The use of procedural sedation and analgesia to perform painful orthopaedic interventions in the emergency department has become a standard practice over the last decade as a result of the convenience, accessibility, and cost-effectiveness of performing these procedures in the emergency department rather than in the operating room1. There are many available anesthetic agents suitable for procedural sedation and analgesia. The combination of midazolam and ketamine (referred to in the present study as midazolam/ketamine) is a popular choice and is a safe and effective regimen for the induction of sedation in children1-4. Few studies have investigated the safety profile and effectiveness of this regimen in adults, although the complementary pharmacological properties of the two agents make their combination attractive for routine use in the emergency department5,6. The ultra-short-acting sedative propofol is extensively used in emergency medicine but is still considered by many to be unsafe unless administered by a trained anesthesiologist. Evidence suggests that the propofol safety profile is similar to or better than those of other traditionally used procedural sedation and analgesia regimens, even when it is administered by nonanesthesiologists1,7-11. Propofol offers the advantages of rapid induction and short recovery time that may expedite patient management in the emergency department1,12. However, propofol is associated with a considerable prevalence of respiratory and hemodynamic adverse events, especially in patients older than sixty-five years of age13. In addition to sedation and analgesia, amnesia should also be provided by an effective sedative regime. Recall rates of 16% to 20% have been reported following sedation with propofol14,15, but we are not aware of any published data on recall rates following sedation with midazolam/ketamine.
Our hypothesis was that procedural sedation and analgesia with propofol would save time in a busy emergency department with a similar rate of adverse events as compared with midazolam/ketamine.
The primary purpose of the present study was to evaluate the recovery time and total sedation time associated with procedural sedation and analgesia with use of propofol as compared with midazolam/ketamine in the emergency department. A secondary purpose was to evaluate the safety profile and effectiveness measures of sedation with propofol as compared with midazolam/ketamine.
Study Design and Setting
The present randomized, prospective, single-blind study was conducted in the adult emergency department of a tertiary care, university-affiliated medical center with approximately 130,000 emergency department visits per year. Patients who had been managed in the emergency department were enrolled from November 2008 to September 2009 as a convenience sample based on the availability of the first two authors (O.U., E.B.) during all hours of the day and all days of the week. This study was planned according to the International Conference on Harmonisation guidelines for good clinical practice, was approved by our Institutional Ethics Committee (reference number, 0272-08), and was registered at ClinicalTrials.gov (“Procedural Sedation Using Propofol Versus Midazolam/Ketamine in the Adult Emergency Department,” NCT00784498).
Selection of Participants
All patients with orthopaedic injuries requiring painful manipulation (reduction of a fracture or dislocation, suture of an extensive laceration) were candidates for inclusion in this study.
The decision to perform the procedure with the patient under sedation or with use of another type of analgesia (e.g., local anesthesia, hematoma or intra-articular block, intravenous opiates) was made by the treating orthopaedic surgeon (one of the first two authors). The inclusion criteria were (1) an age of eighteen to sixty-five years, (2) an American Society of Anesthesiologists (ASA) score of 1 or 2, (3) a systolic blood pressure of >90 mm Hg before the initiation of sedation, (4) the willingness and ability to provide informed consent, (5) no known hypersensitivity to any of the study medications, (6) no evidence of intoxication, (7) no solid food two hours before the induction of sedation, and (8) a nonpregnant status (for women).
Study Protocol
Patients who required sedation were informed, by a nonanesthesiologist in the emergency department, about the benefits and risks of undergoing sedation, including the risk of respiratory adverse events and the necessity of emergency intubation. After providing signed informed consent, each patient was randomized to receive either (1) a slow intravenous (IV) push of midazolam (0.1 mg/kg) until spontaneous eye closure (or up to 5 mg) followed by a slow IV push of ketamine (1 mg/kg, up to 100 mg) to achieve the desired level of sedation (a modified Ramsay sedation level of 5 to 611,16) or (2) a titrated slow IV push of propofol in 10 mg/10 second boluses, titrated to achieve the desired sedation level (up to 200 mg), followed by repeat dosing at 10 mg/10 seconds as needed. Randomization was achieved by opening consecutively numbered envelopes containing equal numbers of propofol and midazolam/ketamine assignments (prepared by personnel not directly involved in the study). A peripheral intravenous line was placed in the forearm region, and an infusion of a 0.9% normal saline solution was ready for infusion if indicated. All patients received supplemental oxygen via a 100% nonrebreathing reservoir mask. Heart rate and oxygen saturation were monitored continually by a registered nurse. Heart rate, oxygen saturation, and blood pressure values were documented before the initiation of sedation, every two minutes from the induction of sedation to the completion of the procedure, and then every five minutes until the return to baseline mental status. The return to baseline mental status was determined by an emergency department registered nurse, who was trained in the assessment of the mental status of sedated patients and was unaware of the patient assignment group. Return to baseline mental status was defined as the patient being fully awake and able to converse coherently and having returned to his or her pre-sedation ambulatory status. The patient remained under observation in the emergency department for at least one hour after the return to baseline mental status.
Three 10 × 10-cm pictures were presented to the patient to assess recall and amnesia. The first picture (an image of a cat) was shown just before induction. The second picture (an image of a dog) was shown during recovery to a modified Ramsay sedation level 3. The last picture (an image of a mouse) was shown on return to baseline mental status. The patient was asked to verbally identify the presented image.
All sedations and manipulations were performed by one of two senior orthopaedic residents experienced in performing the orthopaedic interventions. These residents had been formally authorized and certified by our institutional management to perform sedation in adults with either propofol or midazolam/ketamine and were capable of performing emergency intubation. Their training included (1) satisfactory completion of Advanced Trauma Life Support (ATLS) and Advanced Cardiac Life Support (ACLS) courses; (2) a three-month rotation in the intensive care unit, during which the residents performed many intubations; (3) a special institutional instructional course on performing sedation; and (4) inductions of sedation in the emergency department under the direct supervision of a trained anesthesiologist. A crash cart and an intubation kit were always available in the sedation room according to the study protocol and our emergency department protocol. In addition, trained emergency physicians and anesthesiologists were always available within a response time of one to three minutes.
Outcome Measures
Recovery time was defined as the time interval from the completion of the orthopaedic manipulation procedure to the return of the pre-sedation mental status. Total sedation time was defined as the time interval from the induction of sedation to the return to the pre-sedation mental status. The safety profile of each regimen was evaluated on the basis of the prevalence of respiratory and hemodynamic adverse events. A registered nurse monitored the patient's vital signs during the procedure and recovery. Absence of breathing was assessed clinically by the registered nurse and was measured as accurately as possible. Minor adverse events were defined as (1) apnea (i.e., the absence of breathing for twenty seconds or more) that responded to airway repositioning (e.g., chin lift, jaw thrust), (2) hypoxemia (i.e., an O2Sat of <90% for sixty seconds or more as measured with pulse oximetry) that resolved spontaneously or with airway repositioning, (3) hypotension (i.e., systolic blood pressure of <90 mm Hg) that resolved spontaneously or with intravenous fluid administration, and (4) bradycardia (i.e., heart rate of <50 bpm for sixty seconds or more) that resolved spontaneously. Major adverse events included (1) apnea that required oral airway insertion, bag valve mask ventilation, or tracheal intubation; (2) hypoxemia that required oral airway insertion or intubation; (3) hypotension not responding to IV fluid administration within two minutes; (4) bradycardia lasting more than two minutes; and (5) seizures. All other adverse events were recorded as well but were not considered to be safety issues (e.g., agitation, euphoria). All clinical decisions and treatment for adverse events were made by one of the first two authors (O.U., E.B), who performed all sedations and manipulations.
The effectiveness of the procedural sedation and analgesia was assessed on the basis of (1) the procedure success rate (i.e., whether or not the goal of the procedure had been achieved), (2) physician's comfort in performing the procedure as assessed with a 100-mm visual analog scale (VAS), (3) observed behavioral distress on the part of the patient as assessed by a registered nurse who was unaware of the patient assignment group with use of a 100-mm VAS score, and (4) satisfaction with the orthopaedic treatment in the emergency department as assessed by the patient with use of a 100-mm VAS after returning to the baseline mental status. The specific patient-satisfaction question was “On a 0 to 100 scale, how much are you satisfied with the orthopaedic treatment you received?”
To assess the recall rate, patients were asked, on return to the baseline mental status, to describe what they remembered about the procedure and the events prior to it and to describe the three pictures that had been presented to them.
Data Analysis
Continuous parameters were described as the mean and the standard deviation (SD) with 95% confidence intervals (CI). Categorical parameters were described with proportions and 95% confidence intervals. Comparisons between the groups were performed with a two-tailed t test for continuous parameters and the Fisher exact test for categorical parameters. Statistical analysis was performed with use of SPSS for Windows (version 16.0; IBM, Chicago, Illinois). With the alpha level set at 0.05, it was determined prospectively that twenty-four patients per group would give 90% power to identify a ten-minute difference in recovery and total sedation time.
Source of Funding
No external funding was received for this study.
Demographic Characteristics
Two hundred and thirty adults underwent painful orthopaedic manipulations that were performed by one of two authors (O.U., E.B.) in our emergency department from August 2008 to September 2009. Sixty-four patients were judged to require procedural sedation and analgesia and were randomly enrolled in the study. The indication for sedation in more than half of the patients was for the reduction of a shoulder dislocation. Four patients were excluded from further analysis because of missing data, leaving sixty patients for the study (thirty in each group) (Fig. 1). The mean age (and standard deviation) of the patients in the study group was 45 ± 17 years (range, twenty-two to sixty-five years), and 58% of the patients were male. The demographic and clinical data for each group are summarized in Table I.
Dosing
The mean induction dose was 1.5 ± 0.4 mg/kg for propofol, 0.05 ± 0.02 mg/kg for midazolam, and 1.0 ± 0.3 mg/kg for ketamine. Ten patients (33.3%) in the propofol group required a top-up dose for maintenance of adequate sedation, compared with no patients in the midazolam/ketamine group (difference, 33.3% [95% CI, 15% to 51%]; p < 0.001). These ten patients received a single top-up dose of propofol that averaged 0.7 ± 0.3 mg/kg within a mean of 4.4 ± 1.1 minutes after the induction of sedation. Four patients in the midazolam/ketamine group had received intravenous morphine at an average dose of 8.8 mg (approximately 0.1 mg/kg) approximately thirty minutes prior to the induction of procedural sedation and analgesia, which did not appear to influence the outcome measures. No patient in the propofol group received opioid analgesics.
Sedation Time
The mean recovery time was significantly shorter in the propofol group as compared with the midazolam/ketamine group (7.8 ± 3.7 minutes compared with 30.7 ± 10.1 minutes, representing a difference of 22.9 minutes [95% CI, 18.9 to 26.8 minutes]; p < 0.001), as was the total sedation time (16.2 ± 3.8 minutes compared with 41.6 ± 10.7 minutes, representing a difference of 25.4 minutes [95% CI, 21.2 to 29.5 minutes]; p < 0.001). The mean induction time (i.e., the induction of sedation to the modified Ramsay sedation level of 5 to 6) was also significantly shorter in the propofol group as compared with the midazolam/ketamine group (3.2 ± 1.1 minutes compared with 4.5 ± 1.3 minutes, representing a difference of 1.3 minutes [95% CI, 0.7 to 1.9 minutes]; p < 0.001). Sedation time values are summarized in Table II.
Adverse Events
Respiratory and hemodynamic adverse events were observed in six (20%) of the thirty patients in the propofol group and in three (10%) of the thirty patients in the midazolam/ketamine group. Specifically, four patients in the propofol group exhibited a brief apnea of less than one minute that responded to airway repositioning, and two patients in the propofol group had temporary hypotension (a drop in the systolic blood pressure to 80 mm Hg) that occurred two minutes following the induction of sedation and resolved spontaneously. Two patients in the midazolam/ketamine group had short apnea of less than one minute that responded to airway repositioning, and a third patient had more serious apnea that necessitated an oral airway insertion and a brief (less-than-one-minute) bag valve mask ventilation. There were no cases of hypoxemia (O2Sat <90%), and no patient required endotracheal intubation (Table III). Because the study was not powered to examine differences in the rates of adverse events, we were not able to determine whether this difference was significant.
Both systolic and diastolic blood pressure dropped significantly following propofol induction. Systolic blood pressure dropped from a pre-sedation level of 147.2 ± 25.1 mm Hg to a desired sedation level of 132.8 ± 22.8 mm Hg, representing a difference of 14.4 mm Hg (95% CI, 11.8 to 16.9 mm Hg; p = 0.023), and diastolic blood pressure dropped from a pre-sedation level of 87.1 ± 13.3 mm Hg to a desired sedation level of 79.6 ± 12.3 mm Hg, representing a difference of 7.5 mm Hg (95% CI, 5.6 to 9.3 mm Hg; p = 0.027). The lowest level of blood pressure (both systolic and diastolic) following propofol administration was observed on attaining the desired level of sedation at a mean of 3.2 ± 1.1 minutes after induction. No significant drop in blood pressure was observed following midazolam/ketamine administration (see Appendix).
Effectiveness Measures and Procedure Duration Effect
Effectiveness outcomes were similar for both groups (Table IV). However, the duration of the procedure seemed to influence effectiveness when sedation was induced with propofol. Specifically, eleven patients (37%) in the propofol group and twelve patients (40%) in the midazolam/ketamine group (representing a difference of 3% [95% CI, −20% to 26%]; p = 0.99) underwent procedures that lasted for more than five minutes. Of these patients, ten (91%) of eleven in the propofol group required an additional dose for adequate sedation compared with zero of twelve patients in the midazolam/ketamine group (representing a difference of 91% [95% CI, 53% to 98%]; p = 0.005). These were the only patients in this study who required an additional dose for the maintenance of sedation. The physician satisfaction score averaged 80.5 ± 21.3 mm for the procedures in the propofol group that lasted more than five minutes, compared with 96.6 ± 9.6 mm for those that were completed within five minutes (representing a difference of 16.1 mm [95% CI, 14.1 to 18.1 mm]; p < 0.001). Inversely, the observed behavioral distress score was 39.5 ± 17.4 mm during those relatively longer procedures, compared with 20.5 ± 22.2 mm during procedures completed within five minutes (representing a difference of 19.0 mm [95% CI, 16.7 to 21.3 mm]; p < 0.001). The duration of the procedure did not influence the effectiveness outcomes in the midazolam/ketamine group.
Amnesia
The patient immediate recall rate of the events surrounding the procedure after the patient return to baseline mental status was similar for both groups. Four of the thirty patients in the midazolam/ketamine group and three of the thirty patients in the propofol group reported some recall of the procedure (a difference of 3.3% [95% CI, −14% to 21%]; p = 0.99), mostly describing nonspecific sounds and sensations of touch and pain. Twenty-five of the thirty patients in the midazolam/ketamine group demonstrated short-term anterograde amnesia during recovery (i.e., an inability to recall the object in the picture that had been presented on recovery to a modified Ramsay stage 3), compared with only three of the thirty patients in the propofol group (representing a difference of 73% [95% CI, 50% to 84%]; p < 0.001). All patients could recall which specific picture had been presented to them on returning to baseline mental status. All patients had clear recall of the management in the emergency department prior to the procedural sedation and analgesia, and all of them recalled the picture that had been presented to them just prior to the induction of sedation.
In the present randomized prospective trial, the sedation time and the rate of adverse events associated with the use of propofol were compared with those associated with the use of midazolam/ketamine in the emergency department. Our findings demonstrated that the recovery time and total sedation time were significantly shorter following sedation with propofol as compared with sedation with midazolam/ketamine. The approximately twenty-three-minute difference in recovery time and the twenty-five-minute difference in total sedation time were meaningful and supported our hypothesis that sedation with propofol can expedite patient management in the emergency department when sedation is required. Our results were similar to those of Miner et al., who recently reported a significantly shorter recovery time following sedation with propofol17. In that study, however, propofol was compared with ketamine (as a single drug), which has been associated with a high rate of recovery agitation (“emergence phenomena”) in adults, a factor that has made this agent less favored in the adult population1,5,18. Godambe et al. reported on a recovery time of fifty-four minutes following sedation with midazolam/ketamine (0.04 and 1.99 mg/kg, respectively) and of twenty minutes following sedation with propofol/fentanyl (4.55 mg/kg and 1.21 mcg/kg, respectively) in children19. The recovery time in our study was shorter, perhaps because of the administration of lower sedative doses. We used an induction protocol of 10 mg of propofol every ten seconds until adequate sedation was achieved. We preferred to avoid larger intravenous boluses because smaller and slower boluses of propofol may reduce adverse events13. The mean initial dose of propofol in our study (1.5 mg/kg) was dictated by patient response during drug titration and was similar to doses administered in previous studies20. A protocol of close titration of sedatives should be followed to achieve the desired level of sedation with the minimum required sedative dose. The orthopaedic procedure was successful in all but two patients in the propofol group who sustained periprosthetic dislocations (one after total hip arthroplasty and one after total knee arthroplasty) that could not be reduced by closed means. Adequate sedation was achieved in all patients.
Propofol tends to redistribute from the blood within three to five minutes with a rapidly resolving clinical effect8. This may explain the fact that ten of the eleven patients in the propofol group who had a procedure lasting for longer than five minutes required a top-up dose for adequate sedation within 4.4 minutes after induction. Our intervention protocol did not define a specific timing for the administration of top-up doses of propofol, and determining the necessity of providing it was based on the patient response to pain, indicative of a transition to a lower level of sedation. This suboptimal level of sedation resulted in a significantly higher score of observed patient behavioral distress. The necessity of administering sedatives while performing a complex orthopaedic procedure by the same physician resulted in a significantly lower physician satisfaction score with the procedure. The duration of the procedure did not influence these measures in the midazolam/ketamine group, probably because of the relatively longer activity of this regimen, which provides adequate sedation following the induction dose, even for longer procedures. The administration of top-up doses every three to four minutes following induction, as suggested by Miner et al.17 and supported by our findings, may help to ensure adequate sedation. However, this does not alter the complexity of a single physician simultaneously providing sedation and performing procedures.
The American College of Emergency Physicians sedation policy states that the literature does not provide clear evidence on the number of personnel necessary to safely provide procedural sedation and analgesia21. Rex et al.9, in a study of 36,743 cases of propofol sedation for endoscopic procedures in which sedation was administered by a two-person team (a registered nurse supervised only by the endoscopist), reported that no patient required endotracheal intubation or had serious sequelae. Similarly, in our study, all sedations and manipulations were performed by a two-person team (an orthopaedic resident who performed both sedation and manipulation and a registered nurse who monitored the patient), with no patient requiring emergency intubation. In our study, patients under sedation were continuously monitored by the registered nurse and in any case of an adverse event the orthopaedic resident in charge was experienced enough to perform emergency intubation or resuscitation.
The orthopaedic manipulation had to be abandoned to treat adverse events when encountered. Maintaining sedation with propofol, performing a complex orthopaedic manipulation, and treating adverse events simultaneously by a single physician in this study resulted in considerable inconvenience that negatively affected the physician satisfaction score and the observed behavioral distress score. Nevertheless, this inconvenience did not adversely affect patient safety. Performing procedural sedation in the emergency department by a team of three individuals, including a physician who administers the sedation, a registered nurse who monitors the physiological state of the patient, and a physician who performs the procedure, may address the technical inconvenience encountered in this study and improve patient care.
The overall rate of respiratory adverse events in our study was 20% in the propofol group and 10% in the midazolam/ketamine group. Because the study was not powered to examine differences in the rates of adverse events, we were not able to determine the significance of this difference, and larger studies are required for further examination of the safety of these interventions. All but one of these adverse events (in the midazolam/ketamine group) were mild and resolved spontaneously, suggesting that the benefits may still outweigh the risks. This relatively low prevalence of serious adverse events is encouraging and suggests that procedural sedation with use of either a propofol or a midazolam/ketamine regimen can be safely implemented by specifically trained nonanesthesiologists. Nevertheless, it must be emphasized that the therapeutic index is narrower for propofol than it is for midazolam/ketamine because of the potential risk of serious respiratory adverse events. Therefore, propofol should only be administered by those who are capable of performing an emergency endotracheal intubation if needed, and a crash cart and intubation kit should be available when propofol is used.
All patients in our study maintained normal oxygen saturation levels throughout the procedure, including those who had overt respiratory depression. This was not surprising considering the fact that all of them received supplemental oxygen via a 100% nonrebreathing reservoir mask. Substantial evidence has suggested that the routine use of supplemental oxygen during sedation delays oxygen desaturation by several minutes when respiratory depression occurs, thus paradoxically rendering blood oximetry insensitive for monitoring ventilatory depression7,22,23. End-tidal carbon dioxide (CO2) monitoring has been recommended as a more accurate and sensitive device for the early detection of respiratory depression6,24,25. Miner et al., with use of end-tidal CO2 monitoring, detected subclinical respiratory depression in twenty (40%) of fifty patients and thirty (63%) of forty-seven patients during sedation with propofol and ketamine, respectively17. We are not aware of the meaning of this subclinical respiratory depression.
We found a significant but clinically inconsequential decrease in the rates of systolic blood pressure and diastolic blood pressure following the administration of propofol, and no patient required treatment for hypotension.
Short-lived anterograde amnesia was more common in the midazolam/ketamine group, probably because of the amnesic properties of midazolam. However, the immediate recall rate was similar for both groups (three of thirty in the propofol group and four of thirty in the midazolam/ketamine group). We did not find any signs of retrograde amnesia among any of our patients, a finding consistent with other reports14,15.
Our study had several limitations. First, the study physicians were aware of the patient assignment group. The relatively low doses of midazolam that were administered in this study possibly resulted from their familiarity with the properties of midazolam and their attempts to avoid prolonging time to recovery from anesthesia. However, the fact that midazolam was closely titrated according to the patient clinical response would make this less likely. Second, the administration of supplemental oxygen to all of our patients precluded the early detection of respiratory depression with pulse oximetry. Thus, respiratory adverse events in this study were detected by means of clinical evaluation alone, which may have resulted in an underestimation of respiratory adverse events.
In conclusion, the use of propofol for a required procedural sedation in the emergency department expedites patient management and saves time in comparison with the use of midazolam/ketamine. Both these regimens can be safely implemented by specially trained and approved nonanesthesiologists.