Procedural sedation and analgesia in a Canadian adult tertiary care emergency department: a case series

EM Advances

CJEM 2006;8(2):85-93

Abstract

Résumé

Introduction

Procedural sedation and analgesia (PSA) in the emergency department (ED) has allowed patients to undergo unpleasant or painful procedures that, in earlier times, would have entailed a general anesthetic in the operating room, or would have been performed with inadequate analgesia, for fear of uncontrolled cardiac or respiratory depression, or airway compromise.^1-3 In North America, PSA agents were traditionally comprised of a combination of benzodiazepine sedatives and opioid analgesics. Prolonged onset and recovery times⁴ delayed respiratory depression,⁵ and the variable efficacy of these agents promoted the search for better options.^3,6,7 The use of propofol, ketamine and etomidate have become commonplace in the ED as a result.^2,8-13 The literature has supported this change for patients of all ages,^12-14 and even for those with significant pre-existing disease.^15,16 Evidence supports the use of such agents even in settings outside the ED or operating room by supervised nurses or by physicians without formal advanced airway or cardiac life support training.^4,16-20 Despite this, the risk of inadvertent deep sedation or general anesthesia, with the attendant respiratory and hemodynamic depression^3,21 has raised concerns about the safety of these drugs. Opinions as to their appropriateness in the ED vary in the literature.^1,22,23

Objections by anesthesiologists have resulted in restrictions of drugs like propofol to the operating room in some hospitals.^11,22 These objections are often based on the relative paucity of published data when compared with that accrued in the operating room.²² In their 2002 "Practice guidelines for sedation and analgesia by non-anesthesiologists," the American Society of Anesthesiologists (ASA) place emergency physicians under the same umbrella as dentists, radiologists and gastroenterologists.²⁴ They recommend that non-anesthesiologists providing PSA should understand the pharmacology of the agents used, that an individual with advanced life skills be immediately available, and that resuscitation equipment should be present during PSA.²⁴ A call has been made for further data to support the use of potent sedative agents in the ED.¹⁹ A 1999 consensus statement by the Canadian Association of Emergency Physicians summarized the use of PSA in the ED and made clear recommendations regarding monitoring, administration and dosage of agents, and discharge instructions. This paper is still extremely useful for ED PSA today.¹

At the Queen Elizabeth II Health Sciences Centre in Halifax, NS, a regional tertiary care referral centre with approximately 70 000 adult (>15 years of age) ED visits per year, approximately 80 patients receive PSA every month. Drug administration and patient monitoring is conducted by advanced level paramedics (Advanced Care Paramedics [ACPs]) trained in PSA, under the supervision of an emergency physician. The ACPs receive both specific didactic and on-the-job training in the use of the medication options for PSA before assuming the role of PSA facilitator. Didactic education covers the aims, objectives and dangers of PSA, the mechanisms and potential adverse effects of PSA drugs, and criteria for the selection of appropriate patients for PSA (ASA Class I and II²⁴). ACPs are required to attend a specific advanced airway management course with particular emphasis on airway assessment, basic airway manoeuvres, bag-valve-mask ventilation and laryngoscopy. On-the-job training includes accompanying a Primary Care Paramedic with PSA experience for 4-8 shifts with graduated responsibility for drug administration and monitoring. PSAs are conducted according to a specific ED PSA protocol (Appendix 1). They administer medications only on the order of a physician instructed in the use of PSA who is at the bedside. In August 2003, a dedicated PSA patient care record was introduced to document the process of each PSA conducted in the department (Appendix 2).

The purpose of this paper was to collect data prospectively in order to ascertain the frequency of recorded adverse events associated with PSA, and to review our current practice with regard to PSA drug choice in our institution.

Methods

We performed a chart review of all PSA records between Aug. 1, 2004, and July 3, 2005. Data gathered prospectively, on the standardized PSA form (Appendix 2) included:

• evaluation of the patient's suitability for ED PSA;
• indication(s) for PSA;
• medications and doses used;
• level of consciousness; and
• vital signs every 5 minutes during PSA.

Data were transcribed from the forms directly into MS Access 2000. This study was approved by the institutional research ethics board.

The a priori definition of an adverse event included: oxygen saturation (SaO 2 ) of <90% at any time during the procedure in any patient with a baseline SaO 2 of ≥95%; systolic blood pressure (SBP) of <85 mm Hg in any patient with a baseline (pre-procedure) systolic blood pressure of 100 mm Hg or greater; evidence of aspiration; endotracheal intubation; or death. Descriptive statistical analysis of the data were performed to determine if we could assess whether adverse events were more common with either of the 2 most commonly used medication combinations. The null hypothesis was that there is no statistical difference.

Results

PSA was carried out on 979 patients during the study period; 481 (49.2%) were recorded as women, 484 (49.4%) as men. In 14 patient records (1.4%), the age and gender of the patient was not entered on the chart (these patients were excluded from any calculations involving gender or age). Two hundred and ten (21.8%) of the patients were >65 years of age. Indications for PSA were classified as "orthopedic" in 786 (80.3%), followed by "incision and drainage" in 61 (6.2%) (Table 1).

Adverse events

Adverse events were documented in 17 patients: 9/900 (1%) with oxygen desaturation and 8/969 (0.8%) with hypotension. One case of emergence agitation was recorded in a patient receiving ketamine. No pressor agents were prescribed during PSA, and no cases of aspiration, endotracheal intubation or death were recorded.

The lowest SaO 2 recorded during PSA was 64%, occurring in a 71-year-old, 75-kg woman (with baseline SaO 2 of 74% due to severe chronic pulmonary disease) who was undergoing reduction of a right shoulder dislocation and proximal humeral fracture, using fentanyl alone. The lowest SBP recorded was 42 mm Hg recorded in a 82-year-old, 55-kg woman with a baseline SBP of 96 mm Hg who was undergoing emergency cardioversion under etomidate and fentanyl.

Patterns of medication use

Fentanyl was used in 94.0% of cases (marginal error = 1.49%). Propofol and midazolam were the most frequently used sedative agents, at 61.2% and 42.5% use, respectively. Ketamine was used in only 26 cases. Nine hundred and twelve (93.2%) patients received 2 or more drugs, and 90 (9.2%) received 3 or more. The commonest drug combinations were propofol (P) and fentanyl (F) (i.e., P/F) in 487 (49.7%) and midazolam (M) and fentanyl (F) (i.e., M/F) in 324 (33.1%.) In 71 cases (7.3%), fentanyl was used in combination with both midazolam and propofol (F/M/P). When ketamine (K) was used, it was the sole agent in 11 cases, and was combined with a benzodiazepine in 7 cases: with propofol in 3 (K/P), with fentanyl in 2 (K/F), with both midazolam and fentanyl in 2 (K/M/F), and midazolam and propofol (K/M/P) in 1. Etomidate was only used in 1 case, in combination with fentanyl. The breakdown of medication combinations is shown in Table 2. Other medications recorded as being administered during PSA were, lorazepam (1 patient), morphine (5) diphenhydrinate (7), metoclopramide (1) and succinylcholine (1). The administration of succinylcholine was done by a staff anesthesiologist for a difficult reduction by an orthopedic resident. The ACP was present to document the procedure and assist with the monitoring.

Association between medications and adverse events or sub-optimal SaO 2 or systolic blood pressure

The medications associated with oxygen desaturation in patients with an initial SaO 2 of ≥95% were P/F in 5, M/F in 2, and one each of M/K and fentanyl. There was no significant difference (p = 0.7422) in the incidence of desaturation between the P/F (1.0%; confidence interval [CI] 0.1%-1.9%) and M/F (0.6%; CI 0.0%-1.5%) treated groups. This study had 80% power to detect an absolute risk difference of 2.6% between the 2 groups.

In 79 patients (8.1% of the total), in whom the baseline SaO 2 was less than 95%, medications chosen were M/F (38), P/F (25), M/P/F (8) and 8 other regimens (including fentanyl alone in 4, and F/K/P in 2, and one each of propofol and midazolam alone).

M/F was used in the 6 patients who had baseline SaO 2 <90%. Of these 6 patients, 3 had only M/F and 3 had M/F in combination with other medications.

Thirteen (1.3%) patients had an SBP of 85 mm Hg or less recorded during the procedure, of whom 5 started with a SBP of <100 mm Hg. Of these, M/F was used in 2, M/P/F in 1, P/F in 1, and fentanyl alone in 1. Of the 8 who became hypotensive during the procedure, M/F and P/F were each used in 4.

The medications used for the 10 patients (1.0%) with a baseline SBP of <100 mm Hg were P/F in 4, M/F in 5, ketamine in 1, and M/P/F in 1.

Additional associations with medication choice

Interestingly, P/F was used more frequently in male patients (58.7% [284/484]) than in female patients (41.6% [200/481]). M/F was used in 41% of women but used in only 25.8% of men (125/484). Both of these differences were statistically significant (p < 0.001).

Medication preferences were also found to vary with age. In patients >65 years of age, M/F was used in 56.7% (119/210) compared with P/F use of 25.2% (53/210). This difference was statistically significant (p < 0.0001).

Discussion

This paper is the largest published series describing PSA use in a Canadian adult ED setting.¹⁹ Our data show that midazolam, propofol and fentanyl are frequently used for PSA and are associated with few adverse effects when used by well trained practitioners following a standardized protocol. We were unable to demonstrate the difference in side-effect profiles of the different agents demonstrated in a pediatric series.²⁵ Surprisingly, the use of ketamine, which has some advantages with regard to lower rates of respiratory depression or hypotension,^11,26,27 and from which the hazards of psychic emergence reactions have been exaggerated,²⁸ was uncommon in this case series. We reported 1 case in which succinylcholine was used by an anesthesiologist in the ED because it was entered in to our database. The use of this drug as part of ED PSA is not "standard of care," and we do not support the use of succinylcholine in the ED for any objective, apart from rapid sequence intubation.

We were interested to find an association between medication selection by the physician and the age and sex of the patient. This likely reveals a sense of greater comfort with more "traditional" PSA drugs and patients who might, consciously or subconsciously, be considered more vulnerable. The numbers of adverse events and patients with suboptimal baseline SaO 2 and SBP were too low to demonstrate an association between these features and drug choice.

In publishing these results, we are not advocating the indiscriminate use of potent sedatives in the ED. PSA utilization in our institution is facilitated by ACPs with special training and expertise in the use of the drugs under the supervision of and in partnership with a dedicated emergency physician or resident. In addition, PSAs are performed according to a strict protocol. All PSAs in our ED are conducted after an appropriate physical (including a systematic airway) assessment of the patient, and with resuscitation and advanced airway management equipment ready at the bedside. The paramedic's principal responsibility is to administer medication and monitor the patient. He or she is not involved in performing the actual procedure. Facilitating PSA is one of the many ED duties of the ACP (their duties include assisting with resuscitations, casting and obtaining difficult venous access); PSA facilitation has, however, become their primary responsibility. Most of the paramedics, therefore, had gained significant experience in the use of these medications before the time studied. In spite of our low adverse event rate, the reported number of PSAs with the use of these drugs is still relatively small, so caution is always appropriate when introducing unfamiliar drugs into the ED. In this paper we have defined and measured the use of PSA using paramedic-physicians and have demonstrated safe use of this essential clinical intervention in a tertiary care ED. PSA remains a key skill for all emergency physicians, and its use demands a safe and clear protocol with close monitoring.

Limitations

Although the data were recorded prospectively on a standardized form, this study still suffers from many of the limitations of retrospective chart audits. In addition to this, the review illustrated a number of limitations of the PSA record used in our ED. The intended depth of sedation was not recorded, so we were unable to ascertain the degree of "overshoot" that occurred. Although the record included check boxes to gather data regarding the patient's suitability for ED PSA (namely ASA grade, and risk factors for basic and advanced airway management), these were used in an inconsistent fashion, making these particular data unreliable. Specific complications that might be feared during PSA, such as death, endotracheal intubation, pulmonary aspiration, or hospital admission as a result of an event attributable to the PSA, were not specifically solicited, and for these we had to rely on a free text area on the form. Although we doubt that the ACP would neglect to enter a significant adverse event on the form, and note that none of the investigators (who are aware of the vast majority of ED mishaps) heard of any significant complication, it is possible that such a complication may not have been recorded. A thorough assessment of these factors is considered mandatory, and our patient care record has been revised to capture this more reliably.

Because end-tidal carbon dioxide monitoring is not used routinely for PSA in our ED, subclinical respiratory depression may have been missed. In an effort to identify all such cases, we used an SaO 2 of ≤90% at any stage during the procedure as indication of respiratory depression, while other authors used this saturation level for at least 30 seconds as a definition of desaturation.²⁹ Soto and colleagues³⁰ showed that apnea of >20 seconds occurred in 26% of the PSAs performed in their study, all of which were diagnosed by capnography. None were detected by the provider monitoring the patient.³⁰ While recognizing apnea is important, reacting to very brief periods identified with capnography could induce caregivers to use positive pressure ventilation before oxygen desaturation, which could actually increase the danger of aspiration. If brief apnea was missed in some of our patients, the universally good outcome in our series suggests that detecting such apnea is probably unnecessary.²³

The post-hoc power calculation demonstrated that we had a chance of detecting an absolute risk difference of 0.026 (2.6%) in desaturation rates between groups treated with P/F and M/F. However, this should be interpreted with caution; patients were not randomly assigned to their treatment groups, so it is difficult to conclude that medications used in the study would have equivalent safety in all circumstances .

Finally, the assistance of ACPs in the ED is still atypical in Canada, and our results might not mirror those that would be expected in a busy ED where a single physician is expected to administer medication, perform the procedure and monitor the patient. The findings may also not apply to PSA facilitated by paramedics without specific training in ED PSA and advanced airway management.

Conclusion

PSA is being conducted safely in our tertiary care ED. Midazolam and propofol, both in combination with fentanyl are frequently used for PSA, without significant adverse effects. Further registry-based research will tell us more about preferences of PSA agents by physicians of different disciplines, and the incidence of levels of sedation deeper than originally intended. PSA remains a key skill for all ED MDs, and its use demands a safe and clear protocol with close monitoring.