Aliment. Pharmacol. Therap. (1990) 4, 35-42.
The value of flumazenil in the reversal of midazolarn-induced sedation for upper gastrointestinal endoscopy .
A. A. DUNK, A. C. N O R T O N , * M. HUDSON, C. R. DUNDAS," & N. ASHLEY G. M O W A T Departments of Medicine and * Anaesthesia, Aberdeen Royal Infirmary, Foresferhill, Aberdeen, UK Accepted for publication 6 October 1989
SUMMARY
Fifty patients who underwent diagnostic upper gastrointestinal endoscopy after midazolam sedation were randomized to receive (after completion of the examination) either the benzodiazepine receptor antagonist flumazenil or an identical-looking placebo. The speed of recovery from sedation was assessed by reaction time testing, measurement of critical flicker fusion frequency, and the semi-quantitative SOCA scoring system. Measurements were made up to 6 h post examination in all subjects, and at 12 and 24 h in all in-patients (n = 2'0). Flumazenil-treated patients were significantly more alert than those who received placebo at 10 min, 30 min, 1 h and 2 h ( P < 0.001 in all instances). Thereafter the two groups were similar. There was no evidence of recurrence of sedation in flumazenil-treated patients, nor did this drug adversely affect the period of anterograde amnesia between the administration of midazolam and flumazenil.
Correspondence to: Dr A. A. Dunk, Eastboume District General Hospital, Kings Drive, Eastboume, East Sussex, UK. SOCA = Sedation, Orientation, Comprehension-Collaboration, Amnesia. 35
36
A. A. D U N K et al.
INTRODUCTION Patients undergoing upper gastrointestinal endoscopy are usually sedated with an i.v. benzodiazepine compound such as diazepam or midazolam. Whilst these agents are effective and safe, they have relatively long elimination half-lives and may cause sedation which persists for several hours after completion of the examination. This is a disadvantage not only to the patient, who usually has to miss time from work and has to refrain from driving for at least 24 h, but also disadvantages the busy gastrointestinal unit, which may have to limit the number of procedures performed because of the time recovery-beds are occupied by sedated patients. In such circumstances it would clearly be advantageous to be able to return patients to full alertness upon completion of their endoscopic examination. We report a double-blind placebo-controlled study designed to evaluate whether the specific benzodiazepine receptor antagonist, flumazenil' ('Anexate ', Roche Products Ltd, Welwyn, UK), is of value in rapidly reversing the sedative effects of the benzodiazepine midazolam ('Hypnovel ', Roche Products Ltd, Welwyn, UK) in patients undergoing endoscopic examination of the upper gastrointestinal tract.
PATIENTS A N D M E T H O D S Fifty patients who underwent diagnostic upper gastrointestinal endoscopy were studied. Each gave informed consent and the study was approved by the Joint Ethics Committee of the Grampian Health Board and the University of Aberdeen. We did not study patients who fell outside the age-range 18-75 years, those with a history of hepatic or neurological disease, alcohol abusers, or those who required drugs with either sedative effects or effects on hepatic drug-metabolizing enzymes. Twenty of those studied were in-patients. Overall alertness was assessed by the measurement of the speed of reaction to light, sound and choice stimuli,' by the measurement of critical flicker fusion frequency (CFFF),3and by the estimation of SOCA score.4 Prior to sedation, speed and accuracy of reaction to light, sound and choice stimuli were measured using the Vienna reaction-time apparatus,' choice reactiontime being simply the time taken for the patient to press one of two buttons on the reaction timer, the decision depended on whether a light or sound stimulus was given. Critical flicker fusion frequency was measured by asking the patient to view a green neon bulb sited at the end of an 0.7-m long, 5-cm diameter, filtered black plastic tube, constructed such that the angle of light subtended on the fovea was less than 1' and that no reflected light was visible. The initial flicker frequency of the light was set at 25 Hz and gradually increased until the patient perceived the flicker to have disappeared. The frequency at which this occurs is affected by drugs which stimulate or depress the central nervous ~ y s t e mReaction .~ times and CFFF
FLUMAZENIL REVERSAL OF MIDAZOLAM-INDUCED SEDATION 37
were each measured six times at all study-time points (see below) and the mean values used for analysis. Patients were sedated using a constant i.v. infusion of 0.075 mg kg-‘ min-’ of midazolam. This was given until the patient was no longer able to press a handheld audible signal generator in synchrony with a similar generator ‘bleeping’ at a frequency of 1Hz. This fixed end-point of sedation allowed satisfactory endoscopy to be performed on all patients. Upon completion of the endoscopy patients were shown three 12 x 12 cm picture cards of familiar everyday objects and asked to remember them. Ten minutes after completion of the midazolam infusion, an i.v, infusion of either 0.1 mg/ml flumazenil or an identical-looking placebo was begun. These were administered at a rate of 1 ml/min until either the patient appeared alert and was able to respond in synchrony with the signal generator again, or until a total volume of 10 ml had been infused. Neuropsychological testing began 10 min after the end of the infusion of flumazenil/placebo and was repeated at 30 min, I, 2, 3, 4 and 6 h in all patients (n = 50), with additional measurements at 12 and 24 h in in-patients (n = 20). At all time-points reaction times and critical flicker fusion frequency were measured as described. The SOCA score4was also calculated. This is a simple semi-quantitative measure of overall alertness, the total score ranging from 0 to.10, and made up of the sum of individual scores (ranging from 0 to either 2 or 3 ) for four parameters, namely sedation, orientation, comprehension/collaboration, and the degree of amnesia. Amnesia was scored at the 10-min time point by asking the patient to recall the three picture cards shown immediately upon completion of the endoscopy. If none could be recalled a score of 0 was given and a maximum score of 3 was obtained if all pictures were recalled. At each time-point a further three picture cards were shown and recall for them was assessed at the subsequent time-point, thus allowing SOCA scores to be calculated for all time-points except the initial pre-sedation one. In order to avoid the hazards of multiple significance testing, the effect of treatment on critical flicker fusion frequency and each reaction time test was examined collectively at each time-point as follows. At all time-points, each variable (patient score in a particular test) was initially considered separately, and the result ranked. Thus patients had four ranks for every time-point, which reflected their performance in all tests. These four ranks were then averaged for every timepoint, thus providing patients with an average rank reflecting ’global ’ performance in the test battery for all time-points. The differences in average ranks between the placebo and flumazenil groups were formally tested using an analysis of covariance, with the baseline average rank as the co-variate, taking into account possible differences between in-patients and out-patients and the interaction between the type of patient and treatment. This analysis was performed at each time-point separately. The SOCA score results have not been subjected to formal statistical analysis, as they are semi-quantitative only and a relatively crude measure of degree of sedation when compared to the other parameters tested.
38
A. A. DUNK ef
d. Table 1. The SOCA scores in the flumazeniland placebo-treated patients (median and range)
Time
Flumazenil (n = 25)
Placebo (n = 25)
7 7-10
5
10
7 0-10
10 rnin
Median Range 30 rnin Median Range
9-10
0-8
i h Median Range 2h Median Range 3h Median Range
10 7-10
8
10 8-10
9
10 7-10
10 8-10
10 8-10
10 9-10
6-10
7-10
4h
Median Range 6h
Median Range 12 h* Median Range 24
10
10
8-10
9-10
10 9-10
10
10-10
h* Median Range
10 9-10
10
8-10
* n = 10 in each group.
RESULTS The flumazenil and placebo groups each contained 25 patients who were closely matched with respect to mean age (47.2 and 43.9 years respectively), and sex (10 and 13 males, respectively). The mean dose of midazolam required for sedation was similar in the flumazenil (11.0 mg; s.d. 2.8 mg, range 5-17 mg) and placebo (10.8 mg; s.d. 2.9 mg, range 6-18 mg) groups. The mean dose of flumazenil administered was 0.54 mg (s.d. 0.14 mg, range 0.35-0.80mg), and this was well tolerated by all subjects.
FLUMAZENIL REVERSAL OF MIDAZOLAM-INDUCED S E D A T I O N 39
Figure 1. Critical flicker fusion frequenFies of the flumazeniland placebo-treated patients (mean _+ s.e.m.)
0.0 L 1.6
I
-0
-
+ Placebo
+ Flurnarenll
--
7.
0.6
g
0.4
-
0
6 0.2 0.0 L
Figure 2. Reaction-times to light, choice and sound stimuli in flumazenil- and placebotreated patients (mean _+ s.e.m.).
0
40
A. A. D U N K ef al.
The SOCA score results of both groups are shown in Table 1. Application of this semi-quantitative parameter revealed that flumazenil-treated patients appeared to be more alert than those who received placebo up to 1 h after receiving flumazenil, thereafter the two groups appearing to be similar. At 10 min all flumazenil-treated patients were awake and fully mobile about the patient recovery area. Their median SOCA score of 7 at this time-point reflects the fact that the majority of flumazenil-treated patients (24/25) were totally amnesic for the picture cards, which were shown to them upon completion of the endoscopy but prior to L flumazenil administration. Critical flicker fusion frequency and reaction-time results for both groups are displayed in Figures I and 2, respectively. It should be noted that 10 and 6 placebotreated patients were unable to perform the critical flicker fusion frequency and reaction-time tests respectively at 10 min, whereas all flumazenil-treated patients could do so. At 30 min one placebo-treated patient remained incapable of performing either test. These patients are not represented at 10 min and 30 min in Figures 1 and 2, the data points therefore being overestimates of the performance of placebo-treated patients at these times, Analysis of the results together, as previously described, confirms that flumazenil-treated patients recovered more quickly from sedation than those who received placebo, the differences between the groups being statistically significant at 10 min, 30 min, 1h and 2 h ( P < 0.001 in all instances). There was no evidence of recurrence of sedation in either group of patients when studied at 12 and 24 h, although placebo-treated patients were significantly less alert ( P = 0.03) than the flumazenil group at 1 2 h.
DISCUSSION This study of patients who underwent diagnostic upper gastrointestinal endoscopy, demonstrates that the specific benzodiazepine receptor antagonist, flumazenil, can reverse promptly the sedative effects of the benzodiazepine compound midazolam. All the subjects who received flumazenil awoke during infusion of the drug and were fully ambulant within 5-10 min of drug administration. These findings are in broad agreement with those of other studies performed in patients undergoing a variety of diagnostic and thereapeutic procedures under benzodiazepine sedation.’t5-’ As assessed from initial SOCA score data, flumazenil did not affect the period of midazolam-induced anterograde amnesia between the administration of the two drugs. Thus 24 of the 25 flumazenil-treated patients were unable to remember events that occurred at the time of endoscopy, and this is of importance when considered in terms of patient acceptability of the need for repeated examinations. As re-sedation following discharge from the endoscopy unit is undesirable and potentially hazardous, knowledge of the likely duration of action of flumazenil, when used in this setting, is clearly required. This question has not yet been fully addressed by the endoscopy studies performed so far,5,7#yas their maximum
FLUMAZENIL REVERSAL OF MIDAZOLAM-INDUCED S E D A T I O N 41
duration of in-patient assessment has been only 4 h post-endoscopy.' It is known that the duration of action of flumazenil is dependent on the dose and type of benzodiazepine used (for example, diazepam or midazolam), the time-interval between benzodiazepine and flumazenil administration, and the dose of flumazenil used." We chose to study midazolam rather than diazepam primarily because of the shorter half-life of the former and because, unlike diazepam, it does not have active metabolites which may cause re-sedation."i12 A mean dose of approximately 11 mg of midazolam was given to both groups and, 10 min later, flumazenil, the plasma half-life of which is slightly shorter than that of midazolam at 0.7-1.3 h,' was administered in a mean dose of 0.54 mg to those in the active-treatment group. Using these drugs in this way we were unable to demonstrate any evidence of re-sedation during a maximum follow-up period of 24 h. In contrast, Holloway & Logan did find evidence of re-sedation when patients who had been endoscoped under diazepam sedation were given flumazenil and discharged from the endoscopy unit.' Although alertness was not formally tested beyond 4 h, they reported that 3 of 40 patients who received flumazenil fell asleep at home after discharge from hospital. When using flumazenil to reverse benzodiazepine-induced sedation in the day-care setting, it would therefore seem wise to use a benzodiazepine with a short plasma half-life such as midazolam, thus avoiding the risk of re-sedation. We conclude that flumazenil promptly reverses the sedative effects of midazolam when used in patients who are undergoing out-patient diagnostic endoscopy. The drug was well tolerated, did not affect the initial anterograde amnesia induced by midazolam, and there was no evidence of late re-sedation when used as described. Whilst there is little to be gained by administering flumazenil routinely to hospital in-patients following benzodiazepine sedation, in the setting of the busy out-patient endoscopy unit with a limited number of recovery beds, the routine use of flumazenil post endoscopy may represent a practical solution to this common logistical problem. ACKNOWLEDGEMENTS
We would like to thank Dr P. W. Brunt and Dr T. S. Sinclair for allowing us to study patients under their care. Statistical analyses were kindly performed by Dr Sarah Little. Flumazenil and placebo were provided by Roche Products Ltd, Welwyn, UK. REFERENCES 1 Klotz U, Kanto J. Pharmacokinetics and clinical use of flumazenil (Ro15-1788). Clin Pharmacol 1988; 14: 1-12. 2 Klensch H. Die diagnostische Valenz de Reaktionszeitmessung bei verschiedenen
zerebralen Erkrankungen. Fortschr Neurol Psychiatr 1973; 41: 575-81. 3 Simonson E, Brozek J. Flicker fusion frequency : background and applications. Physiol Rev 1952; 32: 349-78. 4 Alon E, Baitella L, Hossli E. Double-blind study of the reversal of rnidazolam-
42
5
6
7
8
A. A. D U N K ef d. supplemented general anaesthesia with RO15-1788. Br J Anaesth 1987: 59: 455-8. Kirkegaard L, Knudsen L, Jensen S, Kruse A. Benzodiazepine antagonist Ro15-1788 : antagonism of diazepam sedation in outpatients undergoing gastroscopy. Anaesthesia 1986; 41: 1184-88. Ricou B, Forster A, Brucher A, Chastonay P, Gemperle M. Clinical evaluation of a specific benzodiazepine antagonist (Ro151788). Br J Anaesth 1986; 58: 1005-11. Jensen S, h u d s e n L, Kirkegaard L. Flumazenil used in the antagonising of diazepam and midazolam sedation in outpatients undergoing gastroscopy. Eur J 'Anaesthesiol 1988 (Suppl. 2): 161-6. Rosenbaum N L, Hooper P A. The use of flumazenil as an antagonist to midazolam in intravenous sedation for dental procedures.
Eur J Anaesthesiol 1988 (Suppl. 2): 183-90. 9 Holloway A M , Logan D A . The use of flumazenil to reverse diazepam sedation after endoscopy. Eur J Anaesthesiol 1988 (SUPPI.2): 191-4. R, Hetzel W, Hartmann D, Lorscheid T. Clinical pharmacology of flumazenil. Eur J Anaesthesiol 1988 (Suppl. 2): 65-77. 11 Dundee J W, Samule I P, Toner W, Howard P J. Midazolam: a water soluble benzodiazepine-studies in volunteers. Anaesthesia 1980; 35 : 454-8. 12 Smith M T, Eadie M J, ORouke Brophy T. The pharmacokinetics of midazolam in man. Eur J Clin Pharmacol 1981; 19: 271-8.
10 Amrein
The value of flumazenil in the reversal of midazolarn-induced sedation for upper gastrointestinal endoscopy .
A. A. DUNK, A. C. N O R T O N , * M. HUDSON, C. R. DUNDAS," & N. ASHLEY G. M O W A T Departments of Medicine and * Anaesthesia, Aberdeen Royal Infirmary, Foresferhill, Aberdeen, UK Accepted for publication 6 October 1989
SUMMARY
Fifty patients who underwent diagnostic upper gastrointestinal endoscopy after midazolam sedation were randomized to receive (after completion of the examination) either the benzodiazepine receptor antagonist flumazenil or an identical-looking placebo. The speed of recovery from sedation was assessed by reaction time testing, measurement of critical flicker fusion frequency, and the semi-quantitative SOCA scoring system. Measurements were made up to 6 h post examination in all subjects, and at 12 and 24 h in all in-patients (n = 2'0). Flumazenil-treated patients were significantly more alert than those who received placebo at 10 min, 30 min, 1 h and 2 h ( P < 0.001 in all instances). Thereafter the two groups were similar. There was no evidence of recurrence of sedation in flumazenil-treated patients, nor did this drug adversely affect the period of anterograde amnesia between the administration of midazolam and flumazenil.
Correspondence to: Dr A. A. Dunk, Eastboume District General Hospital, Kings Drive, Eastboume, East Sussex, UK. SOCA = Sedation, Orientation, Comprehension-Collaboration, Amnesia. 35
36
A. A. D U N K et al.
INTRODUCTION Patients undergoing upper gastrointestinal endoscopy are usually sedated with an i.v. benzodiazepine compound such as diazepam or midazolam. Whilst these agents are effective and safe, they have relatively long elimination half-lives and may cause sedation which persists for several hours after completion of the examination. This is a disadvantage not only to the patient, who usually has to miss time from work and has to refrain from driving for at least 24 h, but also disadvantages the busy gastrointestinal unit, which may have to limit the number of procedures performed because of the time recovery-beds are occupied by sedated patients. In such circumstances it would clearly be advantageous to be able to return patients to full alertness upon completion of their endoscopic examination. We report a double-blind placebo-controlled study designed to evaluate whether the specific benzodiazepine receptor antagonist, flumazenil' ('Anexate ', Roche Products Ltd, Welwyn, UK), is of value in rapidly reversing the sedative effects of the benzodiazepine midazolam ('Hypnovel ', Roche Products Ltd, Welwyn, UK) in patients undergoing endoscopic examination of the upper gastrointestinal tract.
PATIENTS A N D M E T H O D S Fifty patients who underwent diagnostic upper gastrointestinal endoscopy were studied. Each gave informed consent and the study was approved by the Joint Ethics Committee of the Grampian Health Board and the University of Aberdeen. We did not study patients who fell outside the age-range 18-75 years, those with a history of hepatic or neurological disease, alcohol abusers, or those who required drugs with either sedative effects or effects on hepatic drug-metabolizing enzymes. Twenty of those studied were in-patients. Overall alertness was assessed by the measurement of the speed of reaction to light, sound and choice stimuli,' by the measurement of critical flicker fusion frequency (CFFF),3and by the estimation of SOCA score.4 Prior to sedation, speed and accuracy of reaction to light, sound and choice stimuli were measured using the Vienna reaction-time apparatus,' choice reactiontime being simply the time taken for the patient to press one of two buttons on the reaction timer, the decision depended on whether a light or sound stimulus was given. Critical flicker fusion frequency was measured by asking the patient to view a green neon bulb sited at the end of an 0.7-m long, 5-cm diameter, filtered black plastic tube, constructed such that the angle of light subtended on the fovea was less than 1' and that no reflected light was visible. The initial flicker frequency of the light was set at 25 Hz and gradually increased until the patient perceived the flicker to have disappeared. The frequency at which this occurs is affected by drugs which stimulate or depress the central nervous ~ y s t e mReaction .~ times and CFFF
FLUMAZENIL REVERSAL OF MIDAZOLAM-INDUCED SEDATION 37
were each measured six times at all study-time points (see below) and the mean values used for analysis. Patients were sedated using a constant i.v. infusion of 0.075 mg kg-‘ min-’ of midazolam. This was given until the patient was no longer able to press a handheld audible signal generator in synchrony with a similar generator ‘bleeping’ at a frequency of 1Hz. This fixed end-point of sedation allowed satisfactory endoscopy to be performed on all patients. Upon completion of the endoscopy patients were shown three 12 x 12 cm picture cards of familiar everyday objects and asked to remember them. Ten minutes after completion of the midazolam infusion, an i.v, infusion of either 0.1 mg/ml flumazenil or an identical-looking placebo was begun. These were administered at a rate of 1 ml/min until either the patient appeared alert and was able to respond in synchrony with the signal generator again, or until a total volume of 10 ml had been infused. Neuropsychological testing began 10 min after the end of the infusion of flumazenil/placebo and was repeated at 30 min, I, 2, 3, 4 and 6 h in all patients (n = 50), with additional measurements at 12 and 24 h in in-patients (n = 20). At all time-points reaction times and critical flicker fusion frequency were measured as described. The SOCA score4was also calculated. This is a simple semi-quantitative measure of overall alertness, the total score ranging from 0 to.10, and made up of the sum of individual scores (ranging from 0 to either 2 or 3 ) for four parameters, namely sedation, orientation, comprehension/collaboration, and the degree of amnesia. Amnesia was scored at the 10-min time point by asking the patient to recall the three picture cards shown immediately upon completion of the endoscopy. If none could be recalled a score of 0 was given and a maximum score of 3 was obtained if all pictures were recalled. At each time-point a further three picture cards were shown and recall for them was assessed at the subsequent time-point, thus allowing SOCA scores to be calculated for all time-points except the initial pre-sedation one. In order to avoid the hazards of multiple significance testing, the effect of treatment on critical flicker fusion frequency and each reaction time test was examined collectively at each time-point as follows. At all time-points, each variable (patient score in a particular test) was initially considered separately, and the result ranked. Thus patients had four ranks for every time-point, which reflected their performance in all tests. These four ranks were then averaged for every timepoint, thus providing patients with an average rank reflecting ’global ’ performance in the test battery for all time-points. The differences in average ranks between the placebo and flumazenil groups were formally tested using an analysis of covariance, with the baseline average rank as the co-variate, taking into account possible differences between in-patients and out-patients and the interaction between the type of patient and treatment. This analysis was performed at each time-point separately. The SOCA score results have not been subjected to formal statistical analysis, as they are semi-quantitative only and a relatively crude measure of degree of sedation when compared to the other parameters tested.
38
A. A. DUNK ef
d. Table 1. The SOCA scores in the flumazeniland placebo-treated patients (median and range)
Time
Flumazenil (n = 25)
Placebo (n = 25)
7 7-10
5
10
7 0-10
10 rnin
Median Range 30 rnin Median Range
9-10
0-8
i h Median Range 2h Median Range 3h Median Range
10 7-10
8
10 8-10
9
10 7-10
10 8-10
10 8-10
10 9-10
6-10
7-10
4h
Median Range 6h
Median Range 12 h* Median Range 24
10
10
8-10
9-10
10 9-10
10
10-10
h* Median Range
10 9-10
10
8-10
* n = 10 in each group.
RESULTS The flumazenil and placebo groups each contained 25 patients who were closely matched with respect to mean age (47.2 and 43.9 years respectively), and sex (10 and 13 males, respectively). The mean dose of midazolam required for sedation was similar in the flumazenil (11.0 mg; s.d. 2.8 mg, range 5-17 mg) and placebo (10.8 mg; s.d. 2.9 mg, range 6-18 mg) groups. The mean dose of flumazenil administered was 0.54 mg (s.d. 0.14 mg, range 0.35-0.80mg), and this was well tolerated by all subjects.
FLUMAZENIL REVERSAL OF MIDAZOLAM-INDUCED S E D A T I O N 39
Figure 1. Critical flicker fusion frequenFies of the flumazeniland placebo-treated patients (mean _+ s.e.m.)
0.0 L 1.6
I
-0
-
+ Placebo
+ Flurnarenll
--
7.
0.6
g
0.4
-
0
6 0.2 0.0 L
Figure 2. Reaction-times to light, choice and sound stimuli in flumazenil- and placebotreated patients (mean _+ s.e.m.).
0
40
A. A. D U N K ef al.
The SOCA score results of both groups are shown in Table 1. Application of this semi-quantitative parameter revealed that flumazenil-treated patients appeared to be more alert than those who received placebo up to 1 h after receiving flumazenil, thereafter the two groups appearing to be similar. At 10 min all flumazenil-treated patients were awake and fully mobile about the patient recovery area. Their median SOCA score of 7 at this time-point reflects the fact that the majority of flumazenil-treated patients (24/25) were totally amnesic for the picture cards, which were shown to them upon completion of the endoscopy but prior to L flumazenil administration. Critical flicker fusion frequency and reaction-time results for both groups are displayed in Figures I and 2, respectively. It should be noted that 10 and 6 placebotreated patients were unable to perform the critical flicker fusion frequency and reaction-time tests respectively at 10 min, whereas all flumazenil-treated patients could do so. At 30 min one placebo-treated patient remained incapable of performing either test. These patients are not represented at 10 min and 30 min in Figures 1 and 2, the data points therefore being overestimates of the performance of placebo-treated patients at these times, Analysis of the results together, as previously described, confirms that flumazenil-treated patients recovered more quickly from sedation than those who received placebo, the differences between the groups being statistically significant at 10 min, 30 min, 1h and 2 h ( P < 0.001 in all instances). There was no evidence of recurrence of sedation in either group of patients when studied at 12 and 24 h, although placebo-treated patients were significantly less alert ( P = 0.03) than the flumazenil group at 1 2 h.
DISCUSSION This study of patients who underwent diagnostic upper gastrointestinal endoscopy, demonstrates that the specific benzodiazepine receptor antagonist, flumazenil, can reverse promptly the sedative effects of the benzodiazepine compound midazolam. All the subjects who received flumazenil awoke during infusion of the drug and were fully ambulant within 5-10 min of drug administration. These findings are in broad agreement with those of other studies performed in patients undergoing a variety of diagnostic and thereapeutic procedures under benzodiazepine sedation.’t5-’ As assessed from initial SOCA score data, flumazenil did not affect the period of midazolam-induced anterograde amnesia between the administration of the two drugs. Thus 24 of the 25 flumazenil-treated patients were unable to remember events that occurred at the time of endoscopy, and this is of importance when considered in terms of patient acceptability of the need for repeated examinations. As re-sedation following discharge from the endoscopy unit is undesirable and potentially hazardous, knowledge of the likely duration of action of flumazenil, when used in this setting, is clearly required. This question has not yet been fully addressed by the endoscopy studies performed so far,5,7#yas their maximum
FLUMAZENIL REVERSAL OF MIDAZOLAM-INDUCED S E D A T I O N 41
duration of in-patient assessment has been only 4 h post-endoscopy.' It is known that the duration of action of flumazenil is dependent on the dose and type of benzodiazepine used (for example, diazepam or midazolam), the time-interval between benzodiazepine and flumazenil administration, and the dose of flumazenil used." We chose to study midazolam rather than diazepam primarily because of the shorter half-life of the former and because, unlike diazepam, it does not have active metabolites which may cause re-sedation."i12 A mean dose of approximately 11 mg of midazolam was given to both groups and, 10 min later, flumazenil, the plasma half-life of which is slightly shorter than that of midazolam at 0.7-1.3 h,' was administered in a mean dose of 0.54 mg to those in the active-treatment group. Using these drugs in this way we were unable to demonstrate any evidence of re-sedation during a maximum follow-up period of 24 h. In contrast, Holloway & Logan did find evidence of re-sedation when patients who had been endoscoped under diazepam sedation were given flumazenil and discharged from the endoscopy unit.' Although alertness was not formally tested beyond 4 h, they reported that 3 of 40 patients who received flumazenil fell asleep at home after discharge from hospital. When using flumazenil to reverse benzodiazepine-induced sedation in the day-care setting, it would therefore seem wise to use a benzodiazepine with a short plasma half-life such as midazolam, thus avoiding the risk of re-sedation. We conclude that flumazenil promptly reverses the sedative effects of midazolam when used in patients who are undergoing out-patient diagnostic endoscopy. The drug was well tolerated, did not affect the initial anterograde amnesia induced by midazolam, and there was no evidence of late re-sedation when used as described. Whilst there is little to be gained by administering flumazenil routinely to hospital in-patients following benzodiazepine sedation, in the setting of the busy out-patient endoscopy unit with a limited number of recovery beds, the routine use of flumazenil post endoscopy may represent a practical solution to this common logistical problem. ACKNOWLEDGEMENTS
We would like to thank Dr P. W. Brunt and Dr T. S. Sinclair for allowing us to study patients under their care. Statistical analyses were kindly performed by Dr Sarah Little. Flumazenil and placebo were provided by Roche Products Ltd, Welwyn, UK. REFERENCES 1 Klotz U, Kanto J. Pharmacokinetics and clinical use of flumazenil (Ro15-1788). Clin Pharmacol 1988; 14: 1-12. 2 Klensch H. Die diagnostische Valenz de Reaktionszeitmessung bei verschiedenen
zerebralen Erkrankungen. Fortschr Neurol Psychiatr 1973; 41: 575-81. 3 Simonson E, Brozek J. Flicker fusion frequency : background and applications. Physiol Rev 1952; 32: 349-78. 4 Alon E, Baitella L, Hossli E. Double-blind study of the reversal of rnidazolam-
42
5
6
7
8
A. A. D U N K ef d. supplemented general anaesthesia with RO15-1788. Br J Anaesth 1987: 59: 455-8. Kirkegaard L, Knudsen L, Jensen S, Kruse A. Benzodiazepine antagonist Ro15-1788 : antagonism of diazepam sedation in outpatients undergoing gastroscopy. Anaesthesia 1986; 41: 1184-88. Ricou B, Forster A, Brucher A, Chastonay P, Gemperle M. Clinical evaluation of a specific benzodiazepine antagonist (Ro151788). Br J Anaesth 1986; 58: 1005-11. Jensen S, h u d s e n L, Kirkegaard L. Flumazenil used in the antagonising of diazepam and midazolam sedation in outpatients undergoing gastroscopy. Eur J 'Anaesthesiol 1988 (Suppl. 2): 161-6. Rosenbaum N L, Hooper P A. The use of flumazenil as an antagonist to midazolam in intravenous sedation for dental procedures.
Eur J Anaesthesiol 1988 (Suppl. 2): 183-90. 9 Holloway A M , Logan D A . The use of flumazenil to reverse diazepam sedation after endoscopy. Eur J Anaesthesiol 1988 (SUPPI.2): 191-4. R, Hetzel W, Hartmann D, Lorscheid T. Clinical pharmacology of flumazenil. Eur J Anaesthesiol 1988 (Suppl. 2): 65-77. 11 Dundee J W, Samule I P, Toner W, Howard P J. Midazolam: a water soluble benzodiazepine-studies in volunteers. Anaesthesia 1980; 35 : 454-8. 12 Smith M T, Eadie M J, ORouke Brophy T. The pharmacokinetics of midazolam in man. Eur J Clin Pharmacol 1981; 19: 271-8.
10 Amrein
