Intravenous sedation for cardiac procedures can be administered safely and cost-effectively by non-anesthesia personnel
- First Online:
- Cite this article as:
- Kezerashvili, A., Fisher, J.D., DeLaney, J. et al. J Interv Card Electrophysiol (2008) 21: 43. doi:10.1007/s10840-007-9191-0
- 151 Views
Primary: to determine the safety and efficacy of intravenous sedation for cardiac procedures administered by non-anesthesia personnel. Secondary: to assess cost effectiveness of such sedation.
Anesthesiologists trained non-anesthesia personnel, and established our sedation protocol, which was then used in 9,558 patients who had cardiac procedures with sedation by non-anesthesia personnel, recorded on a computerized database. Most sedation used was midazolam (MID) and morphine (MOR). Complications and problems were derived from the database and quality assurance committee records. Doses were based on desired level of sedation and procedure duration; highest dose used: MID 78 mg, MOR 84 mg.
Data included catheterization (n = 3,819) and transesophageal echo procedures (n = 260); and overall electrophysiology (n = 5,479) and selected subsets. There were complications or problems in only 9 patients (0.1%), a strong safety statement. There were 3 deaths in electrophysiology related procedures, 2 deaths in catheterization related procedures, all in very sick patients and not definitely related to sedation; 4 others developed clinical instability (hives, hypotension and heart failure—all with no sequellae), 2 of which needed reversal medications. Three patients (<0.03%) proved difficult to sedate, and their procedures were completed with help from the anesthesia department; by protocol this was not a complication. A total of $5,365,691 was saved during the last decade on cardiac procedures performed with conscious sedation.
Non-anesthesia personnel can administer intravenous sedation for cardiac procedures in cardiac settings, with safety and cost-effectiveness demonstrated over many years. Anesthesia services are still appropriate for selected cases.
KeywordsConscious sedationIntravenous sedationCardiac proceduresCardioversionCardiology
Until the early 1990s, cardiac procedures were performed with little or no sedation unless an anesthetist or anesthesiologist was present. This situation changed in much of the United States because of two simultaneous developments. First, anesthesia services underwent a period of contraction due to manpower needs assessments and altered reimbursement policies. Second, several cardiac procedures expanded in frequency, duration, or potential patient discomfort, driving a demand for satisfactory sedation in the absence of anesthesia services. At our institution, guidelines for patient monitoring, drug administration and protocols for dealing with potential complications were developed in close collaboration with the Department of Anesthesiology in accordance with published standards of practice [1–3]. Initial reports on relatively small numbers of patients have indicated that such guidelines may be effective [4, 5]. Nevertheless, some cardiologists or institutions have been reluctant to decrease their dependence on anesthesia services. This report describes the outcome of a standardized sedation protocol involving nearly 10,000 cardiac procedures to determine the safety, efficacy, and cost-effectiveness of intravenous sedation by non-anesthesia personnel.
This is a consecutive series of 9,558 patients, (with exclusions as described below), undergoing both elective and urgent cardiac procedures in a large, unselected patient population whose data were entered into the computerized data base.
Exclusions Patients routinely scheduled for general anesthesia including lead extractions, most atrial fibrillation ablations (at the discretion of the operator), children under 15 years of age undergoing electrophysiology procedures, patients with conditions such as severe obstructive sleep apnea (OSA), and those with known intolerance to sedation. In the case of sleep apnea, continuous pulmonary airway pressure (CPAP) is not permitted in the electrophysiologic (EP) lab per our hospital policy. If patients reported that their OSA was severe enough to always require CPAP, then anesthesia was used.
The primary outcome measure was to determine the safety and efficacy of intravenous sedation for cardiac procedures administered by non-anesthesia personnel in a consecutive series of patients other than those with predetermined exclusions. Safety was defined as absence of complications described in a subsequent paragraph. Efficacy was defined as successful completion of the procedure using sedation. The secondary outcome was to assess cost effectiveness of such sedation.
Institutional sedation administration policies and requirements
Levels of sedation for cardiac procedures with NASPE/HRS sedation (SED) score 
Responds to verbal
Responds to physical
ALDRETE score 
Move 4/2/0 extremities
Breath deeply & cough/dyspnea/apnea
BP vs baseline
±20%, 20–50%, 50%
Drugs for sedation approved by Montefiore Department of Anesthesiology for use during cardiac procedures by non-anesthesia personnel
Narcotics & tranquilizers
Nurses (RNs) administered all sedation under the general direction of the attending physician. For some procedures, such as diagnostic catheterization, only light sedation  (SED score 8) was intended. Moderate approaching deep sedation was the default target for electrical (DCCV) cardioversions (SED score 3–4). Variable levels of sedation were planned for electrophysiologic studies (EP), particularly for lengthy ablations and implantable cardioverter-defibrillator (ICD) implantations or ICD checks and some DCCVs. For these, brief periods touching on deep sedation (SED score >2) were used. The target was the desired level of sedation and not a pre-specified dose of medication.
The most complex sedation protocols were developed for EP studies. Initially 1.0 mg of morphine and 1.0 mg of midazolam were administered, and repeated until the desired level of sedation is attained. If more than 8–10 mg of morphine (or equivalent narcotic) were used during a procedure, droperidol 1.25–2.5 mg was added for combined sedation and anti-emetic effect. The depth of sedation was assessed by using SED score. After the procedure, other antiemetics were prescribed prophylactically at the cardiologist’s discretion, e.g. prochlorperazine or metaclopramide. If morphine was ineffective, fentanyl in 25–50 mcg dose increments was substituted. As an adjunct or for patients who developed restlessness with midazolam, diphenhydramine 25–50 mg was added. Reversal agents were not used routinely. These agents were used judiciously during procedures, e.g. if oxygen saturation was noted to be 5% lower than baseline for a few minutes or if a patient remained sedated for too long after the procedure. Use of reversal agents that did not interrupt or otherwise affect a procedure was not considered to be a complication. After each procedure, patients were observed in the EP lab or an intensive monitoring area until they were awake, responsive to minimal verbal or tactile stimulation, and met an Aldrete score of at least 8, or returned to pre-procedure baseline.
A computerized database was available for EP studies including ICD implants (but not pacemakers) from 1993. A database for catheterization (CATH), trans-esophageal echocardiograms (TEE) and DCCVs became available in 2004. For pacemaker implants (PPM), 123 consecutive cases from 2006 were entered. In addition to the complications fields in the computerized database, we reviewed the records of the cardiology quality-assurance committee which tracks and acts on all complications. More detailed data was collected on EP-related patients because they received higher doses of sedation.
Tabulation included sex and age, type of procedure, medications and doses, including use of reversal agents. For a subset of EP procedures, procedure duration and rate of medication administration were recorded.
Complications included use of reversal agents that did interrupt or otherwise affect a procedure. Thus, requirements for emergency endotracheal intubation were counted as complications, in contrast to calls for anesthesia services for help with a non-sedatable but otherwise stable patient (high Aldrete or SED score). The latter were not counted as complications, but did represent failure of sedation. Complications also included low Aldrete or SED score findings that required interruption, termination, urgent/repeated use of reversal agents, or post-procedure discharge delays attributable to the effects of sedation. Death or significant hypotension, requiring fluid resuscitation, were also included as complications.
Cost analysis: was computed based on the fees billed and collected for each procedure by the Anesthesiology Department, using the 2005 version of the American Medical Association procedural codes (CPT). Charges are based on: a) pre-defined billing “base units” linked to specific procedures, plus b) billing units of 15 min each (average times in Table 5). Assuming a 40% reimbursement rate, each unit produced collections of $33.00.
The data were largely descriptive and tabular. Group data are presented as means and standard deviations. Differences between groups were assessed with ANOVA. A value of p < 0.05 was considered significant.
Procedures performed and medication amounts used (mean ± standard deviation)
All EP studies
7 ± 5
6 ± 5
Full diagnostic EP
5 ± 3
4 ± 3
13 ± 8
13 ± 8
Follow up EP
2 ± 2
2 ± 2
5 ± 2
5 ± 2
7 ± 3
7 ± 3
ICD check with shock
5 ± 2
4 ± 2
5 ± 4
5 ± 4
2 ± 2
4 ± 3
1 ± 1
1 ± 1
There were a total of 9 complications recorded in the 9,558 procedures. This number amounts to 0.1% of all the patients who underwent cardiac procedures under intravenous sedation administered by non-anesthesia personnel at Montefiore Medical Center. The computerized database allowed for inclusion of immediate complications, but not those that appeared post-procedure. Hence the computer data was cross-checked with the quality assurance committee reports. There were no discrepancies on procedure-related events; late events as described in the text below were gleaned only from the quality assurance reports. Less than 0.03% of all patients required unscheduled anesthesia services; all of these were related to EP procedures. A total of 5 patients died (0.05%). Three were EP patients (0.03%) and two CATH patients (0.02%); a possible role of sedation could not be excluded.
(Five of the nine complications). Two EP deaths occurred with severe hypotension requiring pressors and CPR respectively, while one died from PEA arrest. 1) One 75-year-old male underwent urgent PPM implantation for recurrent sinus arrest. After an un-eventful procedure, the patient developed hypoxia that did not resolve with naloxone and flumazenil administration. He was intubated and developed hypotension requiring pressor support, but expired 48 h later. 2) An 84-year-old male who underwent a CATH for a myocardial infarction and ventricular tachycardia arrest underwent an ICD implant. During pacing threshold testing, the patient became hypotensive and expired after an attempt at CPR. 3) A 75-year-old male underwent EP study followed by atrial flutter ablation and developed mild hypoxia prior to conclusion of the procedure. Naloxone and flumazenil were administered with a good response. Minutes after the procedure, the patient had a seizure followed by pulseless electrical activity. The patient died in spite of prompt CPR.
Of the CATH patients: 1) A 73-year-old male underwent a left cardiac CATH, after which he developed acute tubular necrosis that later resolved. Two days later, the patient underwent successful angioplasty/stenting, but 3 days later was found to have hypoglycemia, followed by unresponsiveness and pulselessness; he died 5 days after the CATH. 2) A 79-year-old male underwent an emergent right and left heart CATH. Following the right coronary injections he developed ventricular fibrillation then became hypotensive, requiring pressors and intubation. After the procedure the patient developed refractory ventricular fibrillation and died the following day.
(Four of the nine complications). 1) A 15-year-old patient developed hives during an EP procedure after receiving 1 mg of morphine and 1 mg of midazolam. All symptoms resolved with diphenhydramine and hydrocortisone. Hypotension (60 mmHg or lower) occurred in two patients during diagnostic EP studies. 2) One patient was an 81 year old male undergoing EP study for VT arrest. 3) Another was a 44 year old male undergoing EP ablation of wide complex tachycardia. Both were given reversal agents (naloxone and flumazenil) and fluids with full restoration of blood pressure. 4) Severe shortness of breath occurred during PCI of in-stent restenosis in an 80 year old male at the end of a procedure, and attributed to acute heart failure.
Unscheduled anesthesia use in stable patients: (Not considered as complications)
Unscheduled anesthesia involvement during an EP procedure was needed in three stable patients (not considered as complications). These patients were difficult to sedate with morphine, midazolam and diphenhydramine. 1) A 12 year old female undergoing WPW ablation was difficult to sedate despite morphine, midazolam and diphenhydramine. Anesthesia was called in and patient intubated. 2) Another patient was a 65 year old male who after morphine, fentanyl and midazolam was difficult to sedate and anesthesia had to sedate and subsequently intubate the patient. 3) A 78 year old male patient had morphine, versed and diphenhydramine, but remained restless, resulting in rescheduled procedure with anesthesia involvement. Two of the three were completed with anesthesia assistance; one procedure was rescheduled. A small (unrecorded) number of patients who were difficult to sedate had successful completion of their procedures. These were not recorded as complications, but are presented here anecdotally to convey the idea that although the procedures were safely completed with sedation, anesthesia support would have been appreciated.
Select procedures, rate of medicine administration
A subgroup analysis of cardioversion patients showed that female patients received a significantly lower amount of both morphine and midazolam for sedation. For females, 3.4 mg of morphine and 4.4 mg of midazolam were used, while male patients received 5.5 mg of morphine and 5.3 mg of midazolam (p ≤ 0.001 for both). Information on the patients’ weight was not available.
Estimated cost savings*
Savings per case
33.0 (7 10) = $562
33.0 × (15 9) = $792
33.0 × (10 6) = $528
33.0 × (15 18) = $1,089
33.0 × (7 6) = $429
Safety and efficacy
Safety was defined as absence of defined complications (“Methods” section). Nine complications in 9,558 patients translates to 99.91% safety.
Efficacy was defined as successful completion of the procedure using sedation. Twelve incomplete cases amounts to 99.87% efficacy.
Since the early 1990s, there has been a change to non-thoracotomy implantable defibrillators, with transvenous lead systems [11–15] and catheter based ablations. Most invasive cardiac procedures can be done in the EP lab without the need of anesthesia services (an anesthesiologist or anesthetist). There has been some reluctance to give up formal anesthesia support , and concerns have been raised in the popular press about who is providing anesthesia and sedation . Yet, general anesthesia has its own complications in both cardiologic and other surgical procedures [15, 17]. The present study demonstrates the safety of conscious sedation by trained non-anesthesia personnel, applied systematically to consecutive patients requiring cardiac procedures. Conscious sedation in the absence of an anesthesiologist is widely used for colonoscopy, pancreatography, dental, ureteroscopic and orthopedic procedures [18–23]. Conscious sedation and its safety have been examined in multiple cardiac procedures in the past decade. Although a few centers report conscious sedation to be safe in a spectrum of invasive cardiac procedures [4, 24–33], most report only one type of procedure and are under-powered to provide general reassurance for conscious sedation. A few reports with moderate numbers evaluated sedation in patients undergoing various EP procedures, and concluded deep sedation to be safe [9, 34, 35]. Most studies included middle aged to elderly patients; a few looked at the safety of conscious sedation in children [36, 37].
The present report is a large, single center, consecutive patient population study that reviews a database of nearly 10,000 cardiologic procedures. This study shows that non-anesthesia personnel can safely and efficaciously administer intravenous sedation for a wide range of elective and urgent cardiac procedures in a cost-effective fashion. The patients in this study were non-selected other than those meeting pre-determined exclusion criteria and sedation followed a protocol as described above. Ages ranged from <1 to 104 years (CATH patients at both extremes), and we believe that the large numbers are representative of patients who receive treatment in other American referral centers.
It is uncertain whether any of the five deaths (0.05%) were sedation-related. Seven of the nine complications reported occurred in the patients undergoing EP studies, which require the highest amount of sedatives used and longest procedure times (Tables 4 and 5). It is clear that all mortalities occurred in very elderly and sick patients (multiple myocardial infarctions, recurrent sinus arrest, refractory ventricular fibrillation). The rash in a 15 year-old patient, after 1 mg of morphine was likely secondary to release of histamine, a known effect of morphine .
Medication amount and infusion rates
The significant difference (p < 0.0001) in the amounts of midazolam and morphine (Table 4) used in each procedure can be attributed to the variable length of time for each procedure (Table 5), and differences in individual patient requirements. The time to perform TEE and follow-up EP was significantly shorter than the time for ablations and detailed diagnostic studies. As the procedure time increased, the dose of sedation medication increased accordingly. However the rate (mg/min) was actually higher for shorter procedures such as an ICD check (Table 5). Despite the faster rate of infusion and relatively large amount of sedatives used in a short period of time, no intubations were required in any of the ICD checks. In addition, there were no complications among the 35 procedures that took 9–10 h to complete. These results confirm that conscious sedation is safe despite the variability of medication amounts, infusion rates and length of the procedures.
The use of conscious sedation has been shown to be cost effective in several studies [13, 14, 27, 29, 39, 40]. The practice of conscious sedation also resulted in significant cost-savings while maintaining a high level of safety. Our estimates of societal cost are conservative, as only the anesthesiologist’s fees were considered. Routine participation by anesthesiology personnel would also produce additional costs for delayed scheduling with possible longer length of stay in the hospital, and use of other anesthetic agents. Differences in age, gender and cardiovascular morbidity among patients and in different hospital settings may confound the analysis of cost, because such factors may affect duration of procedure, complication rate and hospital stay. Variable and evolving reimbursement plans over the years make cost analyses of anesthesia for procedures difficult.
Sedation amount gender difference
Gender related differences were calculated in cardioversions only, since these require rather uniform and short procedure times. The finding that male patients required larger amounts of both morphine and anesthesia (p < 0.001) may reflect their greater body mass (though we do not have data on this); previous studies have suggested that males may require less sedation per kilogram [41–43]. One study showed the potency of strong opioids such as morphine to be higher in males . Two other studies showed females using higher doses (+11%) of morphine for postoperative pain  and 30% more morphine to achieve a similar degree of analgesia compared with men . Sex-related differences in the perception of pain and susceptibility to opioids remains a matter of debate.
The data in this study were derived primarily from the cardiology division computerized database. Data were complete on items such as type of procedure, and doses of midazolam and morphine. However other data including procedure duration were not always filled in (hence the “selected” data in Table 5). Patient weight was not recorded, so doses per/kilogram/minute could not be reported. Cardioversions and pacemakers were not in the computer system early in the study, thus limiting the numbers of these procedures in this report.
We report a large study based on single center database that covers all of the cardiac procedures that are performed under conscious sedation without an anaesthesiologist present. Our data confirms the safety and cost effectiveness of sedation administered by non-anesthesia personnel shown in previous studies [4, 25–34]. To our knowledge, this is the largest study done so far to deal with the issues of safety and cost in a consecutive non-selected patient population at a single medical center. We also believe that anesthesia services should be used routinely for certain types of procedures (as per exclusions in the “Methods” section) and must be available as backup on an “as needed” basis.
Conflict of Interest