- Email: [email protected]
National use of the transesophageal echocardiographic-guided approach to cardioversion for patients in atrial fibrillation
National Use of the Transesophageal Echocardiographic-Guided Approach to Cardioversion for Patients in Atrial Fibrillation R. Daniel Murray, PhD, Ariel S. Goodman, BA, Elizabeth A. Lieber, BA, Susan E. Jasper, BSN, Richard A. Grimm, DO, Mario J. Garcia, MD, Deborah M. Miller, PhD, and Allan L. Klein, MD Transesophageal echocardiographic (TEE)-guided cardioversion of patients in atrial fibrillation (AF) of >2 days’ duration is used as an alternative to conventional therapy. The purpose of this study was to investigate practice patterns employed for stroke prophylaxis in patients with AF who underwent cardioversion, and to determine the relative use of conventional and TEEguided management strategies. We forwarded regionally stratified survey questionnaires to 947 clinical practices within the United States. The 10-question questionnaire queried demographic and clinical practice volumes and practices for managing patients with AF who underwent cardioversion. In addition, we used historical data to determine longitudinal use patterns of the TEE-guided approach for a large institution over 7 years. The 197 completed and returned surveys yielded a return rate of 20.8%. The TEE-guided approach was em-
ployed in approximately 12% of total cardioversions, but 75% of practices indicated that they employed transesophageal echocardiography only occasionally. The TEE-guided approach was associated with community size (r ⴝ 0.19; p <0.008), type of practice (r ⴝ 0.26; p ⴝ 0.001), total use of transesophageal echocardiography (r ⴝ 0.48; p <0.001), and volume of cardioversions (r ⴝ 0.28; p <0.001). Importantly, there was little consensus on the most appropriate clinical indications for TEE-guided cardioversions, and the proportions of TEE-guided cardioversion to total number of electrical cardioversions remained stable over 7 years. Practice volume and physician training may be the most important variables in the adoption of the TEE approach. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;85:239 –244)
ardioversion of patients from atrial fibrillation (AF) to normal sinus rhythm is frequently perC formed to relieve symptoms, improve cardiac perfor-
and allowing earlier cardioversion and a shorter duration of anticoagulation.4 – 6 Although the general perception that transesophageal echocardiographic (TEE) use in cardioversion is widespread and that TEE use is continuing to increase in popularity, we have postulated that the vast majority of medical centers employ tranesophageal echocardiography sparingly at best and that costs of technology and physician training limit the use of the TEE-guided approach in many institutions. The purpose of this study is to determine the current use of TEE-guided and conventional approaches to cardioversion among cardiologists practicing in widespread regions of the continental United States. The data presented will also outline demographic and professional variables that likely influenced the use of one approach versus the other. Trends for utilization of TEE-guided cardioversion are presented with data from a single institution.
mance, and decrease cardioembolic risk. However, the cardioversion procedure itself has an inherent risk of stroke, presumably due to embolization of extant thrombus or post-cardioversion thrombogenesis in the left atrium.1,2 To decrease this risk, patients with AF of ⬎2 days’ duration who undergo cardioversion are conventionally treated with therapeutic anticoagulation (international normalized ratio 2.0 to 3.0) for 3 weeks before and 4 weeks following cardioversion. This conventional practice follows the American College of Chest Physicians’ consensus statement on antithrombotic therapy in AF.3 As an alternative, transesophageal echocardiography has been employed to guide anticoagulation management in patients with AF by screening for thrombi From the Departments of Cardiology, Biostatistics and Epidemiology, and Neurology, The Cleveland Clinic Foundation, Cleveland, Ohio. This study was supported, in part, by an American Medical Association (AMA) Educational Research Grant, Chicago, Illinois; and an American Society of Echocardiography (ASE) Outcomes Research Grant, Raleigh, North Carolina. Manuscript received May 27, 1999; revised manuscript received and accepted August 20, 1999. Address for reprints: Allan L. Klein, MD, Department of Cardiology, Desk F15, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195. E-mail: [email protected]. ©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 85 January 15, 2000
METHODS To assess the prevalence of the TEE-guided approach in the United States, we surveyed clinical centers using a simple 10 question questionnaire (Figure 1). Clinical centers were selected from 3 membership directories: (1) the American College of Cardiology (ACC), (2) the North American Society of Pacing and Electrophysiology (NASPE), and (3) the 0002-9149/00/$–see front matter PII S0002-9149(99)00637-2
239
FIGURE 1. Survey instrument mailed to 947 United States clinical practices regarding the use of the TEE-guided approach to cardioversion of patients with AF of >2 days duration.
240 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
JANUARY 15, 2000
TABLE I Summary of Responses to 10-Question Survey on the Use of TEE-Guided Cardioversion from 197 Clinical Practices in the United States Frequency and Percent of Categorical Responses to Survey Questions Survey Question
Question Topic
1 2 3 4 5 6 7 8 9 10
Practice’s community size Type of practice Volume of TEE’s performed Volume of AF Volume of CV performed Proportion: inpatients/outpatients Proportion of CV without AC* Proportion of conventional CV† Proportion of TEE-guided CV‡ Indications for TEE-guided CV
Response a
Response b
Response c
Response d
Response e
Frequency Percent Frequency Percent Frequency Percent Frequency Percent Frequency Percent 3 141 16 60 3 20 125 3 50 39
1.5 71.9 8.1 30.5 1.5 10.2 63.5 1.5 25.4 20
12 34 55 113 107 69 46 1 72 68
6.1 17.3 27.9 57.4 54.3 35.2 23.4 0.5 36.5 34.9
37 13 81 17 74 62 12 2 28 80
18.8 6.6 41.1 8.6 37.6 31.6 6.1 1 14.2 41.2
75 8 34 7 11 20 8 7 22 69
38.1 4.1 17.3 3.6 5.6 10.2 4.1 3.6 11.2 35.4
70 NA 11 NA 2 25 11 184 25 36
35.5 NA 5.6 NA 1 12.8 5.6 93.4 12.7 13.3
*Mean ⫾ SD of respondent-provided values ⫽ N/A; †mean ⫾ SD of respondent-provided values ⫽ 89.9 ⫾ 13.8%; ‡mean ⫾ SD of respondent-supplied values ⫽ 47.5 ⫾ 23.1%. AC ⫽ anticoagulation; CV ⫽ cardioversion.
American Society of Echocardiography (ASE). Only physicians were selected as addressees and duplicates within clinical practices and between affiliated organization were eliminated. The selection process was performed by an independent technician unfamiliar with the practice patterns of the physicians. A selection goal was to survey as many group practices as possible over a widespread geographic region within each state using membership directories. Respondents were requested to answer the questions on behalf of their practice and not their individual preferences. Regions of the United States selected for sampling included the states of Alabama, California, Florida, Kansas, Ohio, New York, and Washington. In summary, 947 survey questionnaires were mailed using the US Post Office first-class service between September 1, and October 15, 1998. The survey questionnaire was mailed with a cover letter and an addressed, first-class postage-paid, return envelope. Survey questionnaires were mailed with a cover letter to clinical practices with assurances of anonymity. Anonymity was important to augment return rate and to improve the impartiality of responses. Only the respondent’s state (region of the United States) and membership affiliation were requested on the returned survey. In an additional effort to determine the temporal trend for frequency of use of the TEE-guided approach, we reviewed Cleveland Clinic Foundation historical information from 1991 to 1997. We specifically sought to determine the extent to which the volume of electrical cardioversions and the number of TEE-guided cardioversions changed over time in a large tertiary health care center. Oracle database records from the electrophysiology laboratory were accessed to obtain an absolute number of direct current cardioversions. Electrophysiology laboratory handwritten logbooks indicating the TEE-guided cardioversions were manually summarized. Patients participating in the Assessment of Cardioversion Using Transesophageal Echocardiography
(ACUTE) studies6,7 were excluded from summaries from this laboratory. Statistical analyses: To facilitate analysis (mean and 95% confidence intervals [CI]), interval midpoints from the categorical data were used when interval data were needed to determine prevalence of the approaches. Chi-square tests and Spearman’s correlation coefficients were used to assess and quantify associations for categorical data. Chi-square tests were used to compare response proportions among surveyed subgroups. All tests were 2-tailed. A p value of ⬍0.05 was considered statistically significant.
RESULTS
Survey response: We received 197 of 947 surveys (20.8%) submitted to different United States medical centers of various sizes. The returned and tallied surveys were anonymous and represented different types of clinical practices with varying patient volume. Table I presents the descriptive data (questions 1 through 10 consisting of categorical and continuous data) summarized from the clinical practices responding to the survey. There was no significant difference in the response rate from different regions (states) of the United States (Table II), and therefore, the respondent’s geographic region had no effect on the likelihood of response to this survey. Conversely, we found differences in response rates when respondents were pooled by their affiliation with ACC, NASPE, or ASE (Table III). ACC and ASE affiliates were more likely to respond to the survey compared with the NASPE affiliates (20.3%, 26.0%, and 14.2% respectively; p ⫽ 0.001).
Prevalence of the transesophageal echocardiographic-guided approach: Survey questions 7 to 9 addressed
the frequency of various approaches in the management of patients in AF of ⬎2 days’ duration who required cardioversion. These proportions represented midpoint estimates based on respondent-supplied “percentage intervals” for questions 7 to 9. Because the data from questions on prevalence represented a percentage of cardioversions from the respondents’
METHODS/NATIONAL USE OF TEE-GUIDED CARDIOVERSION
241
ences in absolute volume of cardioversions (question 5) among the respondents, the proportion of TEESurveys Mailed Surveys Returned guided cardioversions relative to the Response Rate total number of cardioversions was No. (%) No. (%) to Survey 12.1% (95% CI 11.6 to 12.6%) (FigAlabama 58 (6.1) 12 (6.1) 20.7% ure 2). California 234 (24.7) 51 (25.9) 21.8% Likewise, an estimate of the prevFlorida 169 (17.8) 25 (12.7) 14.8% Kansas 48 (5.1) 8 (4.1) 16.7% alence of cardioversions using the New York 212 (22.5) 49 (24.9) 23.1% conventional approach was reOhio 150 (15.8) 33 (16.8) 22.0% quested in question 8. The unadWashington 76 (8.0) 19 (9.6) 25.0% justed mean prevalence of the conTotals 947 (100) 197 (100) 20.8% ventional approach to cardioversion p ⫽ NS for all comparisons using chi-square test. at these sampled practices was 84.5% (95% CI 81.0 to 87.9%). By adjusting for the differences in absoTABLE III Membership Association for Mailing and Receipt of 10-Question Survey lute number of cardioversions (queson TEE Guided Cardioversion tion 5) among the respondents, the Surveys Mailed Surveys Returned proportion of conventional cardioResponse Rate versions relative to the total number No. (%) No. (%) to Survey of cardioversion procedures was ACC 439 (46.4) 89 (45.2) 20.3% 78% (95% CI 77.3 to 78.6%) (Figure NASPE 204 (21.5) 29 (14.7)* 14.2% 2). ASE 304 (32.1) 79 (40.1) 26.0% Finally, question 7 requested an Totals 947 (100) 197 (100) 20.8% estimate of the percentage of patients *p ⫽ 0.001 for return rate of surveys from NASPE versus ACC and ASE membership associations with AF of ⬎2 days’ duration who using chi-square test. were cardioverted without prior anticoagulation. The unadjusted mean was 2.9% (95% CI 1.7 to 4.0%). However, the mean response to question 7, adjusted for differences in cardioversion volume among respondents (question 5), showed that 3.7% of total patients (95% CI 3.5 to 4.0%) represented by this survey were cardioverted without anticoagulation (Figure 2). Importantly, the responses to questions 7 to 9 were complementary and could approximate unity. The sum of the unadjusted means for the 3 questions closely approached unity at 96.7%, and thus reflected a good measurement of the comprehension of the exclusivity of these 3 questions by the 197 respondents. Only 50 (25.4%) of the 197 respondents indicated response “a” (i.e., zero) to question 9, indicating that they never employed the TEE-guided approach to cardioversion for any indication. Hence, 147 of the respondents (74.6%) at least occasionally used the TEE-guided approach in their practice. Notably, 72 of FIGURE 2. The relative volume of the TEE-guided approach to the 147 respondents (49%) employed the TEE-guided cardioversion, the conventional approach to cardioversion, and approach in only 1% to 5% of all cardioversions. cardioversion without prior anticoagulation in the management of patients with AF undergoing cardioversion. Data represent Figure 3 represents the use the TEE-guided approach proportions of these approaches (questions 7 to 9) adjusted for in the 197 surveyed practices. TABLE II Geographic Distribution for Mailing and Receipt of 10-Question Survey on TEE-Guided Cardioversion
differences in volume of cardioversion (question 5) reported by the 197 respondents to the survey. No AC ⴝ no anticoagulation.
institution, these data have been adjusted for the practices’ estimated cardioversion volume provided in question 5. A graph representing the volume-adjusted means is presented as Figure 2. The estimated percent of TEE-guided cardioversion at the respondent’s institution was requested by question 9. The unadjusted mean of this response was 9.3% (95% CI 6.7 to 11.6). By adjusting for differ242 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
Factors influencing the prevalence of the TEE guided approach: Respondents’ community size (question 1),
type of practice or clinical setting (question 2), and total institutional use of both tranesophageal echocardiography (question 3) and cardioversion (question 4) all may have influenced employment of the TEE approach (questions 8 and 9). There was a statistically significant positive correlation between the proportion of the TEE-guided approach and community size (r ⫽ 0.19; p ⬍0.008), type of practice (r ⫽ 0.26; p ⫽ 0.001), total volume of TEE (r ⫽ 0.48; p ⬍0.001), and JANUARY 15, 2000
FIGURE 3. The proportion of clinical practices that never used the TEE-guided approach (50 of 197 or 25.4%), and those clinical practices using the TEE-guided approach at least occasionally (147 of 197 or 74.6%). Of the clinical practices using the TEEguided approach at least occasionally, approximately one-half use the TEE-guided approach in >5% of total cardioversions.
volume of cardioversions (r ⫽ 0.28; p ⬍0.001). These data show that institutional size and volume as well as the degree of academic and/or research activity at an institution may ultimately influence patient management. Among the 7 states selected for this survey, there were no differences in proportion of TEE-guided cardioversions. Likewise, there was no significant difference in the employment of the TEE-guided approach shown among the respondents from ACC, NASPE, and ASE. In summary, neither geographic region nor membership affiliation of the respondent influenced the employment of TEE-guided approach to cardioversion. Management of atrial fibrillation as inpatient versus outpatient: There was considerable variation among
the reporting clinical practices in the ratio of patients treated as outpatient versus inpatient (question 6) (Table I). Region of the country, community population size, or type of institution had no affect on whether patients were typically managed as inpatient versus outpatient. However, institutions reporting a prevalence of ⬎5% for the TEE-guided approach (question 9; responses c to e) had a significantly more heterogenous “mix” of inpatients and outpatients (question 6; responses b to d) (p ⬍0.004).
Clinical indications for the TEE-guided approach:
Physician response to the question concerning the circumstances under which tranesophageal echocardiography was prescribed (question 10) varied widely. Fifty-nine of the 197 respondents took the liberty of responding to ⬎1 answer for question 10. Of the 197 respondents, 68 (34.5%) indicated that the TEEguided approach was preferred for low-risk patients, and 80 respondents (40.6%) indicated that the TEEguided approach was preferred for high-risk patients (Table I). Likewise, we received different responses to the inpatient and/or outpatient segment of question 10. Of the 197 respondents, 67 (34.0%) indicated that an inpatient was an appropriate circumstance for the
TEE-guided approach, whereas 24 (12.2%) of the respondents indicated that an outpatient was the preferred circumstance (Table I). Temporal changes in clinical practice: The frequency of total direct current (electrical) cardioversions and total TEE-guided electrical cardioversions have been followed at the Cleveland Clinic for the last several years. Our data show that the number of electrical cardioversions has increased steadily since 1991. In fact, between 1995 and 1997, the number of electrical cardioversions increased from 633 to 937 per year, emphasizing the popularity of the procedure as a therapy for restoring sinus rhythm. Correspondingly, the total number of TEE-guided cardioversions increased at approximately the same rate. However, the proportion of TEE-guided cardioversions to total cardioversions has remained quite constant, ranging from 12% to 14% over the last 7 years (Figure 4).
DISCUSSION The data from this survey suggest a modest national use (12%) of the TEE-guided approach of all cardioversion procedures performed, although the accumulated data suggest that comparable efficacy and possible early cardioversion benefit when TEE is employed.5,6,8 The data from this survey also reveal that nearly 75% of all clinical centers responding used the TEE-guided approach at least occasionally. Thus, these data suggest that the relatively modest use of the TEE-guided approach is not a result of physicians’ lack of familiarity of the procedure. The use the TEE approach was relatively low in clinical private practices and in smaller communities. Conversely, the highest use of the TEE-guided approach was found at academic institutions and tertiary referral centers. Notably, the employment of the TEEguided approach was positively correlated with total volume of TEE procedures performed at the institution. Hence, institutions that performed a larger number of TEE procedures were also more likely to perform TEE-guided cardioversions. These data suggest that the employment of the TEE-guided approach may be dependent, at least to some degree, on physician preference and training, as well as the availability of a properly equipped echocardiographic laboratory. Other potential contributing factors such as patient and physician acceptance and/or convenience, or subspecialty interest and/or preference were not assessed by this survey. Our data also indicated that the use of TEE-guided cardioversion with short-term anticoagulation of AF varied widely among institutions. Conversely, trends from the Cleveland Clinic showed little variability over time, although there have been changes in personnel and patient volume. Although the volume of electrical cardioversion and the volume of transesophageal echocardiographies have increased steadily over the last 7 years, the proportion of TEE-guided cardioversions to total electrical cardioversions have remained relatively stable. Which patients should undergo TEE-guided cardioversion? Physicians from 197 United States medi-
METHODS/NATIONAL USE OF TEE-GUIDED CARDIOVERSION
243
TEE screening before cardioversion. In addition, inpatients and outpatients were selected as indications for prescribing the TEE-guided approach to cardioversion (Table I). The lack of consensus on question 10 may also be due to the respondents’ problem of reducing a complex clinical issue (appropriate indication for TEE-guided cardioversion) to a simple catagorical response. We await the results of the ACUTE Multicenter Study to help answer these questions.7 Study limitations: The primary limitation of this survey is that sampled clinical practices were not selected in a truly random fashion. However, the selection of names from membership directories was performed “blinded” to the clinical practices of the physician. Names representing clinical practices of varying size and complexity were chosen with an interest in obtaining a state-wide geographic distribution of practices within the membership lists. The bias in selection of names and clinical practices was consequently negligible. The second limitation of this study is that the data obtained from the surveys reflect the physicians’ understanding of their own clinical practice rather than actual data derived from the clinical practice. Consequently, physicians were requested to provide not just a summary of their own clinical preferences, but rather provide a summary of baseline volume and management strategies of their entire clinical practice. Finally, the 3 membership groupings did not respond to the survey with the same frequency. NASPE members were under-represented in this sampling. However, this may have had little effect on the reported proportions of TEE-guided cardioversions because we found no difference among membership lists in the reported use of TEE-guided cardioversion. 1. Goldman MJ. The management of chronic atrial fibrillation. Prog Cardiovasc
Dis 1960;2:465– 479. 2. Grimm RA, Stewart WJ, Maloney JD, Pearce GL, Salcedo EE. Impact of
FIGURE 4. Direct current cardioversion (DCC) procedures and TEE-guided procedures completed at the Cleveland Clinic Foundation from 1991 to 1997. (A) Total volume of DCCs performed. (B) Total volume of TEE-guided DCCs performed. (C) The proportion of TEE-guided DCC to total DCC volume. Note that the volume of both DCC procedures and TEE-guided DCC procedures dramatically increased over 7 years, but the proportion of TEE-guided DCC to total DCC did not change markedly.
cal centers showed great variation in response to question 10. There was a comparable frequency of response for “b) low-risk patients” and “c) high–risk patients,” indicating the lack of a clear indication for TEE-guided cadioversion, or at least lack of consensus on the most appropriate patient population for
244 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
electrical cardioversion for atrial fibrillation on left atrial appendage function and spontaneous echo contrast: characterization by simultaneous transesophageal echocardiography. J Am Coll Cardiol 1993;22:1359 –1366. 3. Laupacis A, Albers G, Dalen J, Dunn MI, Jacobson A, Singer, DE. Antithrombotic therapy in atrial fibrillation. Chest 1998;114:579S-589S. 4. Manning WJ, Silverman DI, Gordon SP, Krumholz HM, Douglas PS. Cardioversion from atrial fibrillation without prolonged anticoagulation with use of transesophageal echocardiography to exclude the presence of atrial thrombi. New Engl J Med 1993;328:750 –755. 5. Manning WJ, Silverman DI, Keighley CS, Oettgen P, Douglas PS. Transesophageal echocardiographically facilitated early cardioversion from atrial fibrillation using short-term anticoagulation: final results of a prospective 4.5-year study. J Am Coll Cardiol 1995;25:1354 –1361. 6. Klein AL, Grimm RA, Black IW, Leung DY, Chung MK, Vaughn SE, Murray RD, Miller D, Arheart KL. Cardioversion guided by transesophageal echocardiography: the ACUTE Pilot Study. A randomized, controlled trial. Ann Intern Med 1997;126:200 –209. 7. ACUTE Steering and Publications Committees. Design of a clinical trial for the Assessment of Cardioversion Using Transesophageal Echocardiography (the ACUTE Multicenter Study). Am J Cardiol 1998;81:877– 883. 8. Stoddard MF, Dawkins PR, Prince CR, Longaker RA. Transesophageal echocardiographic guidance of cardioversion in patients with atrial fibrillation. Am Heart J 1995;129:1204 –1215.
JANUARY 15, 2000
ployed in approximately 12% of total cardioversions, but 75% of practices indicated that they employed transesophageal echocardiography only occasionally. The TEE-guided approach was associated with community size (r ⴝ 0.19; p <0.008), type of practice (r ⴝ 0.26; p ⴝ 0.001), total use of transesophageal echocardiography (r ⴝ 0.48; p <0.001), and volume of cardioversions (r ⴝ 0.28; p <0.001). Importantly, there was little consensus on the most appropriate clinical indications for TEE-guided cardioversions, and the proportions of TEE-guided cardioversion to total number of electrical cardioversions remained stable over 7 years. Practice volume and physician training may be the most important variables in the adoption of the TEE approach. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;85:239 –244)
ardioversion of patients from atrial fibrillation (AF) to normal sinus rhythm is frequently perC formed to relieve symptoms, improve cardiac perfor-
and allowing earlier cardioversion and a shorter duration of anticoagulation.4 – 6 Although the general perception that transesophageal echocardiographic (TEE) use in cardioversion is widespread and that TEE use is continuing to increase in popularity, we have postulated that the vast majority of medical centers employ tranesophageal echocardiography sparingly at best and that costs of technology and physician training limit the use of the TEE-guided approach in many institutions. The purpose of this study is to determine the current use of TEE-guided and conventional approaches to cardioversion among cardiologists practicing in widespread regions of the continental United States. The data presented will also outline demographic and professional variables that likely influenced the use of one approach versus the other. Trends for utilization of TEE-guided cardioversion are presented with data from a single institution.
mance, and decrease cardioembolic risk. However, the cardioversion procedure itself has an inherent risk of stroke, presumably due to embolization of extant thrombus or post-cardioversion thrombogenesis in the left atrium.1,2 To decrease this risk, patients with AF of ⬎2 days’ duration who undergo cardioversion are conventionally treated with therapeutic anticoagulation (international normalized ratio 2.0 to 3.0) for 3 weeks before and 4 weeks following cardioversion. This conventional practice follows the American College of Chest Physicians’ consensus statement on antithrombotic therapy in AF.3 As an alternative, transesophageal echocardiography has been employed to guide anticoagulation management in patients with AF by screening for thrombi From the Departments of Cardiology, Biostatistics and Epidemiology, and Neurology, The Cleveland Clinic Foundation, Cleveland, Ohio. This study was supported, in part, by an American Medical Association (AMA) Educational Research Grant, Chicago, Illinois; and an American Society of Echocardiography (ASE) Outcomes Research Grant, Raleigh, North Carolina. Manuscript received May 27, 1999; revised manuscript received and accepted August 20, 1999. Address for reprints: Allan L. Klein, MD, Department of Cardiology, Desk F15, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, Ohio 44195. E-mail: [email protected]. ©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 85 January 15, 2000
METHODS To assess the prevalence of the TEE-guided approach in the United States, we surveyed clinical centers using a simple 10 question questionnaire (Figure 1). Clinical centers were selected from 3 membership directories: (1) the American College of Cardiology (ACC), (2) the North American Society of Pacing and Electrophysiology (NASPE), and (3) the 0002-9149/00/$–see front matter PII S0002-9149(99)00637-2
239
FIGURE 1. Survey instrument mailed to 947 United States clinical practices regarding the use of the TEE-guided approach to cardioversion of patients with AF of >2 days duration.
240 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
JANUARY 15, 2000
TABLE I Summary of Responses to 10-Question Survey on the Use of TEE-Guided Cardioversion from 197 Clinical Practices in the United States Frequency and Percent of Categorical Responses to Survey Questions Survey Question
Question Topic
1 2 3 4 5 6 7 8 9 10
Practice’s community size Type of practice Volume of TEE’s performed Volume of AF Volume of CV performed Proportion: inpatients/outpatients Proportion of CV without AC* Proportion of conventional CV† Proportion of TEE-guided CV‡ Indications for TEE-guided CV
Response a
Response b
Response c
Response d
Response e
Frequency Percent Frequency Percent Frequency Percent Frequency Percent Frequency Percent 3 141 16 60 3 20 125 3 50 39
1.5 71.9 8.1 30.5 1.5 10.2 63.5 1.5 25.4 20
12 34 55 113 107 69 46 1 72 68
6.1 17.3 27.9 57.4 54.3 35.2 23.4 0.5 36.5 34.9
37 13 81 17 74 62 12 2 28 80
18.8 6.6 41.1 8.6 37.6 31.6 6.1 1 14.2 41.2
75 8 34 7 11 20 8 7 22 69
38.1 4.1 17.3 3.6 5.6 10.2 4.1 3.6 11.2 35.4
70 NA 11 NA 2 25 11 184 25 36
35.5 NA 5.6 NA 1 12.8 5.6 93.4 12.7 13.3
*Mean ⫾ SD of respondent-provided values ⫽ N/A; †mean ⫾ SD of respondent-provided values ⫽ 89.9 ⫾ 13.8%; ‡mean ⫾ SD of respondent-supplied values ⫽ 47.5 ⫾ 23.1%. AC ⫽ anticoagulation; CV ⫽ cardioversion.
American Society of Echocardiography (ASE). Only physicians were selected as addressees and duplicates within clinical practices and between affiliated organization were eliminated. The selection process was performed by an independent technician unfamiliar with the practice patterns of the physicians. A selection goal was to survey as many group practices as possible over a widespread geographic region within each state using membership directories. Respondents were requested to answer the questions on behalf of their practice and not their individual preferences. Regions of the United States selected for sampling included the states of Alabama, California, Florida, Kansas, Ohio, New York, and Washington. In summary, 947 survey questionnaires were mailed using the US Post Office first-class service between September 1, and October 15, 1998. The survey questionnaire was mailed with a cover letter and an addressed, first-class postage-paid, return envelope. Survey questionnaires were mailed with a cover letter to clinical practices with assurances of anonymity. Anonymity was important to augment return rate and to improve the impartiality of responses. Only the respondent’s state (region of the United States) and membership affiliation were requested on the returned survey. In an additional effort to determine the temporal trend for frequency of use of the TEE-guided approach, we reviewed Cleveland Clinic Foundation historical information from 1991 to 1997. We specifically sought to determine the extent to which the volume of electrical cardioversions and the number of TEE-guided cardioversions changed over time in a large tertiary health care center. Oracle database records from the electrophysiology laboratory were accessed to obtain an absolute number of direct current cardioversions. Electrophysiology laboratory handwritten logbooks indicating the TEE-guided cardioversions were manually summarized. Patients participating in the Assessment of Cardioversion Using Transesophageal Echocardiography
(ACUTE) studies6,7 were excluded from summaries from this laboratory. Statistical analyses: To facilitate analysis (mean and 95% confidence intervals [CI]), interval midpoints from the categorical data were used when interval data were needed to determine prevalence of the approaches. Chi-square tests and Spearman’s correlation coefficients were used to assess and quantify associations for categorical data. Chi-square tests were used to compare response proportions among surveyed subgroups. All tests were 2-tailed. A p value of ⬍0.05 was considered statistically significant.
RESULTS
Survey response: We received 197 of 947 surveys (20.8%) submitted to different United States medical centers of various sizes. The returned and tallied surveys were anonymous and represented different types of clinical practices with varying patient volume. Table I presents the descriptive data (questions 1 through 10 consisting of categorical and continuous data) summarized from the clinical practices responding to the survey. There was no significant difference in the response rate from different regions (states) of the United States (Table II), and therefore, the respondent’s geographic region had no effect on the likelihood of response to this survey. Conversely, we found differences in response rates when respondents were pooled by their affiliation with ACC, NASPE, or ASE (Table III). ACC and ASE affiliates were more likely to respond to the survey compared with the NASPE affiliates (20.3%, 26.0%, and 14.2% respectively; p ⫽ 0.001).
Prevalence of the transesophageal echocardiographic-guided approach: Survey questions 7 to 9 addressed
the frequency of various approaches in the management of patients in AF of ⬎2 days’ duration who required cardioversion. These proportions represented midpoint estimates based on respondent-supplied “percentage intervals” for questions 7 to 9. Because the data from questions on prevalence represented a percentage of cardioversions from the respondents’
METHODS/NATIONAL USE OF TEE-GUIDED CARDIOVERSION
241
ences in absolute volume of cardioversions (question 5) among the respondents, the proportion of TEESurveys Mailed Surveys Returned guided cardioversions relative to the Response Rate total number of cardioversions was No. (%) No. (%) to Survey 12.1% (95% CI 11.6 to 12.6%) (FigAlabama 58 (6.1) 12 (6.1) 20.7% ure 2). California 234 (24.7) 51 (25.9) 21.8% Likewise, an estimate of the prevFlorida 169 (17.8) 25 (12.7) 14.8% Kansas 48 (5.1) 8 (4.1) 16.7% alence of cardioversions using the New York 212 (22.5) 49 (24.9) 23.1% conventional approach was reOhio 150 (15.8) 33 (16.8) 22.0% quested in question 8. The unadWashington 76 (8.0) 19 (9.6) 25.0% justed mean prevalence of the conTotals 947 (100) 197 (100) 20.8% ventional approach to cardioversion p ⫽ NS for all comparisons using chi-square test. at these sampled practices was 84.5% (95% CI 81.0 to 87.9%). By adjusting for the differences in absoTABLE III Membership Association for Mailing and Receipt of 10-Question Survey lute number of cardioversions (queson TEE Guided Cardioversion tion 5) among the respondents, the Surveys Mailed Surveys Returned proportion of conventional cardioResponse Rate versions relative to the total number No. (%) No. (%) to Survey of cardioversion procedures was ACC 439 (46.4) 89 (45.2) 20.3% 78% (95% CI 77.3 to 78.6%) (Figure NASPE 204 (21.5) 29 (14.7)* 14.2% 2). ASE 304 (32.1) 79 (40.1) 26.0% Finally, question 7 requested an Totals 947 (100) 197 (100) 20.8% estimate of the percentage of patients *p ⫽ 0.001 for return rate of surveys from NASPE versus ACC and ASE membership associations with AF of ⬎2 days’ duration who using chi-square test. were cardioverted without prior anticoagulation. The unadjusted mean was 2.9% (95% CI 1.7 to 4.0%). However, the mean response to question 7, adjusted for differences in cardioversion volume among respondents (question 5), showed that 3.7% of total patients (95% CI 3.5 to 4.0%) represented by this survey were cardioverted without anticoagulation (Figure 2). Importantly, the responses to questions 7 to 9 were complementary and could approximate unity. The sum of the unadjusted means for the 3 questions closely approached unity at 96.7%, and thus reflected a good measurement of the comprehension of the exclusivity of these 3 questions by the 197 respondents. Only 50 (25.4%) of the 197 respondents indicated response “a” (i.e., zero) to question 9, indicating that they never employed the TEE-guided approach to cardioversion for any indication. Hence, 147 of the respondents (74.6%) at least occasionally used the TEE-guided approach in their practice. Notably, 72 of FIGURE 2. The relative volume of the TEE-guided approach to the 147 respondents (49%) employed the TEE-guided cardioversion, the conventional approach to cardioversion, and approach in only 1% to 5% of all cardioversions. cardioversion without prior anticoagulation in the management of patients with AF undergoing cardioversion. Data represent Figure 3 represents the use the TEE-guided approach proportions of these approaches (questions 7 to 9) adjusted for in the 197 surveyed practices. TABLE II Geographic Distribution for Mailing and Receipt of 10-Question Survey on TEE-Guided Cardioversion
differences in volume of cardioversion (question 5) reported by the 197 respondents to the survey. No AC ⴝ no anticoagulation.
institution, these data have been adjusted for the practices’ estimated cardioversion volume provided in question 5. A graph representing the volume-adjusted means is presented as Figure 2. The estimated percent of TEE-guided cardioversion at the respondent’s institution was requested by question 9. The unadjusted mean of this response was 9.3% (95% CI 6.7 to 11.6). By adjusting for differ242 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
Factors influencing the prevalence of the TEE guided approach: Respondents’ community size (question 1),
type of practice or clinical setting (question 2), and total institutional use of both tranesophageal echocardiography (question 3) and cardioversion (question 4) all may have influenced employment of the TEE approach (questions 8 and 9). There was a statistically significant positive correlation between the proportion of the TEE-guided approach and community size (r ⫽ 0.19; p ⬍0.008), type of practice (r ⫽ 0.26; p ⫽ 0.001), total volume of TEE (r ⫽ 0.48; p ⬍0.001), and JANUARY 15, 2000
FIGURE 3. The proportion of clinical practices that never used the TEE-guided approach (50 of 197 or 25.4%), and those clinical practices using the TEE-guided approach at least occasionally (147 of 197 or 74.6%). Of the clinical practices using the TEEguided approach at least occasionally, approximately one-half use the TEE-guided approach in >5% of total cardioversions.
volume of cardioversions (r ⫽ 0.28; p ⬍0.001). These data show that institutional size and volume as well as the degree of academic and/or research activity at an institution may ultimately influence patient management. Among the 7 states selected for this survey, there were no differences in proportion of TEE-guided cardioversions. Likewise, there was no significant difference in the employment of the TEE-guided approach shown among the respondents from ACC, NASPE, and ASE. In summary, neither geographic region nor membership affiliation of the respondent influenced the employment of TEE-guided approach to cardioversion. Management of atrial fibrillation as inpatient versus outpatient: There was considerable variation among
the reporting clinical practices in the ratio of patients treated as outpatient versus inpatient (question 6) (Table I). Region of the country, community population size, or type of institution had no affect on whether patients were typically managed as inpatient versus outpatient. However, institutions reporting a prevalence of ⬎5% for the TEE-guided approach (question 9; responses c to e) had a significantly more heterogenous “mix” of inpatients and outpatients (question 6; responses b to d) (p ⬍0.004).
Clinical indications for the TEE-guided approach:
Physician response to the question concerning the circumstances under which tranesophageal echocardiography was prescribed (question 10) varied widely. Fifty-nine of the 197 respondents took the liberty of responding to ⬎1 answer for question 10. Of the 197 respondents, 68 (34.5%) indicated that the TEEguided approach was preferred for low-risk patients, and 80 respondents (40.6%) indicated that the TEEguided approach was preferred for high-risk patients (Table I). Likewise, we received different responses to the inpatient and/or outpatient segment of question 10. Of the 197 respondents, 67 (34.0%) indicated that an inpatient was an appropriate circumstance for the
TEE-guided approach, whereas 24 (12.2%) of the respondents indicated that an outpatient was the preferred circumstance (Table I). Temporal changes in clinical practice: The frequency of total direct current (electrical) cardioversions and total TEE-guided electrical cardioversions have been followed at the Cleveland Clinic for the last several years. Our data show that the number of electrical cardioversions has increased steadily since 1991. In fact, between 1995 and 1997, the number of electrical cardioversions increased from 633 to 937 per year, emphasizing the popularity of the procedure as a therapy for restoring sinus rhythm. Correspondingly, the total number of TEE-guided cardioversions increased at approximately the same rate. However, the proportion of TEE-guided cardioversions to total cardioversions has remained quite constant, ranging from 12% to 14% over the last 7 years (Figure 4).
DISCUSSION The data from this survey suggest a modest national use (12%) of the TEE-guided approach of all cardioversion procedures performed, although the accumulated data suggest that comparable efficacy and possible early cardioversion benefit when TEE is employed.5,6,8 The data from this survey also reveal that nearly 75% of all clinical centers responding used the TEE-guided approach at least occasionally. Thus, these data suggest that the relatively modest use of the TEE-guided approach is not a result of physicians’ lack of familiarity of the procedure. The use the TEE approach was relatively low in clinical private practices and in smaller communities. Conversely, the highest use of the TEE-guided approach was found at academic institutions and tertiary referral centers. Notably, the employment of the TEEguided approach was positively correlated with total volume of TEE procedures performed at the institution. Hence, institutions that performed a larger number of TEE procedures were also more likely to perform TEE-guided cardioversions. These data suggest that the employment of the TEE-guided approach may be dependent, at least to some degree, on physician preference and training, as well as the availability of a properly equipped echocardiographic laboratory. Other potential contributing factors such as patient and physician acceptance and/or convenience, or subspecialty interest and/or preference were not assessed by this survey. Our data also indicated that the use of TEE-guided cardioversion with short-term anticoagulation of AF varied widely among institutions. Conversely, trends from the Cleveland Clinic showed little variability over time, although there have been changes in personnel and patient volume. Although the volume of electrical cardioversion and the volume of transesophageal echocardiographies have increased steadily over the last 7 years, the proportion of TEE-guided cardioversions to total electrical cardioversions have remained relatively stable. Which patients should undergo TEE-guided cardioversion? Physicians from 197 United States medi-
METHODS/NATIONAL USE OF TEE-GUIDED CARDIOVERSION
243
TEE screening before cardioversion. In addition, inpatients and outpatients were selected as indications for prescribing the TEE-guided approach to cardioversion (Table I). The lack of consensus on question 10 may also be due to the respondents’ problem of reducing a complex clinical issue (appropriate indication for TEE-guided cardioversion) to a simple catagorical response. We await the results of the ACUTE Multicenter Study to help answer these questions.7 Study limitations: The primary limitation of this survey is that sampled clinical practices were not selected in a truly random fashion. However, the selection of names from membership directories was performed “blinded” to the clinical practices of the physician. Names representing clinical practices of varying size and complexity were chosen with an interest in obtaining a state-wide geographic distribution of practices within the membership lists. The bias in selection of names and clinical practices was consequently negligible. The second limitation of this study is that the data obtained from the surveys reflect the physicians’ understanding of their own clinical practice rather than actual data derived from the clinical practice. Consequently, physicians were requested to provide not just a summary of their own clinical preferences, but rather provide a summary of baseline volume and management strategies of their entire clinical practice. Finally, the 3 membership groupings did not respond to the survey with the same frequency. NASPE members were under-represented in this sampling. However, this may have had little effect on the reported proportions of TEE-guided cardioversions because we found no difference among membership lists in the reported use of TEE-guided cardioversion. 1. Goldman MJ. The management of chronic atrial fibrillation. Prog Cardiovasc
Dis 1960;2:465– 479. 2. Grimm RA, Stewart WJ, Maloney JD, Pearce GL, Salcedo EE. Impact of
FIGURE 4. Direct current cardioversion (DCC) procedures and TEE-guided procedures completed at the Cleveland Clinic Foundation from 1991 to 1997. (A) Total volume of DCCs performed. (B) Total volume of TEE-guided DCCs performed. (C) The proportion of TEE-guided DCC to total DCC volume. Note that the volume of both DCC procedures and TEE-guided DCC procedures dramatically increased over 7 years, but the proportion of TEE-guided DCC to total DCC did not change markedly.
cal centers showed great variation in response to question 10. There was a comparable frequency of response for “b) low-risk patients” and “c) high–risk patients,” indicating the lack of a clear indication for TEE-guided cadioversion, or at least lack of consensus on the most appropriate patient population for
244 THE AMERICAN JOURNAL OF CARDIOLOGY姞
VOL. 85
electrical cardioversion for atrial fibrillation on left atrial appendage function and spontaneous echo contrast: characterization by simultaneous transesophageal echocardiography. J Am Coll Cardiol 1993;22:1359 –1366. 3. Laupacis A, Albers G, Dalen J, Dunn MI, Jacobson A, Singer, DE. Antithrombotic therapy in atrial fibrillation. Chest 1998;114:579S-589S. 4. Manning WJ, Silverman DI, Gordon SP, Krumholz HM, Douglas PS. Cardioversion from atrial fibrillation without prolonged anticoagulation with use of transesophageal echocardiography to exclude the presence of atrial thrombi. New Engl J Med 1993;328:750 –755. 5. Manning WJ, Silverman DI, Keighley CS, Oettgen P, Douglas PS. Transesophageal echocardiographically facilitated early cardioversion from atrial fibrillation using short-term anticoagulation: final results of a prospective 4.5-year study. J Am Coll Cardiol 1995;25:1354 –1361. 6. Klein AL, Grimm RA, Black IW, Leung DY, Chung MK, Vaughn SE, Murray RD, Miller D, Arheart KL. Cardioversion guided by transesophageal echocardiography: the ACUTE Pilot Study. A randomized, controlled trial. Ann Intern Med 1997;126:200 –209. 7. ACUTE Steering and Publications Committees. Design of a clinical trial for the Assessment of Cardioversion Using Transesophageal Echocardiography (the ACUTE Multicenter Study). Am J Cardiol 1998;81:877– 883. 8. Stoddard MF, Dawkins PR, Prince CR, Longaker RA. Transesophageal echocardiographic guidance of cardioversion in patients with atrial fibrillation. Am Heart J 1995;129:1204 –1215.
JANUARY 15, 2000