American Journal of Emergency Medicine xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
American Journal of Emergency Medicine journal homepage: www.elsevier.com/locate/ajem
Brief Reports
Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital☆,☆☆,★ Jennifer L. White, MD a,⁎, Michael B. Heller, MD b, Robert J. Kahoud, MD c, Daniel Slade d, John D. Harding, MD e a
Department of Emergency Medicine, Mayo Clinic, Rochester, MN Department of Emergency Medicine, Beth Israel, New York, NY Department of Pediatric and Adolescent Medicine, Mayo Clinic, Rochester, MN d Temple Medical School, Philadelphia, PA e Department of Electrophysiology, Doylestown Hospital, Doylestown, PA b c
a r t i c l e
i n f o
Article history: Received 11 January 2015 Received in revised form 27 March 2015 Accepted 27 March 2015 Available online xxxx
a b s t r a c t Background: A standard approach to recent onset atrial fibrillation (AF) in the emergency department (ED) in the United States has not been established. Purpose: The purpose of this prospective clinical trial was to determine how an ED protocol emphasizing rhythm control for recent onset AF compared similar patients receiving standard therapy in the same facility. Methods: We enrolled consecutive patients presenting to our community hospital with recent onset AF into a protocol, which called for rhythm control with procainamide and if unsuccessful electrical cardioversion and discharge home. We compared this prospective cohort with matched historical controls. Primary outcome was admission rate. We also compared ED conversion rates and lengths of stay (LOS). We reported 30-day data on the study group including ED recidivism, recurrent AF, outpatient follow-up, and any important adverse events. Results: Fifty-four patients were enrolled in the study group with 4 being admitted compared with 30 of 50 in the historical control group. Ninety-four percent of the study group converted compared with 28% in the historical control. Both hospital and ED LOS were significantly shorter for the study group. Six patients had recurrent AF, and 4 of those returned to the ED. Conclusion: An ED protocol that uses rhythm control decreased hospital admission and LOS, and there were no adverse events at 30 days. © 2015 Elsevier Inc. All rights reserved.
1. Introduction
pharmacologic agent and/or synchronized electrical cardioversion (EC) and discharge home from the ED without anticoagulation [2].
1.1. Background Standard approach in the Unites States to patients with atrial fibrillation (AF) differs from that of many other countries. Patients are often admitted to the hospital and treated with rate control, anticoagulation, and inpatient cardioversion. Outside the Unites States, there is usually a very different approach to the patient with recent onset AF, defined as AF less than 48 hours' duration [1]. This typically involves an emergency department (ED) protocol that incorporates early rhythm control with a
☆ No funding was received; statistics were provided by Mayo Clinic Department of emergency medicine research budget. ☆☆ There are no conflicts of interests. ★ This study was presented at the mid-Atlantic SAEM regional meeting in Philadelphia, PA, on February 22, 2014; the New England regional SAEM meeting in New Haven, CT, March 26, 2014; and at the SAEM national meeting in Dallas, TX, May 14, 2014. ⁎ Corresponding author at: Department of Emergency Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. Tel.: +1 507 255 7002; fax: +1 507 255 6593. E-mail address: [email protected] (J.L. White).
1.2. Importance The burden of AF on the health care system is substantial and is expected to grow [3]. Although the literature is often unclear as to whether symptomatic rapid AF was the proximate cause of the ED visit, it is estimated that 64% of visits to the ED for AF result in admission to the hospital and that 21% of patients presenting have recent onset AF [1,3]. Health care resources and dollars are therefore spent on patients who are otherwise low risk for serious complications, including stroke, who could potentially be discharged from the ED [3]. Other investigators, both in the Unites States and abroad, have safely implemented a rhythm control method, discharging a large percentage of patients presenting to the ED with recent onset AF [4,5]. There has been a trend toward ED observation units managing such patients; however, the emphasis remains on rate control [6]. Protocols emphasizing an ED rhythm control strategy have not been adopted in the United States to date, but with the recent emphasis on cost-efficient management of common problems in cardiology, particularly, the time for examination of such an approach appears propitious.
http://dx.doi.org/10.1016/j.ajem.2015.03.059 0735-6757/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
2
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
1.3. Purpose of this study The purpose of this prospective trial was to compare an expedited rhythm control strategy in select patients with recent onset AF compared with a recent historical control at our own institution. Our main comparator was admission rates. We also looked at conversion rates, ED length of stay (LOS), and total LOS. Our prospective cohort was also followed up 30 days later and assessed for ED recidivism, adverse events such as stroke, and patient satisfaction. Emergency physicians enrolling patients were also polled on their subjective experience. 2. Materials and methods 2.1. Outline of study design We conducted a prospective, observational study of consecutive patients with AF (including 1 with atrial flutter) duration less than 48 hours who presented to the ED from May 1, 2013 to April 30, 2014 who met inclusion criteria. We compared this prospective group with historical controls, consecutive patients with recent onset AF who presented to the ED from January to September 2011 and met inclusion criteria for the protocol. This period was chosen specifically because it represented the most pure time before initiation of the protocol. During this specific period, the treatment of recent onset AF was mainly cardiology consultation, admission, and rate control. This timeframe was chosen, as it was just before any rollout or pilot of the rhythm control method. The protocol was developed in collaboration with all 19 cardiologists at our institution in conjunction with the emergency medicine physicians. The rollout took approximately 2 years, including educating all 12 emergency physicians and nurses on study design and specific inclusion criteria. Meetings, e-mails, and posters in the ED accomplished this. The group of ED physicians consisted of all board-certified ED physicians, and each physician enrolled at least 1 patient during the study period; many had never performed ED cardioversion for stable AF before this protocol.
was excluded. Other exclusion criteria were known ejection fraction (EF) less than 35%, prior coronary artery disease, obvious ischemia on electrocardiogram (ECG), evidence of Wolf Parkinson White syndrome (WPW), unknown duration of symptoms, primary diagnosis not AF, other reason to observe/admit patient, prior cerebrovascular accident (CVA) not on oral anticoagulation, febrile, any antiarrhythmic in the past 72 hours, concurrent QT-prolonging medications (left to discretion of treating provider), known QTc prolongation of 460 milliseconds, and any hospitalizations within the prior 3 months. Prior AF was not an exclusion criterion. Standard consent for procedural sedation and direct current cardioversion was obtained per hospital protocol. Cardiology consultation was not required, as this protocol was specifically designed to be ED driven. All discharged patients were ensured prompt outpatient cardiology follow-up. The protocol was in keeping with the current at the time American Heart Association (AHA)/American College of Cardiology guidelines from 2010 for the management of AF [7,8].
2.5. Specific study protocol All enrolled patients were given 1 g of procainamide intravenous over 1 hour as a standard infusion. Nurses were instructed to stop the infusion if the patient developed symptomatic hypotension that did not respond to a 500 mL bolus of normal saline or if the QRS widened more than 50% of baseline. The nurses obtained vital signs every 15 minutes and rhythm strips during the protocol. Once sinus rhythm was restored, the infusion was stopped, and they were discharged home. Patients who did not convert with procainamide alone were observed for 1-hour postinfusion and then offered ED synchronized cardioversion. The ED physician performed the procedural sedation and cardioversion, and propofol sedation was used. Standard postprocedural sedation protocol was followed, and patients were discharged home after successful cardioversion. If the patient did not convert either with procainamide or EC, the ED physician consulted cardiology for further management. No new medications were prescribed.
2.2. Development of AF protocol A specific order set for the protocol was developed in the computer order entry system to ensure standardization and safety. This was also used as a way to track all patients enrolled in the protocol in case a data collection sheet was not filled out. While we began enrolling patients during the pilot period, these patients were not ultimately included in the final study cohort. This pilot period spanned the 2 years before the start of the study. The historical control group was obtained by chart review by the primary and co-investigators. The chart review period was only 9 months because an a priori sample size was calculated at 46 patients, and we obtained 50 within that timeframe. Research ethics board approval was granted by the institutional hospital institutional review board with a decision to waive the need for informed consent, as this was considered one standard of care.
2.6. Follow-up methods for prospective cohort A 30-day phone call was made to each enrolled patient in the protocol to assess recurrent AF, adverse events, patient satisfaction, and promptness of outpatient cardiology follow-up. This phone call was made by one of the investigators and followed a standardized data collection form that was created a priori. There were 3 attempts to contact each patient via phone and e-mail. If this was unsuccessful, we then obtained some of the follow-up data from their chart, specifically their cardiology appointment after the ED visit. In addition, a standardized satisfaction survey was given to the 8 emergency physicians after study completion to assess their experience with adoption of the protocol.
2.3. Study population and demographics
2.7. Data collection for historical cohort
The ED is part of a nonteaching community hospital with an average ED annual census of 60 000. All emergency physicians were board certified, but only 1 had performed cardioversion for stable AF before initiation of this protocol.
This was compared with a pure historical control group. The rollout of this protocol took close to 2 years with various pilots that were started and stopped, while the protocol was revised. This explains the “pure” historical control group—how the ED was handling recent onset AF before any talk of ED cardioversion and thus was remote from the prospective cohort. The historical controls were consecutive patients abstracted from the charts who presented with recent onset AF (b 48 hours) to our ED. If timing of AF was unclear from the records, then patients were excluded from the protocol. Patients were included in the historical cohort if they met all of the study criteria, although the retrospective chart review methods for identifying historical controls could not identify every exclusionary criterion in every case.
2.4. Inclusion and exclusion criteria If patients met inclusion criteria, the emergency physician rendering care enrolled them in the study. We included consecutive patients presenting to the ED with AF less than 48 hours who did not have any of our specific exclusion criteria. The determination of recent onset AF was left to the provider's discretion; however, if there was any doubt, the patient
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
3
2.8. Outcome measures The primary outcome measure was admission rates. Secondary measures included conversion rates, ED LOS, and total hospital LOS. There is also 30-day descriptive data describing recurrent AF, patient satisfaction score, adverse events (transient ischemic attack, stroke, and death), and promptness of cardiology follow-up. Emergency department physicians and nurses collected data on an ongoing basis, using a standardized data collection sheet kept in the ED. Nurses and physicians received appropriate training on the protocol, inclusion criteria, and data collection. Data were entered into a confidential spreadsheet and deidentified. Emergency department LOS and hospital LOS were abstracted from the chart and reported in minutes. Thirty-day data were obtained by a call back system. The historical cohort was obtained by reviewing consecutive charts and abstracting data by the co-investigators. An initial sample of 10 charts was done concurrently to ensure data abstraction accuracy across investigators. 2.9. Data analysis and a priori sample size calculation An a priori sample size calculation based on an α of .05 and size effect of a 30% reduction in admission rates was calculated at 48 subjects in each arm before initiation of the study based on the study conducted by Jacoby et al [5]. Baseline characteristics were compared. Primary and secondary end points were compared between the 2 study groups and assessed for statistical significance. Descriptive statistics were provided for the 30-day data in the prospective group only. Continuous features were summarized with medians, interquartile ranges, and ranges; categorical features were summarized with frequency counts and percentages. Comparisons between patients in the historical and prospective cohorts were evaluated using Wilcoxon rank sum and χ 2 tests. Statistical analyses were performed using the SAS software package (SAS Institute, Cary, NC). All tests were 2 sided with α set at .05. 3. Results 3.1. Patient enrollment We enrolled 54 total patients into the prospective observational arm of the study. This compares with 50 patients in the historical control, all of which met inclusion criteria. Baseline characteristics of the 2 groups were similar in terms of age, median CHA2DS2VASc scores, and sex. Of the 384 patients who presented to the ED during the study period with a final diagnosis of AF, a total of 68 (17.4%) were eligible, and a total of 54 (14.4%) were enrolled. There were 14 patients who were not enrolled but were eligible. Reasons for not enrolling 11 of those patients were related to provider preference for rate control and cardiology management. No other specifics are known as to why they were not enrolled. The other 3 who were eligible but not enrolled got propafenone, a rhythm control strategy that was used occasionally before the initiation of this protocol (Fig. 1). Patients who were eligible but not enrolled were similar to the prospective cohort. There are no follow-up data on these patients (Fig. 2).
Fig. 1. Flow diagram of prospective study cohort.
3.3. Primary and secondary outcomes There were 4 (7%) patients of 54 in the prospective group who were admitted to the hospital, 3 of which were the patients who did not convert to sinus rhythm, and 1 that did convert but had aberrant conduction and an abnormal postconversion ECG (Table 3). In the historical group, 30 (60%) of the 50 patients were admitted to the hospital (P b .001); this did not include the patients who had transesophageal echocardiogram (TEE) cardioversion in the cardiac catheterization laboratory and discharged from there. This can also be stated as the difference in hospitalization proportions, 0.60 to 0.07 is equal to 0.53. A 95% confidence interval for this difference is 0.35 to 0.66. Of the 54 patients in the prospective group, 51 (94%) converted to sinus rhythm in the ED. Of these, 34 patients (67%) converted with procainamide alone, and the remaining 17 (33%) converted with EC. Time to conversion after initiation of the procainamide drip ranged from 10 to 120 minutes, as we allowed 1-hour postinfusion to convert. This compared with the historical group where 28 (56%) of 50 converted in the ED, all with medication or spontaneously, none with EC. Medications that were used in the historical group included only diltiazem and propafenone. The much higher admission rate accounts for the significant drop in hospital LOS for the prospective group, 274 vs 966 minutes (P b .001). The ED LOS was also significantly shorter in the prospective group, 269 vs 364 minutes (P b .01).
3.2. Baseline demographics 3.4. Emergency department adverse events The median CHA2DS2VASc score for the prospective group was 1 with 43% female and a median of 58.5 (Table 1). The presenting symptom was either “typical AF” or palpitations in all of the patients (Table 2). Patients were in AF between 30 minutes and 18 hours. Although we included patients in AF up to 48 hours, the longest period of AF was 18 hours. This was not intentional, as all patients with AF up to 48 hours could have been enrolled. One of these patients had atrial flutter as the presenting rhythm; all others had AF.
There were 2 patients who developed transient hypotension during procainamide infusion, which resolved with 500 mL bolus of normal saline (NS); and in both cases, the drip did not have to be discontinued. There were 2 patients who developed a wide complex rhythm during the infusion of procainamide, and in both cases the drip was discontinued, and patients were immediately electrically cardioverted without complication. The electrophysiologists who examined the
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
4
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
Fig. 2. Emergency department–driven AF protocol.
ECGs after the ED visit determined both rhythm changes to be aberrancy and not ventricular tachycardia. 3.5. Prospective cohort descriptive data All patients in the prospective group were called 30 days after initial presentation to ED. All 54 patients were followed up. Of those we could not reach by phone, follow-up data were collected based on chart review and cardiology office visit follow-up information. Of the 54 patients contacted, 7 had recurrent AF either recognized on Holter monitor, loop recorder, or subsequent symptomatic episode (Table 4). Of the 6 patients with recurrent AF, 5 returned to the ED in the following 30 days. There were no important adverse events in the study group, specifically death or CVA/transient ischemic attack. Of the 54 patients contacted, 52 had prompt cardiology follow-up determined to be less than 14 days from ED visit. There were 3 patients in the follow-up group who had ablation in the
following 30 days. Patient comments were overwhelmingly positive, and all were satisfied with their experience, particularly with their disposition to home. We used a scale of 1 to 10 to indicate their patient experience at the time of the ED visit, with 1 being “I would never return to that ED again” and 10 being “Was a fantastic experience.” The mean score was a 9.7 of 48 patients who completed the study, due to 1 outlier that was upset over being charged for a
Table 1 Baseline characteristics of prospective and historical cohort Characteristic
Historical controls (CI)
Study (CI)
Median age Male Median CHA2DS2VASc score
66.5 (56-71) 24 2 (1-3)
58.5 (48-67) 31 1 (0-2)
Abbreviation: CI, confidence interval.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx Table 2 Study cohort characteristics
5
Table 4 Thirty-day follow-up data from study cohort
Characteristic
n
Variable
No of patients, n = 54
Median age Male Presenting complaint Palpitations “Typical a-fib” Lightheadedness Chest pain Avg time in AF before presentation Median CHADS2VASC score 0 1 2 3 4 5 Converted in ED Converted with procainamide Converted with EC Complications in ED Transient hypotension Aberrant conduction Number contacted 30 d later Number followed up
58.5 (48-67; 24-86) 31 (57%)
Recurrent AF Return to the ED Stroke Death Medication complication, Coumadin Ablation Cardiology follow-up within 14 d
7 5 0 0 1 4 52
42 (77%) 9 (16%) 3 (6%) 1 (1%) 205 min 1 (0-2; 0-5) 18 (33%) 16 (30%) 9 (17%) 6 (11%) 3 (6%) 2 (4%) 51 (94%) 35 (68.6%) 16 (31.3%) 4 2 2 48 (89%) 54 (100%)
Abbreviations: a-fib, atrial fibrillation; Avg, average.
urine drug screen he did not think was indicated. Additionally, the Emergency physicians were polled regarding their experience (Table 5). All enrolling physicians were overwhelmingly satisfied with the protocol and answered “strongly agree” to the survey, with the exception of one physician that answered “agree”. 4. Discussion We successfully developed and implemented an expedited rhythm control method for patients with recent onset AF in a community hospital. This protocol was both safe (as far as can be judged in a trial of 54 patients) and efficacious. Stiell et al [4] report a decade-long experience of thousands of patients safely enrolled in a similar protocol in Canada. This is now adapted as standard of care for management of patients with recent onset AF. This is the first US study to incorporate this robust experience and challenge the status quo. Although much smaller in terms of patient data, the outcomes are quite similar. Historically, in our institution, patients presenting to the ED with symptomatic AF were treated in a myriad of ways. The predominate management strategy was one of prompt ED cardiology consultation and management per the cardiologist. This generally included rate control and either TEE-guided cardioversion or hospital admission. “The pill in the pocket” approach with attempted ED pharmacologic cardioversion was also used in selected patients. This was directed by cardiology, while the patient remained in the ED. From the standpoint of the emergency physician, the historical strategy was actually less risky. Other studies like this one reported the safety of ED cardioversion and highlighted the potential barriers to implement a practice change [9,10]. The barriers were the same in our study [11]. These include fear of adverse outcomes, specifically stroke and the ability of the ED physician to truly identify the low-risk patient [10]. This study
attempted to select a low-risk cohort of patients who were good candidates for an ED rhythm control method by using multiple exclusion criteria. A better risk predictor is the CHA2DS2VASc score [12,13]. This would be a simpler and more generalizable manner of ensuring a safe ED-driven protocol invoking a rhythm control method, consistent with the current AHA guidelines. Scheuermeyer et al [12] recently looked at how often the ED physician initiates anticoagulation in the ED in patients presenting with AF. This study separated patients according to CHADS2 scores and found that the emergency physician did not routinely offer anticoagulation despite a high CHADS2 score. Despite this, there was only 1 patient that had an ischemic stroke in the 30 days after ED visit. Another barrier to implementing such a strategy is the question of the provider's ability to know exactly when a patient went into AF. This was addressed in our study in that the physician was told to use the longest duration of AF obtained by any source at the time of the ED visit, and if duration was unknown or the patient did not ever feel their AF, they were automatically excluded. This represents a real-world patient encounter. The promptness of outpatient cardiology has been cited as an important predictor of poor outcomes in the patient with AF [14]. Specifically, prompt outpatient cardiology follow-up has been associated with decreased overall morbidity and mortality. The uniqueness of our study is the partnership between the ED physicians and cardiologists, highlighting the benefits of a team approach, which has been proven to improve patient care in many settings. The protocol was developed and implemented by both groups and demonstrated the feasibility of ensuring prompt outpatient cardiology follow-up even without cardiology involvement of care while in the ED. Given the predicted increase in prevalence of AF in the coming decades due to the aging US population, standardized treatment protocols that improve quality and reduce cost will be a necessity [1,15]. Although the primary goal was not to improve cost of care, the reduction in admission rate seen in our study will certainly reduce overall cost of treating AF in the ED if other groups can reproduce our results. Given that other studies suggest increased stroke risk in the 30 days after index AF visit to the ED, prompt follow-up and guideline-based initiation of oral anticoagulation should also reduce thromboembolic events [16,17].
Table 5 Physician satisfaction survey regarding AF protocol Key = 1, strongly agree; 2, agree; 3, slightly agree; 4, neutral; 5, slightly disagree; 6, disagree; 7, strongly disagree 1 2 3
Table 3 Comparison of historical cohort and prospective cohort for primary outcome measures Variable Patients converted In ED Patients admitted Mean ED LOS Mean total LOS
Historical cohort 28 (56%) 30 (60%) 344.5 (240-467) 966 (440-1631)
Study cohort 51 (94%) 4 (7%) 269.5 (213-336) 274 (213-350)
P b.001 b.001 .011 b.001
4
The AF protocol is better than what we were doing prior? 1 2 3 4 5 6 The AF protocol is safe for patients? 1 2 3 4 5 6 I am satisfied with the AF protocol that has been implemented 1 2 3 4 5 6 I would recommend this protocol to other EDs 1 2 3 4 5 6
7 7 7 7
Results of 8 physicians surveyed (100% response rate): Question 1 = mean score, 1.1 (1 physician answered 2). Question 2 = mean score, 1.1 (1 physician answered 2). Question 3 = mean score, 1. Question 4 = mean score, 1.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
6
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
5. Limitations The primary limitation is the study design. This study was born out of a practice change within the institution during development of a hospital-wide AF center. It was designed intentionally to study the new AF protocol as a result of the development of this center. It would have been better to have a randomized controlled trial design. We obtained a historical control instead. Although our constellation of exclusion criteria could be cited as a barrier to enrolling more patients, this can also be viewed as a strength of the study. Reasons for excluding patients who were recently hospitalized or had coronary artery disease were based on cardiology recommendations. Cardiologists at our institution wanted to ensure a cohort of healthy, lowrisk patients. Ultimately, the protocol was revised to risk stratify based more on CHA2DS2-VASC scores. Although limiting patient enrollment, it used standard exclusion criteria that could be applied almost anywhere, as it was safely and effectively applied at a nonteaching hospital. We also noted the average length of AF was only 18 hours in our study cohort, although all patients with duration less than 48 hours could have been enrolled. The patients who were eligible but not enrolled fell within the mean of AF duration of the study cohort, and decision not to enroll them was not based on length of AF. This could be seen as a limitation, as 14 patients were eligible but not enrolled, mostly based on physician preference, and if they were included, it could have changed the data. We do not have outcomes or follow-up for those 14 patients and that would have been helpful. Our decision to use a standard dose of procainamide potentially underestimates the percentage of pharmacologic cardioversion in our study, as procainamide is often used as a weight-based drug. With a weight-based dosing, it is also possible that more arrhythmias may have occurred with higher overall dosing. We were unable to contact any of the historical group via phone per institutional review board guidelines. This limits our ability to subjectively compare 30-day outcomes with the historical group. We also did not use a standardized patient satisfaction scoring system, although we did specifically seek out patient experience via a 1 to 10 numerical scale. Prior studies have not looked at patient satisfaction for an ED-driven rhythm control method; and although our study did look at this, we did not use a standardized scale. Our study did not address the current anticoagulation guidelines as of this publication. The new AHA/American College of Cardiology/ CHEST guidelines were published after initiation of the study [9]. Our ongoing protocol was amended after the new guidelines were published to reflect change in oral anticoagulation guidelines. This new protocol was not used in any of the patients in this study. A potential limitation of the study was that it did not address the current anticoagulation guidelines that were published after the study was ended. The newest version of this protocol incorporates using CHA2DS2VASc scores to risk stratify those patients who leave the ED with oral anticoagulation. Our study did not offer oral anticoagulation to anyone.
6. Conclusions An expedited ED-driven rhythm control method resulted in decreased admission rates, ED, and total LOS. This protocol was easily adopted by community ED physicians and implemented without adverse events and resulted in a high conversion rate compared with
standard therapy. At 30 days, there were no strokes, and patient satisfaction was high. Our institution continues to implement this protocol. Acknowledgments Robert Slade, MD: enrollment of patients and support of study design and concept; Doylestown emergency physicians and nurses for enrollment of patients; Doylestown Hospital Cardiologists for support of study design and implementation; John Mitchell; Robert Sangrigoli, MD; and Stephen Sloan, MD for cardiology specialty consultation, support of study, and design of the AF protocol; Erik Hess, MD—support in preparing manuscript; Christine Lohse—preparing of the statistics; and Ian Stiell MD—permission to use a version of the Ottawa aggressive protocol and mentoring along the way.
References [1] Arco C, Martin A, Laguna P. Analysis of current management of atrial fibrillation in the acute setting: GEFAUR-1 Study. Ann Emerg Med 2005;46:424–30. [2] Besser K, Mills A. Is discharge to home after emergency department cardioversion safe for the treatment of recent-onset atrial fibrillation? Ann Emerg Med 2011;58: 517–20. [3] McDonald AJ, Pelletier AJ, Ellinor PT, Camargo Jr CA. Increasing US emergency department visits rates and subsequent hospital admissions for atrial fibrillation from 1993-2004. Ann Emerg Med 2008;51:58–65. [4] Stiell IG, Clement CM, Perry JJ, Vallancourt C, Symington C, Dickinson G, et al. Association of the Ottawa aggressive protocol with rapid discharge of emergency department patients with recent onset atrial fibrillation or flutter. Can J Emerg Med 2010; 12:181–91. [5] Jacoby JL, Cesta M, Heller MB, Salen P, Reed J. Synchronized emergency department cardioversion of atrial dysrhythmias saves time, money and resources. J Emerg Med 2005; 28:27–30. [6] Decker WW, Smars PA, Vaidyanathan L, Goyal DG, Boie ET, Stead LG, et al. A prospective, randomized trial of an emergency department observation unit for acute onset atrial fibrillation. In: Smars PA, VL, editors. Ann Emerg Med; 2008. p. 322–8. [7] Fuster V, Ryden LE, Cannom DS, Crijns HJ, Curtis AB, Ellenbogen KA, et al. ACC/AHA/ ESC 2006 guidelines for the management of patients with atrial fibrillation. Circulation 2006;114:700–52. [8] JE E, Lowe N, Estes M, group W. Cardiology Foundation/American Heart Association Task Force on practice guidelines on atrial fibrillation (updating the 2006 guidelines). A report of the American College of Cardiology; 2011. [9] January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC, et al. 2014 AHA/ ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the Heart Rhythm Society. J Am Coll Cardiol 2014;14: 01740–9. [10] Oishi ML, Xing S. Atrial fibrillation management strategies in the emergency department. Emerg Med Pract 2013;15:1–26. [11] Kirchhof P, Gunter B, Camm J, Crijns H. Improving outcomes in patients with atrial fibrillation: rationale and design of the early treatment of atrial fibrillation for stroke prevention trial. Am Heart J 2013;166:442–6. [12] Scheuermeyer FX, Innes G, Pourvali R, Dewitt C, Grafstein E, Heslop C, et al. Missed opportunities for appropriate anticoagulation among emergency department patients with uncomplicated atrial fibrillation or flutter. Ann Emerg Med 2013;62: 557–65. [13] Coppens M, Eikelboom JW, Hart RG, Yusuf S, Lip GY, Dorian P, et al. The CHA2DS2VASc score identifies those patients with atrial fibrillation and a CHADS2 score of 1 who are unlikely to benefit from oral anticoagulant therapy. Eur Heart J 2013; 34:170–6. [14] Atzema CL, Austin PC, Chong AS, Dorian P. Factors associated with 90 day death after emergency department discharge for atrial fibrillation. Ann Emerg Med 2013;61: 539–48. [15] Stead LG, Decer WW. Rhythm versus rate control for atrial fibrillation and flutter. Ann Emerg Med 2006;47:496–8. [16] Scheuermeyer FX, Grafstein E, Stenstrom R, Stenstrom R, Innes G, Heslop C, Macphee J, et al. Thirty-day and 1 year outcomes of emergency department patients with atrial fibrillation and no acute underlying cause. Ann Emerg Med 2012;60:755–65. [17] Swaminathan A. Recent onset atrial fibrillation: the case for emergency department cardioversion. Acad Emerg Med 2013;20:408.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
Contents lists available at ScienceDirect
American Journal of Emergency Medicine journal homepage: www.elsevier.com/locate/ajem
Brief Reports
Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital☆,☆☆,★ Jennifer L. White, MD a,⁎, Michael B. Heller, MD b, Robert J. Kahoud, MD c, Daniel Slade d, John D. Harding, MD e a
Department of Emergency Medicine, Mayo Clinic, Rochester, MN Department of Emergency Medicine, Beth Israel, New York, NY Department of Pediatric and Adolescent Medicine, Mayo Clinic, Rochester, MN d Temple Medical School, Philadelphia, PA e Department of Electrophysiology, Doylestown Hospital, Doylestown, PA b c
a r t i c l e
i n f o
Article history: Received 11 January 2015 Received in revised form 27 March 2015 Accepted 27 March 2015 Available online xxxx
a b s t r a c t Background: A standard approach to recent onset atrial fibrillation (AF) in the emergency department (ED) in the United States has not been established. Purpose: The purpose of this prospective clinical trial was to determine how an ED protocol emphasizing rhythm control for recent onset AF compared similar patients receiving standard therapy in the same facility. Methods: We enrolled consecutive patients presenting to our community hospital with recent onset AF into a protocol, which called for rhythm control with procainamide and if unsuccessful electrical cardioversion and discharge home. We compared this prospective cohort with matched historical controls. Primary outcome was admission rate. We also compared ED conversion rates and lengths of stay (LOS). We reported 30-day data on the study group including ED recidivism, recurrent AF, outpatient follow-up, and any important adverse events. Results: Fifty-four patients were enrolled in the study group with 4 being admitted compared with 30 of 50 in the historical control group. Ninety-four percent of the study group converted compared with 28% in the historical control. Both hospital and ED LOS were significantly shorter for the study group. Six patients had recurrent AF, and 4 of those returned to the ED. Conclusion: An ED protocol that uses rhythm control decreased hospital admission and LOS, and there were no adverse events at 30 days. © 2015 Elsevier Inc. All rights reserved.
1. Introduction
pharmacologic agent and/or synchronized electrical cardioversion (EC) and discharge home from the ED without anticoagulation [2].
1.1. Background Standard approach in the Unites States to patients with atrial fibrillation (AF) differs from that of many other countries. Patients are often admitted to the hospital and treated with rate control, anticoagulation, and inpatient cardioversion. Outside the Unites States, there is usually a very different approach to the patient with recent onset AF, defined as AF less than 48 hours' duration [1]. This typically involves an emergency department (ED) protocol that incorporates early rhythm control with a
☆ No funding was received; statistics were provided by Mayo Clinic Department of emergency medicine research budget. ☆☆ There are no conflicts of interests. ★ This study was presented at the mid-Atlantic SAEM regional meeting in Philadelphia, PA, on February 22, 2014; the New England regional SAEM meeting in New Haven, CT, March 26, 2014; and at the SAEM national meeting in Dallas, TX, May 14, 2014. ⁎ Corresponding author at: Department of Emergency Medicine, Mayo Clinic, 200 First Street SW, Rochester, MN 55905. Tel.: +1 507 255 7002; fax: +1 507 255 6593. E-mail address: [email protected] (J.L. White).
1.2. Importance The burden of AF on the health care system is substantial and is expected to grow [3]. Although the literature is often unclear as to whether symptomatic rapid AF was the proximate cause of the ED visit, it is estimated that 64% of visits to the ED for AF result in admission to the hospital and that 21% of patients presenting have recent onset AF [1,3]. Health care resources and dollars are therefore spent on patients who are otherwise low risk for serious complications, including stroke, who could potentially be discharged from the ED [3]. Other investigators, both in the Unites States and abroad, have safely implemented a rhythm control method, discharging a large percentage of patients presenting to the ED with recent onset AF [4,5]. There has been a trend toward ED observation units managing such patients; however, the emphasis remains on rate control [6]. Protocols emphasizing an ED rhythm control strategy have not been adopted in the United States to date, but with the recent emphasis on cost-efficient management of common problems in cardiology, particularly, the time for examination of such an approach appears propitious.
http://dx.doi.org/10.1016/j.ajem.2015.03.059 0735-6757/© 2015 Elsevier Inc. All rights reserved.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
2
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
1.3. Purpose of this study The purpose of this prospective trial was to compare an expedited rhythm control strategy in select patients with recent onset AF compared with a recent historical control at our own institution. Our main comparator was admission rates. We also looked at conversion rates, ED length of stay (LOS), and total LOS. Our prospective cohort was also followed up 30 days later and assessed for ED recidivism, adverse events such as stroke, and patient satisfaction. Emergency physicians enrolling patients were also polled on their subjective experience. 2. Materials and methods 2.1. Outline of study design We conducted a prospective, observational study of consecutive patients with AF (including 1 with atrial flutter) duration less than 48 hours who presented to the ED from May 1, 2013 to April 30, 2014 who met inclusion criteria. We compared this prospective group with historical controls, consecutive patients with recent onset AF who presented to the ED from January to September 2011 and met inclusion criteria for the protocol. This period was chosen specifically because it represented the most pure time before initiation of the protocol. During this specific period, the treatment of recent onset AF was mainly cardiology consultation, admission, and rate control. This timeframe was chosen, as it was just before any rollout or pilot of the rhythm control method. The protocol was developed in collaboration with all 19 cardiologists at our institution in conjunction with the emergency medicine physicians. The rollout took approximately 2 years, including educating all 12 emergency physicians and nurses on study design and specific inclusion criteria. Meetings, e-mails, and posters in the ED accomplished this. The group of ED physicians consisted of all board-certified ED physicians, and each physician enrolled at least 1 patient during the study period; many had never performed ED cardioversion for stable AF before this protocol.
was excluded. Other exclusion criteria were known ejection fraction (EF) less than 35%, prior coronary artery disease, obvious ischemia on electrocardiogram (ECG), evidence of Wolf Parkinson White syndrome (WPW), unknown duration of symptoms, primary diagnosis not AF, other reason to observe/admit patient, prior cerebrovascular accident (CVA) not on oral anticoagulation, febrile, any antiarrhythmic in the past 72 hours, concurrent QT-prolonging medications (left to discretion of treating provider), known QTc prolongation of 460 milliseconds, and any hospitalizations within the prior 3 months. Prior AF was not an exclusion criterion. Standard consent for procedural sedation and direct current cardioversion was obtained per hospital protocol. Cardiology consultation was not required, as this protocol was specifically designed to be ED driven. All discharged patients were ensured prompt outpatient cardiology follow-up. The protocol was in keeping with the current at the time American Heart Association (AHA)/American College of Cardiology guidelines from 2010 for the management of AF [7,8].
2.5. Specific study protocol All enrolled patients were given 1 g of procainamide intravenous over 1 hour as a standard infusion. Nurses were instructed to stop the infusion if the patient developed symptomatic hypotension that did not respond to a 500 mL bolus of normal saline or if the QRS widened more than 50% of baseline. The nurses obtained vital signs every 15 minutes and rhythm strips during the protocol. Once sinus rhythm was restored, the infusion was stopped, and they were discharged home. Patients who did not convert with procainamide alone were observed for 1-hour postinfusion and then offered ED synchronized cardioversion. The ED physician performed the procedural sedation and cardioversion, and propofol sedation was used. Standard postprocedural sedation protocol was followed, and patients were discharged home after successful cardioversion. If the patient did not convert either with procainamide or EC, the ED physician consulted cardiology for further management. No new medications were prescribed.
2.2. Development of AF protocol A specific order set for the protocol was developed in the computer order entry system to ensure standardization and safety. This was also used as a way to track all patients enrolled in the protocol in case a data collection sheet was not filled out. While we began enrolling patients during the pilot period, these patients were not ultimately included in the final study cohort. This pilot period spanned the 2 years before the start of the study. The historical control group was obtained by chart review by the primary and co-investigators. The chart review period was only 9 months because an a priori sample size was calculated at 46 patients, and we obtained 50 within that timeframe. Research ethics board approval was granted by the institutional hospital institutional review board with a decision to waive the need for informed consent, as this was considered one standard of care.
2.6. Follow-up methods for prospective cohort A 30-day phone call was made to each enrolled patient in the protocol to assess recurrent AF, adverse events, patient satisfaction, and promptness of outpatient cardiology follow-up. This phone call was made by one of the investigators and followed a standardized data collection form that was created a priori. There were 3 attempts to contact each patient via phone and e-mail. If this was unsuccessful, we then obtained some of the follow-up data from their chart, specifically their cardiology appointment after the ED visit. In addition, a standardized satisfaction survey was given to the 8 emergency physicians after study completion to assess their experience with adoption of the protocol.
2.3. Study population and demographics
2.7. Data collection for historical cohort
The ED is part of a nonteaching community hospital with an average ED annual census of 60 000. All emergency physicians were board certified, but only 1 had performed cardioversion for stable AF before initiation of this protocol.
This was compared with a pure historical control group. The rollout of this protocol took close to 2 years with various pilots that were started and stopped, while the protocol was revised. This explains the “pure” historical control group—how the ED was handling recent onset AF before any talk of ED cardioversion and thus was remote from the prospective cohort. The historical controls were consecutive patients abstracted from the charts who presented with recent onset AF (b 48 hours) to our ED. If timing of AF was unclear from the records, then patients were excluded from the protocol. Patients were included in the historical cohort if they met all of the study criteria, although the retrospective chart review methods for identifying historical controls could not identify every exclusionary criterion in every case.
2.4. Inclusion and exclusion criteria If patients met inclusion criteria, the emergency physician rendering care enrolled them in the study. We included consecutive patients presenting to the ED with AF less than 48 hours who did not have any of our specific exclusion criteria. The determination of recent onset AF was left to the provider's discretion; however, if there was any doubt, the patient
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
3
2.8. Outcome measures The primary outcome measure was admission rates. Secondary measures included conversion rates, ED LOS, and total hospital LOS. There is also 30-day descriptive data describing recurrent AF, patient satisfaction score, adverse events (transient ischemic attack, stroke, and death), and promptness of cardiology follow-up. Emergency department physicians and nurses collected data on an ongoing basis, using a standardized data collection sheet kept in the ED. Nurses and physicians received appropriate training on the protocol, inclusion criteria, and data collection. Data were entered into a confidential spreadsheet and deidentified. Emergency department LOS and hospital LOS were abstracted from the chart and reported in minutes. Thirty-day data were obtained by a call back system. The historical cohort was obtained by reviewing consecutive charts and abstracting data by the co-investigators. An initial sample of 10 charts was done concurrently to ensure data abstraction accuracy across investigators. 2.9. Data analysis and a priori sample size calculation An a priori sample size calculation based on an α of .05 and size effect of a 30% reduction in admission rates was calculated at 48 subjects in each arm before initiation of the study based on the study conducted by Jacoby et al [5]. Baseline characteristics were compared. Primary and secondary end points were compared between the 2 study groups and assessed for statistical significance. Descriptive statistics were provided for the 30-day data in the prospective group only. Continuous features were summarized with medians, interquartile ranges, and ranges; categorical features were summarized with frequency counts and percentages. Comparisons between patients in the historical and prospective cohorts were evaluated using Wilcoxon rank sum and χ 2 tests. Statistical analyses were performed using the SAS software package (SAS Institute, Cary, NC). All tests were 2 sided with α set at .05. 3. Results 3.1. Patient enrollment We enrolled 54 total patients into the prospective observational arm of the study. This compares with 50 patients in the historical control, all of which met inclusion criteria. Baseline characteristics of the 2 groups were similar in terms of age, median CHA2DS2VASc scores, and sex. Of the 384 patients who presented to the ED during the study period with a final diagnosis of AF, a total of 68 (17.4%) were eligible, and a total of 54 (14.4%) were enrolled. There were 14 patients who were not enrolled but were eligible. Reasons for not enrolling 11 of those patients were related to provider preference for rate control and cardiology management. No other specifics are known as to why they were not enrolled. The other 3 who were eligible but not enrolled got propafenone, a rhythm control strategy that was used occasionally before the initiation of this protocol (Fig. 1). Patients who were eligible but not enrolled were similar to the prospective cohort. There are no follow-up data on these patients (Fig. 2).
Fig. 1. Flow diagram of prospective study cohort.
3.3. Primary and secondary outcomes There were 4 (7%) patients of 54 in the prospective group who were admitted to the hospital, 3 of which were the patients who did not convert to sinus rhythm, and 1 that did convert but had aberrant conduction and an abnormal postconversion ECG (Table 3). In the historical group, 30 (60%) of the 50 patients were admitted to the hospital (P b .001); this did not include the patients who had transesophageal echocardiogram (TEE) cardioversion in the cardiac catheterization laboratory and discharged from there. This can also be stated as the difference in hospitalization proportions, 0.60 to 0.07 is equal to 0.53. A 95% confidence interval for this difference is 0.35 to 0.66. Of the 54 patients in the prospective group, 51 (94%) converted to sinus rhythm in the ED. Of these, 34 patients (67%) converted with procainamide alone, and the remaining 17 (33%) converted with EC. Time to conversion after initiation of the procainamide drip ranged from 10 to 120 minutes, as we allowed 1-hour postinfusion to convert. This compared with the historical group where 28 (56%) of 50 converted in the ED, all with medication or spontaneously, none with EC. Medications that were used in the historical group included only diltiazem and propafenone. The much higher admission rate accounts for the significant drop in hospital LOS for the prospective group, 274 vs 966 minutes (P b .001). The ED LOS was also significantly shorter in the prospective group, 269 vs 364 minutes (P b .01).
3.2. Baseline demographics 3.4. Emergency department adverse events The median CHA2DS2VASc score for the prospective group was 1 with 43% female and a median of 58.5 (Table 1). The presenting symptom was either “typical AF” or palpitations in all of the patients (Table 2). Patients were in AF between 30 minutes and 18 hours. Although we included patients in AF up to 48 hours, the longest period of AF was 18 hours. This was not intentional, as all patients with AF up to 48 hours could have been enrolled. One of these patients had atrial flutter as the presenting rhythm; all others had AF.
There were 2 patients who developed transient hypotension during procainamide infusion, which resolved with 500 mL bolus of normal saline (NS); and in both cases, the drip did not have to be discontinued. There were 2 patients who developed a wide complex rhythm during the infusion of procainamide, and in both cases the drip was discontinued, and patients were immediately electrically cardioverted without complication. The electrophysiologists who examined the
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
4
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
Fig. 2. Emergency department–driven AF protocol.
ECGs after the ED visit determined both rhythm changes to be aberrancy and not ventricular tachycardia. 3.5. Prospective cohort descriptive data All patients in the prospective group were called 30 days after initial presentation to ED. All 54 patients were followed up. Of those we could not reach by phone, follow-up data were collected based on chart review and cardiology office visit follow-up information. Of the 54 patients contacted, 7 had recurrent AF either recognized on Holter monitor, loop recorder, or subsequent symptomatic episode (Table 4). Of the 6 patients with recurrent AF, 5 returned to the ED in the following 30 days. There were no important adverse events in the study group, specifically death or CVA/transient ischemic attack. Of the 54 patients contacted, 52 had prompt cardiology follow-up determined to be less than 14 days from ED visit. There were 3 patients in the follow-up group who had ablation in the
following 30 days. Patient comments were overwhelmingly positive, and all were satisfied with their experience, particularly with their disposition to home. We used a scale of 1 to 10 to indicate their patient experience at the time of the ED visit, with 1 being “I would never return to that ED again” and 10 being “Was a fantastic experience.” The mean score was a 9.7 of 48 patients who completed the study, due to 1 outlier that was upset over being charged for a
Table 1 Baseline characteristics of prospective and historical cohort Characteristic
Historical controls (CI)
Study (CI)
Median age Male Median CHA2DS2VASc score
66.5 (56-71) 24 2 (1-3)
58.5 (48-67) 31 1 (0-2)
Abbreviation: CI, confidence interval.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx Table 2 Study cohort characteristics
5
Table 4 Thirty-day follow-up data from study cohort
Characteristic
n
Variable
No of patients, n = 54
Median age Male Presenting complaint Palpitations “Typical a-fib” Lightheadedness Chest pain Avg time in AF before presentation Median CHADS2VASC score 0 1 2 3 4 5 Converted in ED Converted with procainamide Converted with EC Complications in ED Transient hypotension Aberrant conduction Number contacted 30 d later Number followed up
58.5 (48-67; 24-86) 31 (57%)
Recurrent AF Return to the ED Stroke Death Medication complication, Coumadin Ablation Cardiology follow-up within 14 d
7 5 0 0 1 4 52
42 (77%) 9 (16%) 3 (6%) 1 (1%) 205 min 1 (0-2; 0-5) 18 (33%) 16 (30%) 9 (17%) 6 (11%) 3 (6%) 2 (4%) 51 (94%) 35 (68.6%) 16 (31.3%) 4 2 2 48 (89%) 54 (100%)
Abbreviations: a-fib, atrial fibrillation; Avg, average.
urine drug screen he did not think was indicated. Additionally, the Emergency physicians were polled regarding their experience (Table 5). All enrolling physicians were overwhelmingly satisfied with the protocol and answered “strongly agree” to the survey, with the exception of one physician that answered “agree”. 4. Discussion We successfully developed and implemented an expedited rhythm control method for patients with recent onset AF in a community hospital. This protocol was both safe (as far as can be judged in a trial of 54 patients) and efficacious. Stiell et al [4] report a decade-long experience of thousands of patients safely enrolled in a similar protocol in Canada. This is now adapted as standard of care for management of patients with recent onset AF. This is the first US study to incorporate this robust experience and challenge the status quo. Although much smaller in terms of patient data, the outcomes are quite similar. Historically, in our institution, patients presenting to the ED with symptomatic AF were treated in a myriad of ways. The predominate management strategy was one of prompt ED cardiology consultation and management per the cardiologist. This generally included rate control and either TEE-guided cardioversion or hospital admission. “The pill in the pocket” approach with attempted ED pharmacologic cardioversion was also used in selected patients. This was directed by cardiology, while the patient remained in the ED. From the standpoint of the emergency physician, the historical strategy was actually less risky. Other studies like this one reported the safety of ED cardioversion and highlighted the potential barriers to implement a practice change [9,10]. The barriers were the same in our study [11]. These include fear of adverse outcomes, specifically stroke and the ability of the ED physician to truly identify the low-risk patient [10]. This study
attempted to select a low-risk cohort of patients who were good candidates for an ED rhythm control method by using multiple exclusion criteria. A better risk predictor is the CHA2DS2VASc score [12,13]. This would be a simpler and more generalizable manner of ensuring a safe ED-driven protocol invoking a rhythm control method, consistent with the current AHA guidelines. Scheuermeyer et al [12] recently looked at how often the ED physician initiates anticoagulation in the ED in patients presenting with AF. This study separated patients according to CHADS2 scores and found that the emergency physician did not routinely offer anticoagulation despite a high CHADS2 score. Despite this, there was only 1 patient that had an ischemic stroke in the 30 days after ED visit. Another barrier to implementing such a strategy is the question of the provider's ability to know exactly when a patient went into AF. This was addressed in our study in that the physician was told to use the longest duration of AF obtained by any source at the time of the ED visit, and if duration was unknown or the patient did not ever feel their AF, they were automatically excluded. This represents a real-world patient encounter. The promptness of outpatient cardiology has been cited as an important predictor of poor outcomes in the patient with AF [14]. Specifically, prompt outpatient cardiology follow-up has been associated with decreased overall morbidity and mortality. The uniqueness of our study is the partnership between the ED physicians and cardiologists, highlighting the benefits of a team approach, which has been proven to improve patient care in many settings. The protocol was developed and implemented by both groups and demonstrated the feasibility of ensuring prompt outpatient cardiology follow-up even without cardiology involvement of care while in the ED. Given the predicted increase in prevalence of AF in the coming decades due to the aging US population, standardized treatment protocols that improve quality and reduce cost will be a necessity [1,15]. Although the primary goal was not to improve cost of care, the reduction in admission rate seen in our study will certainly reduce overall cost of treating AF in the ED if other groups can reproduce our results. Given that other studies suggest increased stroke risk in the 30 days after index AF visit to the ED, prompt follow-up and guideline-based initiation of oral anticoagulation should also reduce thromboembolic events [16,17].
Table 5 Physician satisfaction survey regarding AF protocol Key = 1, strongly agree; 2, agree; 3, slightly agree; 4, neutral; 5, slightly disagree; 6, disagree; 7, strongly disagree 1 2 3
Table 3 Comparison of historical cohort and prospective cohort for primary outcome measures Variable Patients converted In ED Patients admitted Mean ED LOS Mean total LOS
Historical cohort 28 (56%) 30 (60%) 344.5 (240-467) 966 (440-1631)
Study cohort 51 (94%) 4 (7%) 269.5 (213-336) 274 (213-350)
P b.001 b.001 .011 b.001
4
The AF protocol is better than what we were doing prior? 1 2 3 4 5 6 The AF protocol is safe for patients? 1 2 3 4 5 6 I am satisfied with the AF protocol that has been implemented 1 2 3 4 5 6 I would recommend this protocol to other EDs 1 2 3 4 5 6
7 7 7 7
Results of 8 physicians surveyed (100% response rate): Question 1 = mean score, 1.1 (1 physician answered 2). Question 2 = mean score, 1.1 (1 physician answered 2). Question 3 = mean score, 1. Question 4 = mean score, 1.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
6
J.L. White et al. / American Journal of Emergency Medicine xxx (2015) xxx–xxx
5. Limitations The primary limitation is the study design. This study was born out of a practice change within the institution during development of a hospital-wide AF center. It was designed intentionally to study the new AF protocol as a result of the development of this center. It would have been better to have a randomized controlled trial design. We obtained a historical control instead. Although our constellation of exclusion criteria could be cited as a barrier to enrolling more patients, this can also be viewed as a strength of the study. Reasons for excluding patients who were recently hospitalized or had coronary artery disease were based on cardiology recommendations. Cardiologists at our institution wanted to ensure a cohort of healthy, lowrisk patients. Ultimately, the protocol was revised to risk stratify based more on CHA2DS2-VASC scores. Although limiting patient enrollment, it used standard exclusion criteria that could be applied almost anywhere, as it was safely and effectively applied at a nonteaching hospital. We also noted the average length of AF was only 18 hours in our study cohort, although all patients with duration less than 48 hours could have been enrolled. The patients who were eligible but not enrolled fell within the mean of AF duration of the study cohort, and decision not to enroll them was not based on length of AF. This could be seen as a limitation, as 14 patients were eligible but not enrolled, mostly based on physician preference, and if they were included, it could have changed the data. We do not have outcomes or follow-up for those 14 patients and that would have been helpful. Our decision to use a standard dose of procainamide potentially underestimates the percentage of pharmacologic cardioversion in our study, as procainamide is often used as a weight-based drug. With a weight-based dosing, it is also possible that more arrhythmias may have occurred with higher overall dosing. We were unable to contact any of the historical group via phone per institutional review board guidelines. This limits our ability to subjectively compare 30-day outcomes with the historical group. We also did not use a standardized patient satisfaction scoring system, although we did specifically seek out patient experience via a 1 to 10 numerical scale. Prior studies have not looked at patient satisfaction for an ED-driven rhythm control method; and although our study did look at this, we did not use a standardized scale. Our study did not address the current anticoagulation guidelines as of this publication. The new AHA/American College of Cardiology/ CHEST guidelines were published after initiation of the study [9]. Our ongoing protocol was amended after the new guidelines were published to reflect change in oral anticoagulation guidelines. This new protocol was not used in any of the patients in this study. A potential limitation of the study was that it did not address the current anticoagulation guidelines that were published after the study was ended. The newest version of this protocol incorporates using CHA2DS2VASc scores to risk stratify those patients who leave the ED with oral anticoagulation. Our study did not offer oral anticoagulation to anyone.
6. Conclusions An expedited ED-driven rhythm control method resulted in decreased admission rates, ED, and total LOS. This protocol was easily adopted by community ED physicians and implemented without adverse events and resulted in a high conversion rate compared with
standard therapy. At 30 days, there were no strokes, and patient satisfaction was high. Our institution continues to implement this protocol. Acknowledgments Robert Slade, MD: enrollment of patients and support of study design and concept; Doylestown emergency physicians and nurses for enrollment of patients; Doylestown Hospital Cardiologists for support of study design and implementation; John Mitchell; Robert Sangrigoli, MD; and Stephen Sloan, MD for cardiology specialty consultation, support of study, and design of the AF protocol; Erik Hess, MD—support in preparing manuscript; Christine Lohse—preparing of the statistics; and Ian Stiell MD—permission to use a version of the Ottawa aggressive protocol and mentoring along the way.
References [1] Arco C, Martin A, Laguna P. Analysis of current management of atrial fibrillation in the acute setting: GEFAUR-1 Study. Ann Emerg Med 2005;46:424–30. [2] Besser K, Mills A. Is discharge to home after emergency department cardioversion safe for the treatment of recent-onset atrial fibrillation? Ann Emerg Med 2011;58: 517–20. [3] McDonald AJ, Pelletier AJ, Ellinor PT, Camargo Jr CA. Increasing US emergency department visits rates and subsequent hospital admissions for atrial fibrillation from 1993-2004. Ann Emerg Med 2008;51:58–65. [4] Stiell IG, Clement CM, Perry JJ, Vallancourt C, Symington C, Dickinson G, et al. Association of the Ottawa aggressive protocol with rapid discharge of emergency department patients with recent onset atrial fibrillation or flutter. Can J Emerg Med 2010; 12:181–91. [5] Jacoby JL, Cesta M, Heller MB, Salen P, Reed J. Synchronized emergency department cardioversion of atrial dysrhythmias saves time, money and resources. J Emerg Med 2005; 28:27–30. [6] Decker WW, Smars PA, Vaidyanathan L, Goyal DG, Boie ET, Stead LG, et al. A prospective, randomized trial of an emergency department observation unit for acute onset atrial fibrillation. In: Smars PA, VL, editors. Ann Emerg Med; 2008. p. 322–8. [7] Fuster V, Ryden LE, Cannom DS, Crijns HJ, Curtis AB, Ellenbogen KA, et al. ACC/AHA/ ESC 2006 guidelines for the management of patients with atrial fibrillation. Circulation 2006;114:700–52. [8] JE E, Lowe N, Estes M, group W. Cardiology Foundation/American Heart Association Task Force on practice guidelines on atrial fibrillation (updating the 2006 guidelines). A report of the American College of Cardiology; 2011. [9] January CT, Wann LS, Alpert JS, Calkins H, Cigarroa JE, Cleveland JC, et al. 2014 AHA/ ACC/HRS guideline for the management of patients with atrial fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on practice guidelines and the Heart Rhythm Society. J Am Coll Cardiol 2014;14: 01740–9. [10] Oishi ML, Xing S. Atrial fibrillation management strategies in the emergency department. Emerg Med Pract 2013;15:1–26. [11] Kirchhof P, Gunter B, Camm J, Crijns H. Improving outcomes in patients with atrial fibrillation: rationale and design of the early treatment of atrial fibrillation for stroke prevention trial. Am Heart J 2013;166:442–6. [12] Scheuermeyer FX, Innes G, Pourvali R, Dewitt C, Grafstein E, Heslop C, et al. Missed opportunities for appropriate anticoagulation among emergency department patients with uncomplicated atrial fibrillation or flutter. Ann Emerg Med 2013;62: 557–65. [13] Coppens M, Eikelboom JW, Hart RG, Yusuf S, Lip GY, Dorian P, et al. The CHA2DS2VASc score identifies those patients with atrial fibrillation and a CHADS2 score of 1 who are unlikely to benefit from oral anticoagulant therapy. Eur Heart J 2013; 34:170–6. [14] Atzema CL, Austin PC, Chong AS, Dorian P. Factors associated with 90 day death after emergency department discharge for atrial fibrillation. Ann Emerg Med 2013;61: 539–48. [15] Stead LG, Decer WW. Rhythm versus rate control for atrial fibrillation and flutter. Ann Emerg Med 2006;47:496–8. [16] Scheuermeyer FX, Grafstein E, Stenstrom R, Stenstrom R, Innes G, Heslop C, Macphee J, et al. Thirty-day and 1 year outcomes of emergency department patients with atrial fibrillation and no acute underlying cause. Ann Emerg Med 2012;60:755–65. [17] Swaminathan A. Recent onset atrial fibrillation: the case for emergency department cardioversion. Acad Emerg Med 2013;20:408.
Please cite this article as: White JL, et al, Performance of an expedited rhythm control method for recent onset atrial fibrillation in a community hospital, Am J Emerg Med (2015), http://dx.doi.org/10.1016/j.ajem.2015.03.059
