusual sedation or amnesia. We thank the staff of the Health Science Library at Christ Hospital for help with literature searches and Ken Scholz, PhD, of the Office of Scientific Communications for help in editing the manuscript.