External cardioversion of atrial arrhythmias to sinusrhythm causes significant pain without adequate seda-tion. Traditionally, anesthesiologists have been askedto sedate patients who require cardioversion, andpropofol is now widely used for this purpose.Recently, electrophysiologists and supervised registerednurses have become accustomed to administering anal-gesics and sedatives for the purpose of “conscioussedation” during electrophysiology studies, pacemakerimplantation, and cardioverter-defibrillator implanta-tion.1 With the advent of managed care and Medicareand Medicaid cuts, cost containment is essential.2

Physicians perform procedures while trying to keepcosts to a minimum but maintain optimum safety stan-dards for the patient. At our institution only anesthesi-

ologists are certified to administer propofol, an expen-sive sedative hypnotic with rapid onset before car-dioversion. However, electrophysiologists certified toadminister benzodiazepines and opioids could admin-ister conscious sedation during elective cardioversions.Moreover, it is often difficult to schedule cardioversionsat a mutually acceptable time for both the anesthesiol-ogist and cardiologist or electrophysiologist.

The purpose of this study was to compare the safetyand efficacy of conscious sedation administered by cer-tified electrophysiologists with propofol administeredby anesthesiologists during elective cardioversion ofatrial fibrillation/flutter to sinus rhythm.

MethodsPatients with hemodynamically stable persistent atrial fibril-

lation and flutter were included in this study. Patients sched-uled to be electively cardioverted from atrial fibrillation orflutter to sinus rhythm with the help of an anesthesiologistusing propofol were assigned to group 1. Patients scheduledto be cardioverted by an electrophysiologist certified in theadministration of conscious sedation (benzodiazepine-opioid)

Electrical cardioversion of atrial fibrillation orflutter with conscious sedation in the age of costcontainmentBruce G. Goldner, MD, Jennifer Baker, RN, Anthony Accordino, RN, Lou Sabatino, RN, Michael DiGiulio, RN,Donna Kalenderian, NP, David Lin, MD, Vincent Zambrotta, Jaime Stechel, Paul Maccaro, MD, and Ram Jadonath,MD Manhasset, NY

Background The purpose of this study was to compare the safety, efficacy, and cost of conscious sedation adminis-tered by electrophysiologists certified in the use of conscious sedation with sedation administered by anesthesiologists duringcardioversion of atrial fibrillation or atrial flutter to sinus rhythm.

Methods and Results Patients with hemodynamically stable persistent atrial fibrillation and flutter were includedin this study. Group 1 patients (n = 33) were sedated by an anesthesiologist and group 2 patients (n = 26) were sedated byan electrophysiologist. Anesthesiologists used propofol and electrophysiologists used midazolam and morphine for sedation.A cost analysis based on professional charges and cost of medications was performed for both groups and compared. Hos-pital charges were similar for both groups and were excluded from the cost analysis. Although time to sedation in group 1was shorter than that in group 2, sedation was adequate in both groups such that no patient in group 1 and only 1 patientin group 2 recalled being shocked. There were no complications in either group. The cost incurred in group 2 was less thanthat in group 1.

Conclusions Sedation administered by electrophysiologists for cardioversion of atrial arrhythmias is safe and costeffective. Midazolam and morphine, the sedative agents administered by electrophysiologists, were effective and well toler-ated by patients. (Am Heart J 1998;136:961-4.)

From the Electrophysiology Section, Department of Medicine, North Shore UniversityHospital.Submitted August 5, 1997; accepted April 2, 1998.Reprint requests: Bruce Goldner, MD, North Shore University Hospital, Electrophysi-ology Section, 300 Community Dr, Manhasset, NY 11030.Copyright © 1998 by Mosby, Inc.0002-8703/98/$5.00 + 0 4/1/90826

were assigned to group 2. The selection of patients into group1 versus group 2 depended on scheduling constraints. Patientswho were hemodynamically unstable were excluded from thestudy. Cardioversions in both groups were performed in aroom having equipment necessary for cardioversion (oxygen,code cart, suction, defibrillator, intubation tray, bag-valvemask, noninvasive blood pressure monitoring equipment,pulse oximeter, and sedatives and analgesics with their antago-nists: midazolam, morphine, flumazenil, and naloxone). Apatent intravenous infusion was established for the duration ofthe procedure and during recovery. During the procedure,vital signs, including level of consciousness and oxygen satu-ration by pulse oximetry, were performed every 5 minutes.

After written informed consent was obtained, patients inboth groups were sedated in the postabsorptive state andunderwent cardioversion to sinus rhythm. Cardioversion ofgroup 2 patients was performed according to the followingprotocol. The patient was attached to a monitor, blood pres-sure cuff, and defibrillator by external pads. With the attend-ing physician present, a nurse certified in the administrationof conscious sedation infused 1 mg of midazolam intra-venously after obtaining baseline vital signs. The patient wasobserved over the next 3 to 5 minutes for level of conscious-ness, assessed by response to mild tactile and verbal stimula-tion. If the patient required more sedation, 1 to 2 mg of mor-phine sulfate and midazolam were administered every 3 to 5minutes to achieve adequate sedation. The patient wasdeemed adequately sedated when there was no response tosoft verbal and mild tactile stimuli yet respiratory drive wasstill intact. In this state, the patients were able to be arousedby vigorous stimuli but quickly fell back to sleep withoutrecollection of the stimuli. An anesthetist was always avail-able for emergencies and known to be less than 5 minutesfrom where the cardioversion took place. The electrophysiol-

ogist and nurse were both certified in advanced cardiac lifesupport, such that they could provide ventilatory supportwith a bag-mask valve while waiting for sedatives to clear, ananesthetist to arrive, or for flumazenil or naloxone to takeeffect. When adequate sedation was achieved, cardioversionwas performed with 50 to 360 J of synchronized energy.Once the cardioversion procedure was completed, thepatient recovered for 3 hours, with vital signs assessed every10 to 15 minutes for the first hour and every 30 minutes forthe second and third hours. Outpatients were then asked toambulate for 1 hour before discharge. Inpatients werereturned to a monitored unit after the first hour of recoveryand were permitted to ambulate after the third hour.

The protocol for group 1 patients was similar to that forgroup 2. However, propofol (1 mg/kg) was used for group 1instead of midazolam and morphine, and an anesthesiologistwas present for group 1 patients. Hospital charges for roomuse and ancillary personnel were the same for groups 1 and2 and were not considered in the cost analysis.

The 2 groups were compared with respect to age, sex,medications at the time of cardioversion, medical history, leftventricular function, pain score, recollection of the event, timefrom initiation of sedation to time of cardioversion, amount ofenergy needed for cardioversion, time to return of conscious-ness, and adverse reactions to the sedatives or anesthesia.

The cost of 1 vial (500 mg/50 mL) of propofol as of July1997 was $31.50, which is greater than the cost of 1 vial (10mg/1 mL) of morphine sulfate ($0.75) and 1 vial (10 mg/10mL) of midazolam ($19.75). In addition, multiuse midazolamvials were used so that if 10 mg of the midazolam was notused, the remainder could be used for another patient, fur-ther reducing waste and cost. The cost of having an anesthe-siologist present was $500.00 and the charge for cardiover-sion was $440.00. The cost incurred by each patient in group1 was obtained by adding the anesthesiologist’s and cardiolo-gist’s fees for cardioversion to the cost of medication used.This sum was then averaged for the entire group. The costincurred by each patient in group 2 was obtained by addingthe charge for cardioversion to the cost of medication used.This sum was then averaged for the entire group.

Data analysisThe Student’s t test was used to compare age, pain score,

time from initiation of sedation to time of cardioversion, andamount of energy needed for cardioversion. A value of P <.05 was considered significant. Proportion of adverse reac-tions was compared by using the Fisher exact test.

ResultsGroup 1 was composed of 33 patients, mean age 69 ±

14 years and 70% men; group 2 was composed of 26patients, mean age 63 ± 12 years and 65% men (Table I).The difference in mean age and percentage of patients

American Heart JournalDecember 1998Goldner et al962

Group 1 Group 2(n = 33) (n = 26)

Age (yr) 69 ± 14 63 ± 12Men (%) 23 (70) 17 (65)Ejection fraction (%) 51 ± 13 57 ± 7Congestive heart failure (%) 5 (15) 1 (4)Open heart surgery (%) 5 (15) 8 (31)Myocardial infarction (%) 3 (9) 4 (15)Hypertension (%) 16 (48) 11 (42)Diabetes mellitus (%) 3 (9) 2 (8)COPD (%) 1 (3) 3 (12)HCM (%) 1 (3) 0 (0)Valvular heart disease (%) 6 (18) 6 (23)CVA/TIA (%) 2 (6) 1 (4)Inpatient (%)* 18 (55) 24 (92)

CVA/TIA, Cerebrovascular accident/transient ischemic attack; COPD, chronic obstruc-tive pulmonary disease; HCM, hypertrophic cardiomyopathy.*P = .001.

Table I. Clinical characteristics

who were men was not statistically significant. Themean ejection fraction of group 1 patients (51% ±13%) was not significantly different from that of group2 (57% ± 7%). There was no significant differencebetween group 1 and group 2 with regard to the otherbaseline characteristics (Table I).

The volume of patients who required cardioversionat our hospital was sometimes large, and outpatients,who were scheduled in advance, filled up availableelective slots when anesthesiologists were available.Inpatients could not be scheduled expeditiously at atime that was mutually convenient for both cardiolo-gist or electrophysiologist and anesthesiologist, butcould easily be incorporated into the electrophysiol-ogy schedule. Therefore there was a significantlygreater proportion of inpatients in group 2 (92%) thanin group 1 (55%) (Table I).

The proportion of patients receiving β-blockers, cal-cium channel blockers, type 1C antiarrhythmic drugs,sotalol, amiodarone, and digoxin was not significantlydifferent between groups (Table II). However, therewas a significantly greater proportion of patients ingroup 2 taking type 1A antiarrhythmic drugs (27% ver-sus 6%) (Table II). Of note, only 30 (51%) of the 59patients included in this study were receiving antiar-rhythmic medications at the time of cardioversion.

Time to sedation was 3 ± 2 minutes for group 1patients compared with 9 ± 4 minutes for group 2patients (P = .0001) (Table III). All patients in bothgroups were successfully cardioverted to sinus rhythm.Time to return of consciousness was 13 ± 6 minutes forgroup 1 patients and 14 ± 8 minutes for group 2patients (P = .56). The mean level of energy necessaryfor cardioversion to sinus rhythm and the proportion ofpatients recalling the cardioversion were not signifi-cantly different between groups. The 1 patient in group2 who recalled being shocked did not perceive theshock as being painful. There was no significant differ-

ence in the proportion of patients complaining of nau-sea (Table III). Group 1 patients required 71 ± 25 mg ofpropofol to achieve adequate sedation, whereas group2 patients required 5 ± 2 mg of midazolam plus 4 ± 2mg of morphine sulfate. No patient required intubationin either group. The mean cost incurred in group 1 was$971.50 compared with that in group 2 of $410.63.

DiscussionThis study demonstrated that sedation can be safely

administered to patients undergoing elective car-dioversion of atrial fibrillation and flutter to sinusrhythm by electrophysiologists certified in administra-tion of conscious sedation. There were no complica-tions in either group during this study. Even thoughanesthesiologists have superior airway skills, theseskills were not required during this study in any of the59 patients who were sedated. In addition, at least 2people certified in advanced cardiac life support werepresent during group 2 cardioversions and were pre-pared to ventilate the patient with a bag-valve maskuntil the anesthetist arrived. Natale et al1 evaluated thesafety of nurse-administered deep sedation for defibril-lator implantation in the electrophysiology laboratory.They concluded that adequate sedation for defibrilla-tor implantation and testing could be administeredsafely by the nursing staff. None of the 53 patients intheir study required intubation during or after the pro-cedure, died, or had any recollection of theprocedure.1

Geiger et al3 administered deep sedation in 536 con-secutive patients for various electrophysiology proce-dures (radiofrequency ablation and defibrillator andpacemaker implantation) without an anesthetist.Although they reported transient reduction in oxygensaturation in 25 patients and hypotension in 14 patients,no patient required intubation and no death occurred.3

Time to sedation is an important parameter, both

American Heart JournalVolume 136, Number 6 Goldner et al 963

Group 1 Group 2

β-Blockers (%) 5 (15) 5 (27)Calcium channel blockers (%) 10 (30) 5 (19)Type IA antiarrhythmic (%)* 2 (6) 7 (27)Type 1C antiarrhythmic (%) 2 (6) 2 (8)Amiodarone (%) 6 (18) 1 (4)Sotalol (%) 7 (21) 3 (12)Digoxin (%) 19 (58) 14 (54)

*P = .04.

Table II. Medications

Group 1 Group 2

Time to sedation (min)* 3 ± 2 9 ± 4Time to consciousness (min) 13 ± 6 14 ± 8Recollection of event (%) 0 (0) 1 (4)Nausea (%) 0 (0) 2 (8)Complications (%) 0 (0) 0 (0)Energy (J) 208 ± 92 174 ± 98

*P = .0001.

Table III. Results of sedation and cardioversion

from the patient perspective and from the point ofview of cost. In this study, time to sedation was shorterwhen sedation was administered by anesthesiologistsbecause of the use of propofol. Time to sedation aver-aged 9 minutes when midazolam and morphine wereused versus 3 minutes when propofol was used. How-ever, it is not inconceivable that electrophysiologists atour institution will some day have access to propofol.

Propofol and midazolam have been compared previ-ously in a prospective and randomized study duringpercutaneous transluminal angioplasty.4 Although bothdrugs resulted in sufficient sedation and anxiolysis,satisfaction of the interventionist was significantlygreater for propofol. In addition, propofol caused lessrespiratory depression than did midazolam.4

The cost of cardioversion was approximately $500.00more when an anesthesiologist was present during car-dioversion. However, not considered in the analysis ofcost was the extra time required to sedate patients withmidazolam and morphine compared with propofol.However, when cardiologists and electrophysiologistshave access to propofol, the issue of time to sedationwill be moot. Alternatively, in the age of cost contain-ment, the cost of the anesthesiologist might be bundledinto the cardioversion procedure so that the cost is thesame with or without the anesthesiologist present.

Only 51% of the patients who underwent cardiover-sion in this study were receiving antiarrhythmic med-ications at the time of cardioversion. The decision toinitiate antiarrhythmic therapy was left to the discre-tion of the electrophysiologist or referring cardiologist.Reasons for not using antiarrhythmic therapy includedconcern about proarrhythmia and side effects, firstattempt to cardiovert, and previously failed attemptwith chemical cardioversion.

Cost containment was one reason that electrophysiol-ogists performed elective cardioversions at our institu-tion without direct anesthesiologist assistance. An evengreater driving force, however, was the need to per-

form these cardioversions in a timely manner. Fre-quently, cardioversions could not be scheduled at amutually agreeable time for both anesthesiologist andelectrophysiologist. Therefore electrophysiologists atour institution performed cardioversions with con-scious sedation. Cost savings came as a byproduct ofthis emerging practice.

This study has several limitations. First, the patientswere assigned to each group on the basis of schedul-ing constraints and were not randomly assigned. Sec-ond, the sample size was small and should beexpanded in future studies.

In conclusion, this study demonstrated that electro-physiologists trained in the administration of conscioussedation can safely and effectively administer sedativesto patients undergoing nonemergency cardioversion ofatrial fibrillation and flutter to sinus rhythm. Moreover,the sedatives administered by electrophysiologists,midazolam and morphine, were effective and well tol-erated by patients. Finally, there is substantial cost sav-ings with the use of conscious sedation. However, wedo not advocate performing cardioversions without ananesthetist in the vicinity of where the cardioversion istaking place. One should be available within minutesshould the need arise.

References1. Natale A, Kearney MM, Brandon MJ, Kent V, Wase A, Newby KH,

et al. Safety of nurse-administered deep sedation for defibrillatorimplantation in the electrophysiology laboratory. J Cardiovasc Elec-trophysiol 1996;301:301-6.

2. DeMaria AN, Lee TH, Leon DF, Ullyot DJ, Wolk MJ, Mills PS, et al.Effect of managed care on cardiovascular specialists: involvement, atti-tudes and practice adaptations. J Am Coll Cardiol 1996;28:1884-95.

3. Geiger MJ, Wase A, Kearney MM, Brandon MJ, Kent V, NewbyKH, et al. Evaluation of the safety and efficacy of deep sedation forelectrophysiology procedures administered in the absence of ananesthetist. PACE 1997;20:1808-14.

4. Wagner HJ, Nowacki J, Klose KJ. Propofol versus midazolam forsedation during percutaneous transluminal angioplasty. J VascInterv Radiol 1996;7(5):673-80.

American Heart JournalDecember 1998Goldner et al964