0016-5107/92/3806-0651$03.00 GASTROINTESTINAL ENDOSCOPY Copyright © 1992 by the American Society for Gastrointestinal Endoscopy
Diazepam versus midazolam for colonoscopy: a prospective evaluation of predicted versus actual dosing requirements Gregory G. Ginsberg, MD, James H. Lewis, Jane E. Gallagher, RN, David E. Fleischer, Firas H. AI-Kawas, MD, Cuong C. Nguyen, Diane J. Mundt, PhD, Stanley B. Benjamin,
MD MD MD MD
Washington, D.C.
We performed a prospective, randomized, double-blind study to evaluate the efficacy of the currently recommended low doses of midazolam for conscious sedation compared with diazepam for colonoscopy. Each agent was administered in a fixed ratio dose in combination with meperidine, and titrated incrementally to allow for adequate sedation prior to initiating and during the procedure. The currently recommended starting dose of midazolam (0.03 mg/kg) proved to be very appropriate for pre-medication. In contrast, the currently recommended starting dose of diazepam (0.10 mg/kg) proved excessive in 21% of patients (especially in those aged >65). The low initial and incremental doses of midazolam compared favorably with diazepam in all efficacy parameters studied and exceeded diazepam in post-procedure amnesia scores (p 0.01). Moreover, the sedative effects of midazolam at these lower doses were not lost despite long duration procedures (>40 min). We conclude that midazolam, given in small incremental doses, in combination with meperidine, produces effective conscious sedation for colonoscopy and exceeds diazepam in its amnestic effect. (Gastrointest Endosc 1992;38:651-656)
=
Diazepam and midazolam are the two most widely used benzodiazepines for endoscopic conscious sedation.1,2 Although midazolam has been cited as having a number of potential advantages over diazepam, such as a shorter elimination half-life, a lower incidence of phlebitis, and superior anterograde amnesia,!·3 it remains controversial whether in fact one or the other is optimal for conscious sedation. 4 • 5 Much of the current controversy centers on the issue of safety. Following the introduction of midazolam in 1986, several serious cardiorespiratory events, including some deaths that were potentially related to its use, were reported to the U.S. Food and Drug Administration (FDA).6 These reports raised concerns that the use of midazolam for endoscopic conscious sedation might Received April 20, 1992. For revision June 1, 1992. Accepted July 28,1992. From the Division of Gastroenterology, Georgetown University Medical Center, Washington, D.C. Reprint requests: Stanley B. Benjamin, MD, Division of Gastroenterology, Georgetown University Medical Center, 3800 Reservoir Road, N. w., Washington, D.C. 20007. VOLUME 38, NO.6, 1992
constitute an overly serious risk. 7 Midazolam is three to four times as potent as diazepam 8 and was in many instances being administered in too large a dose. 9 Although midazolam has a more rapid onset of action compared with diazepam,3 recent studies have demonstrated that the peak drug effect occurs later for midazolam than for diazepam. lO This may have resulted in inadequate intervals of observation between repeat dosing of midazolam, thereby contributing to oversedation and medication associated complications. In response to these issues, it was recommended that the initial and incremental doses of midazolam be decreased for conscious sedation, and a physician education campaign was mounted that emphasized a titrated method of administration, taking into account the patient's age, underlying cardiopulmonary diseases, and the concurrent use of other agents during endoscopy. These efforts led to a significant reduction in the frequency of reported adverse cardiopulmonary events associated with midazolam. 9 Arrowsmith et 651
alY have recently compared the rates of serious cardiorespiratory complications and deaths associated with the use of midazolam or diazepam for endoscopic conscious sedation, and found no significant differences among 21,000 procedures from 17 medical centers that were recorded in the American Society for Gastrointestinal Endoscopy computer-based management system. In spite of midazolam's current widespread acceptance and use as an agent for endoscopic conscious sedation, no prospective study has compared diazepam with midazolam at the lower dosing schedules now considered appropriate for conscious sedation. Several issues remain unanswered: are the lower dosing schedules safer? Is the superior anterograde amnesia produced by midazolam at higher doses maintained at lower doses? Do the newer recommended doses of midazolam correlate with the actual doses required to achieve adequate sedation to initiate endoscopic procedures and allow maintenance of sedation throughout the procedure? Finally, there is concern that at the lower dosing regimens, the efficacy of midazolam may be lost during procedures of longer duration, such as colonoscopy, owing to the drug's shorter elimination half-life. 4 In order to answer questions concerning the efficacy of midazolam at lower doses, we undertook a prospective study of the predicted versus actual dosing requirements for midazolam compared with diazepam during colonoscopy. PATIENTS AND METHODS
This study was a randomized, double-blind comparison of the effectiveness of midazolam and diazepam in combination with meperidine during the performance of colonoscopy. Adult patients scheduled to undergo colonoscopy were eligible for inclusion. Patients were excluded if they were less than 18 years old or pregnant, were allergic to meperidine or benzodiazepines, were unreliable historians, had a history of medication-induced phlebitis or substance abuse, had laboratory abnormalities, or were otherwise unstable. Patients received meperidine in an initial dose of 0.4 mg/
kg intravenously and either 0.03 mg/kg midazolam intravenously or 0.10 mg/kg diazepam intravenously. It is standard in our endoscopy unit that patients receive both a narcotic analgesic and a benzodiazepine for endoscopic procedures. The benzodiazepine was prepared in 5-ml syringes, such that 1 ml contained either 1 mg of midazolam or 3 mg of diazepam to be administered in equal volumes. All medication was infused over 2 min. A 2-min observation interval was maintained between dosing. Sedation endpoints were drowsiness, facial relaxation, slurred speech, and tolerance to insertion of the colonoscope. If additional pre-procedure sedation was required, or the patient required additional dosing during the procedure, the medications were titrated slowly in the fixed ratio of 1.0 mg of midazolam to 10 mg of meperidine, or 3.0 mg of diazepam to 10 mg of meperidine, as needed. Likewise, medications were infused over 2 min, and patients were observed for at least 2 min between interval dosing. All patients were monitored with automated pulse and blood pressure recordings, ECG rhythm strips, and pulse oximetry. Supplemental oxygen was used to maintain an O2 saturation above 90% throughout the procedure, as is standard in our endoscopy unit. Table 1 provides the schedule of events demonstrating the timing of procedure and post-procedure evaluations. Total procedure time, the initial sedative doses, and the total medication doses were recorded. Six efficacy and safety parameters were analyzed: (1) the difference between the initial calculated and the initial actual dose given by drug, (2) the correlation between the duration of the procedure and drug, (3) procedure-related complications, (4) the endoscopist's assessment of the ease or difficulty of the procedure correlated by drug, (5) the patient's assessment of procedural comfort and post-procedure amnesia correlated by drug, and (6) nursing post-anesthesia recovery scores (PARS, see Appendix). Complications were defined as follows: oxygen desaturation as a sustained finger pulse oximetry reading below 90% (for >30 sec, while being stimulated to take deep breaths); bradycardia as pulse rate less than 40 beats per min or a fall in heart rate by more than 50% of baseline; hypotension as fall in blood pressure of 20% of more of the baseline reading. Phlebitis was evaluated by questioning and observation during the procedure, immediately following and at 120-min post-procedure and by telephone contact 24 hours and 10 days post-procedure. Post-procedure consciousness and cir-
Table 1. Schedule of eventsa Test
Baseb 0 Minb 15 Min 30 Min 60 Min 120 Min 24 Hr 10 Days
Vital signs PARS Amnesic line Comfort line Patient questionnaire Physician questionnaire Complication analysis a Patients were assessed continually during the procedure for the incidence of complications related to the medications given. b Base = time preceding the administration of any sedation medications; 0 min = time following the completion of the colonoscopy.
652
GASTROINTESTINAL ENDOSCOPY
culatory assessment were ascertained during the recovery period by the PARS criteria. A patient questionnaire was constructed that asked about events recalled during the procedure, the degree of abdominal pain experienced, and their willingness to undergo the same procedure in the future. Procedural amnesia and comfort were assessed using a 10-point scale, immediately post-procedure, at 60 min, 120 min, and 24 hours post-procedure. All colonoscopies were performed by the staff gastroenterologists at Georgetown University Hospital using either a Fujinon, Olympus, or Pentax video colonoscope. A complete examination to the cecum was accomplished in all patients. Videophotography, biopsy, and polypectomy were performed as indicated. Immediately following the procedure, the physician was asked to complete a questionnaire assessing the ease of the procedure, the degree of sedation achieved, the degree of pain control during the procedure, and the cooperation of the patient. Overall performance of the procedure was rated as easy, moderate, or difficult. Discharge criteria required that the patient be fully awake, oriented to time and place, able to ambulate, and able to tolerate fluids. All patients were accompanied home by a responsible party and were advised not to drive or operate potentially dangerous equipment until the following day. Approval from the Institutional Review Board was obtained and informed consent was provided by all participants. Statistical analysis of the results was performed using the analysis of variance (ANOVA) test for repeated measures, Student's t test for paired samples, Tukey's test for multiple comparisons, and Fischer's exact test as appropriate. Differences with a probability value of less than 0.05 were considered to be significant. RESULTS Sixty patients were prospectively randomized to receive either midazolam or diazepam. Seven individuals were excluded for protocol violations (two because of errors in drug dose administration and five because of accidental unblinding), leaving 53 patients who completed the study. The two groups were similar in size, gender, age, mean total meperidine dose, and mean procedure duration (Table 2). No difference in the mean predicted (calculated) and mean actual dose of meperidine required for initial Table 2. Characteristics of treatment groups
No. of patients M/F Mean age, yr (range) Mean total dose meperidine, mg (SD) Mean total dose benzodiazepine, mg Mean duration of procedure, min (SD)
VOLUME 38. NO.6, 1992
Midazolam
Diazepam
29 18/11 60.8 (31-95) 48.8 (12.0)
24 15/9 60.8 (27-85) 47.8 (11.5)
4.1 30.9 (9.3)
11.1 28.9 (9.7)
Table 3. Initial calculated, initial actual, and total actual doses of meperidine in the two study groups· Dose of meperidine Group
Initial calculated
Initial actual
Total actual
Diazepam
30.0 (22.8-50.8) 30.2 (20.0-46.0)
29.4 (10.0-50.0) 29.3 (20.0-46.0)
47.8 (29.0-70.0) 48.8 (30.0-76.0)
Midazolam
a Mean doses are given in milligrams (range). No statistical difference seen.
Table 4. Initial calculated, initial actual, and total actual doses of benzodiazepines in the two study groups· Dose Diazepam Midazolam
Initial calculated
Initial actual
Total actual
7.5 (5.7-12.7) 2.2 (1.5-3.5)
6.5 (2.0-12.0) 2.2 (1.5-3.5)
11.1 (5.0-19.0) 4.1 (2.0-9.6)
a Mean doses are given in milligrams (range). The difference between the mean calculated and mean initial actual dose of diazepam (1.15 mg) is significant (p = 0.01, t test based on unequal variances) .
sedation was seen (Table 3). Likewise, the mean total dose of meperidine did not differ between the two groups (Table 3). The mean predicted and mean actual dose of midazolam and diazepam for initial sedation for the entire study cohort also did not differ (Table 4). Using a relative potency coefficient of 3/ 2 the mean total dose potency of midazolam (4.1 X 3 = 12.3) was not significantly different from the mean total dose potency of diazepam (11.1 x 1 = 11.1). In a subgroup offive diazepam recipients (21 %), the actual dose of diazepam needed for initial sedation was significantly lower than the calculated dose (8.1 mg mean calculated versus 3.2 mg mean actual dose received). Age appeared to be the discriminating factor in this subgroup (mean age 70.2 years compared with a mean age of 58.8 years for the 19 remaining individuals and 60.8 years for the entire group). No such subgroup emerged among midazolam recipients. With respect to complications, no significant differences were observed between the two groups (Fischer's exact test, p > 0.6) (Table 5), not an unexpected finding given the small numbers in this study. All endoscopic or drug-related complications were considered minor, as no medical or surgical intervention was required (with the exception of increasing supplemental oxygen in three patients). Bradycardia occurred in three patients in the midazolam group and one in the diazepam group. It was felt to be vasovagally induced in all four instances and resolved with 653
Table 5. Incidence of complications by drug group (%)8 O2 desaturaBradycardia Hypotension Phlebitis tion Diazepam Midazolam a
o
2 (8.3)
1 (4.2)
1 (3.4)
3 (10.3)
2 (8.3)
o
1 (3.4)
Differences not significant, p < 0.6.
Table 6. Physician global assessment of ease of procedure by drug (%)8 Difficult
Moderate
Easy
1 (4) 3 (11)
10 (42) 10 (36)
13 (54) 15 (54)
Diazepam Midazolam a
recipients compared with diazepam recipients at all evaluation intervals, and these differences reached statistical significance at 120 min and 24 hours postprocedure (ANOVA test for repeated measures, Tukey's test for multiple comparisons, p = 0.01) (Fig. 2). No correlation could be drawn between the benzodiazepine used and the length of the procedure. Although the study did not consider technical differences among the two study groups (i.e., number of polypectomies, biopsies, etc.) that may have affected procedure length, and while the mean procedure duration was similar, 24% of colonoscopies in the midazolam group went beyond 40 min compared with only 13% in the diazepam group (Fig. 3). In the patient questionnaire completed at 120-min
Differences not significant.
=~I 10
•
';~rt10
8
-
~
0
6 ...
-
e
~
0
()
'iii Q) c
0
I
2 f-
C:rt 0
_
Diazepam o Midazolam
0
4
60min
Diazepam o Midazolam
o
1.-.-"-_ _- ' -_ _-'--_ _-'---'
o
I-
2
120min
I
I 24hr
Time from procedure completion Figure 1. Mean group comfort scores recorded post-procedure by drug group over time, from repeated analysis of variants. No difference seen by ANOVA-repeated measures, p = 0.48.
reduction of excessive looping of the colonoscope. We attributed the low incidence of phlebitis in both groups to the use of a running intravenous line in a large caliber vein. No differences between the two groups were distinguishable by the physician's assessment (Table 6). The colonoscopists were unable to reliably identify which benzodiazepine had been used for conscious sedation. Total PARS scored (specifically, circulation and conscious scores), as recorded by the observation room staff, did not differ between the two groups (ANOVA test for repeated measures; p = 0.69, p = 0.87). Recovery times were similar based on PARS scores. Patient comfort scores were not significantly different at any time post-procedure (p = 0.60 by ANOVA test for repeated measures) (Fig. 1). Amnesia scores, however, demonstrated less recall among midazolam 654
0
E
Diazepam versus midazolam for colonoscopy: a prospective evaluation of predicted versus actual dosing requirements Gregory G. Ginsberg, MD, James H. Lewis, Jane E. Gallagher, RN, David E. Fleischer, Firas H. AI-Kawas, MD, Cuong C. Nguyen, Diane J. Mundt, PhD, Stanley B. Benjamin,
MD MD MD MD
Washington, D.C.
We performed a prospective, randomized, double-blind study to evaluate the efficacy of the currently recommended low doses of midazolam for conscious sedation compared with diazepam for colonoscopy. Each agent was administered in a fixed ratio dose in combination with meperidine, and titrated incrementally to allow for adequate sedation prior to initiating and during the procedure. The currently recommended starting dose of midazolam (0.03 mg/kg) proved to be very appropriate for pre-medication. In contrast, the currently recommended starting dose of diazepam (0.10 mg/kg) proved excessive in 21% of patients (especially in those aged >65). The low initial and incremental doses of midazolam compared favorably with diazepam in all efficacy parameters studied and exceeded diazepam in post-procedure amnesia scores (p 0.01). Moreover, the sedative effects of midazolam at these lower doses were not lost despite long duration procedures (>40 min). We conclude that midazolam, given in small incremental doses, in combination with meperidine, produces effective conscious sedation for colonoscopy and exceeds diazepam in its amnestic effect. (Gastrointest Endosc 1992;38:651-656)
=
Diazepam and midazolam are the two most widely used benzodiazepines for endoscopic conscious sedation.1,2 Although midazolam has been cited as having a number of potential advantages over diazepam, such as a shorter elimination half-life, a lower incidence of phlebitis, and superior anterograde amnesia,!·3 it remains controversial whether in fact one or the other is optimal for conscious sedation. 4 • 5 Much of the current controversy centers on the issue of safety. Following the introduction of midazolam in 1986, several serious cardiorespiratory events, including some deaths that were potentially related to its use, were reported to the U.S. Food and Drug Administration (FDA).6 These reports raised concerns that the use of midazolam for endoscopic conscious sedation might Received April 20, 1992. For revision June 1, 1992. Accepted July 28,1992. From the Division of Gastroenterology, Georgetown University Medical Center, Washington, D.C. Reprint requests: Stanley B. Benjamin, MD, Division of Gastroenterology, Georgetown University Medical Center, 3800 Reservoir Road, N. w., Washington, D.C. 20007. VOLUME 38, NO.6, 1992
constitute an overly serious risk. 7 Midazolam is three to four times as potent as diazepam 8 and was in many instances being administered in too large a dose. 9 Although midazolam has a more rapid onset of action compared with diazepam,3 recent studies have demonstrated that the peak drug effect occurs later for midazolam than for diazepam. lO This may have resulted in inadequate intervals of observation between repeat dosing of midazolam, thereby contributing to oversedation and medication associated complications. In response to these issues, it was recommended that the initial and incremental doses of midazolam be decreased for conscious sedation, and a physician education campaign was mounted that emphasized a titrated method of administration, taking into account the patient's age, underlying cardiopulmonary diseases, and the concurrent use of other agents during endoscopy. These efforts led to a significant reduction in the frequency of reported adverse cardiopulmonary events associated with midazolam. 9 Arrowsmith et 651
alY have recently compared the rates of serious cardiorespiratory complications and deaths associated with the use of midazolam or diazepam for endoscopic conscious sedation, and found no significant differences among 21,000 procedures from 17 medical centers that were recorded in the American Society for Gastrointestinal Endoscopy computer-based management system. In spite of midazolam's current widespread acceptance and use as an agent for endoscopic conscious sedation, no prospective study has compared diazepam with midazolam at the lower dosing schedules now considered appropriate for conscious sedation. Several issues remain unanswered: are the lower dosing schedules safer? Is the superior anterograde amnesia produced by midazolam at higher doses maintained at lower doses? Do the newer recommended doses of midazolam correlate with the actual doses required to achieve adequate sedation to initiate endoscopic procedures and allow maintenance of sedation throughout the procedure? Finally, there is concern that at the lower dosing regimens, the efficacy of midazolam may be lost during procedures of longer duration, such as colonoscopy, owing to the drug's shorter elimination half-life. 4 In order to answer questions concerning the efficacy of midazolam at lower doses, we undertook a prospective study of the predicted versus actual dosing requirements for midazolam compared with diazepam during colonoscopy. PATIENTS AND METHODS
This study was a randomized, double-blind comparison of the effectiveness of midazolam and diazepam in combination with meperidine during the performance of colonoscopy. Adult patients scheduled to undergo colonoscopy were eligible for inclusion. Patients were excluded if they were less than 18 years old or pregnant, were allergic to meperidine or benzodiazepines, were unreliable historians, had a history of medication-induced phlebitis or substance abuse, had laboratory abnormalities, or were otherwise unstable. Patients received meperidine in an initial dose of 0.4 mg/
kg intravenously and either 0.03 mg/kg midazolam intravenously or 0.10 mg/kg diazepam intravenously. It is standard in our endoscopy unit that patients receive both a narcotic analgesic and a benzodiazepine for endoscopic procedures. The benzodiazepine was prepared in 5-ml syringes, such that 1 ml contained either 1 mg of midazolam or 3 mg of diazepam to be administered in equal volumes. All medication was infused over 2 min. A 2-min observation interval was maintained between dosing. Sedation endpoints were drowsiness, facial relaxation, slurred speech, and tolerance to insertion of the colonoscope. If additional pre-procedure sedation was required, or the patient required additional dosing during the procedure, the medications were titrated slowly in the fixed ratio of 1.0 mg of midazolam to 10 mg of meperidine, or 3.0 mg of diazepam to 10 mg of meperidine, as needed. Likewise, medications were infused over 2 min, and patients were observed for at least 2 min between interval dosing. All patients were monitored with automated pulse and blood pressure recordings, ECG rhythm strips, and pulse oximetry. Supplemental oxygen was used to maintain an O2 saturation above 90% throughout the procedure, as is standard in our endoscopy unit. Table 1 provides the schedule of events demonstrating the timing of procedure and post-procedure evaluations. Total procedure time, the initial sedative doses, and the total medication doses were recorded. Six efficacy and safety parameters were analyzed: (1) the difference between the initial calculated and the initial actual dose given by drug, (2) the correlation between the duration of the procedure and drug, (3) procedure-related complications, (4) the endoscopist's assessment of the ease or difficulty of the procedure correlated by drug, (5) the patient's assessment of procedural comfort and post-procedure amnesia correlated by drug, and (6) nursing post-anesthesia recovery scores (PARS, see Appendix). Complications were defined as follows: oxygen desaturation as a sustained finger pulse oximetry reading below 90% (for >30 sec, while being stimulated to take deep breaths); bradycardia as pulse rate less than 40 beats per min or a fall in heart rate by more than 50% of baseline; hypotension as fall in blood pressure of 20% of more of the baseline reading. Phlebitis was evaluated by questioning and observation during the procedure, immediately following and at 120-min post-procedure and by telephone contact 24 hours and 10 days post-procedure. Post-procedure consciousness and cir-
Table 1. Schedule of eventsa Test
Baseb 0 Minb 15 Min 30 Min 60 Min 120 Min 24 Hr 10 Days
Vital signs PARS Amnesic line Comfort line Patient questionnaire Physician questionnaire Complication analysis a Patients were assessed continually during the procedure for the incidence of complications related to the medications given. b Base = time preceding the administration of any sedation medications; 0 min = time following the completion of the colonoscopy.
652
GASTROINTESTINAL ENDOSCOPY
culatory assessment were ascertained during the recovery period by the PARS criteria. A patient questionnaire was constructed that asked about events recalled during the procedure, the degree of abdominal pain experienced, and their willingness to undergo the same procedure in the future. Procedural amnesia and comfort were assessed using a 10-point scale, immediately post-procedure, at 60 min, 120 min, and 24 hours post-procedure. All colonoscopies were performed by the staff gastroenterologists at Georgetown University Hospital using either a Fujinon, Olympus, or Pentax video colonoscope. A complete examination to the cecum was accomplished in all patients. Videophotography, biopsy, and polypectomy were performed as indicated. Immediately following the procedure, the physician was asked to complete a questionnaire assessing the ease of the procedure, the degree of sedation achieved, the degree of pain control during the procedure, and the cooperation of the patient. Overall performance of the procedure was rated as easy, moderate, or difficult. Discharge criteria required that the patient be fully awake, oriented to time and place, able to ambulate, and able to tolerate fluids. All patients were accompanied home by a responsible party and were advised not to drive or operate potentially dangerous equipment until the following day. Approval from the Institutional Review Board was obtained and informed consent was provided by all participants. Statistical analysis of the results was performed using the analysis of variance (ANOVA) test for repeated measures, Student's t test for paired samples, Tukey's test for multiple comparisons, and Fischer's exact test as appropriate. Differences with a probability value of less than 0.05 were considered to be significant. RESULTS Sixty patients were prospectively randomized to receive either midazolam or diazepam. Seven individuals were excluded for protocol violations (two because of errors in drug dose administration and five because of accidental unblinding), leaving 53 patients who completed the study. The two groups were similar in size, gender, age, mean total meperidine dose, and mean procedure duration (Table 2). No difference in the mean predicted (calculated) and mean actual dose of meperidine required for initial Table 2. Characteristics of treatment groups
No. of patients M/F Mean age, yr (range) Mean total dose meperidine, mg (SD) Mean total dose benzodiazepine, mg Mean duration of procedure, min (SD)
VOLUME 38. NO.6, 1992
Midazolam
Diazepam
29 18/11 60.8 (31-95) 48.8 (12.0)
24 15/9 60.8 (27-85) 47.8 (11.5)
4.1 30.9 (9.3)
11.1 28.9 (9.7)
Table 3. Initial calculated, initial actual, and total actual doses of meperidine in the two study groups· Dose of meperidine Group
Initial calculated
Initial actual
Total actual
Diazepam
30.0 (22.8-50.8) 30.2 (20.0-46.0)
29.4 (10.0-50.0) 29.3 (20.0-46.0)
47.8 (29.0-70.0) 48.8 (30.0-76.0)
Midazolam
a Mean doses are given in milligrams (range). No statistical difference seen.
Table 4. Initial calculated, initial actual, and total actual doses of benzodiazepines in the two study groups· Dose Diazepam Midazolam
Initial calculated
Initial actual
Total actual
7.5 (5.7-12.7) 2.2 (1.5-3.5)
6.5 (2.0-12.0) 2.2 (1.5-3.5)
11.1 (5.0-19.0) 4.1 (2.0-9.6)
a Mean doses are given in milligrams (range). The difference between the mean calculated and mean initial actual dose of diazepam (1.15 mg) is significant (p = 0.01, t test based on unequal variances) .
sedation was seen (Table 3). Likewise, the mean total dose of meperidine did not differ between the two groups (Table 3). The mean predicted and mean actual dose of midazolam and diazepam for initial sedation for the entire study cohort also did not differ (Table 4). Using a relative potency coefficient of 3/ 2 the mean total dose potency of midazolam (4.1 X 3 = 12.3) was not significantly different from the mean total dose potency of diazepam (11.1 x 1 = 11.1). In a subgroup offive diazepam recipients (21 %), the actual dose of diazepam needed for initial sedation was significantly lower than the calculated dose (8.1 mg mean calculated versus 3.2 mg mean actual dose received). Age appeared to be the discriminating factor in this subgroup (mean age 70.2 years compared with a mean age of 58.8 years for the 19 remaining individuals and 60.8 years for the entire group). No such subgroup emerged among midazolam recipients. With respect to complications, no significant differences were observed between the two groups (Fischer's exact test, p > 0.6) (Table 5), not an unexpected finding given the small numbers in this study. All endoscopic or drug-related complications were considered minor, as no medical or surgical intervention was required (with the exception of increasing supplemental oxygen in three patients). Bradycardia occurred in three patients in the midazolam group and one in the diazepam group. It was felt to be vasovagally induced in all four instances and resolved with 653
Table 5. Incidence of complications by drug group (%)8 O2 desaturaBradycardia Hypotension Phlebitis tion Diazepam Midazolam a
o
2 (8.3)
1 (4.2)
1 (3.4)
3 (10.3)
2 (8.3)
o
1 (3.4)
Differences not significant, p < 0.6.
Table 6. Physician global assessment of ease of procedure by drug (%)8 Difficult
Moderate
Easy
1 (4) 3 (11)
10 (42) 10 (36)
13 (54) 15 (54)
Diazepam Midazolam a
recipients compared with diazepam recipients at all evaluation intervals, and these differences reached statistical significance at 120 min and 24 hours postprocedure (ANOVA test for repeated measures, Tukey's test for multiple comparisons, p = 0.01) (Fig. 2). No correlation could be drawn between the benzodiazepine used and the length of the procedure. Although the study did not consider technical differences among the two study groups (i.e., number of polypectomies, biopsies, etc.) that may have affected procedure length, and while the mean procedure duration was similar, 24% of colonoscopies in the midazolam group went beyond 40 min compared with only 13% in the diazepam group (Fig. 3). In the patient questionnaire completed at 120-min
Differences not significant.
=~I 10
•
';~rt10
8
-
~
0
6 ...
-
e
~
0
()
'iii Q) c
0
I
2 f-
C:rt 0
_
Diazepam o Midazolam
0
4
60min
Diazepam o Midazolam
o
1.-.-"-_ _- ' -_ _-'--_ _-'---'
o
I-
2
120min
I
I 24hr
Time from procedure completion Figure 1. Mean group comfort scores recorded post-procedure by drug group over time, from repeated analysis of variants. No difference seen by ANOVA-repeated measures, p = 0.48.
reduction of excessive looping of the colonoscope. We attributed the low incidence of phlebitis in both groups to the use of a running intravenous line in a large caliber vein. No differences between the two groups were distinguishable by the physician's assessment (Table 6). The colonoscopists were unable to reliably identify which benzodiazepine had been used for conscious sedation. Total PARS scored (specifically, circulation and conscious scores), as recorded by the observation room staff, did not differ between the two groups (ANOVA test for repeated measures; p = 0.69, p = 0.87). Recovery times were similar based on PARS scores. Patient comfort scores were not significantly different at any time post-procedure (p = 0.60 by ANOVA test for repeated measures) (Fig. 1). Amnesia scores, however, demonstrated less recall among midazolam 654
0
E
