- Email: [email protected]
The Role of the 12-Lead ECG as a Screening Test for Cardiac Source of Embolus
ORIGINAL ARTICLE
Original Article
The Role of the 12-Lead ECG as a Screening Test for Cardiac Source of Embolus Clara Kayei Chow, MBBS, FRACP a,c,∗ , Dominic Mou, MBBS, FRACP a , Anushka Patel, MBBS, FRACP, PhD a,c and David S. Celermajer, MBBS, FRACP, PhD a,b a
Department of Cardiology, Royal Prince Alfred Hospital, Missenden Road, Camperdown, Sydney, NSW 2050, Australia b Department of Medicine, University of Sydney, Australia c The George Institute for International Health, Sydney, Australia Available online 26 August 2005
Background: Investigation for cardiac source of embolus (CSE) is one of the commonest referrals for transthoracic echocardiography (TTE) of hospital inpatients, but has a relatively low-diagnostic yield. We sort to investigate whether 12-lead ECG might be useful in screening patients to obviate the need for TTE, in a subset of patients referred for echocardiographic investigation of cardiac source of embolus. Methods: We collected ECG and echo data for 400 consecutively referred inpatients for TTE investigation of possible cardiac source of embolus. We analysed this data for evidence of cardiac source of embolism on TTE in patients with a normal or abnormal ECG. Results: 41/400 (10%) subjects had possible CSE identified on TTE. Diagnostic yield for CSE was higher for those with abnormal compared with normal ECG (17% versus 6%, p < 0.001). Of 232/400 (58%) patients with a normal ECG, 200 had a normal TTE (86%). Of the 32 with normal ECG and abnormal TTE, echo found a possible embolic source in 13. Of those 168 (42%) with an abnormal ECG, TTE was normal in 73 and abnormal in 95, of whom 28 patients had an echo that identified a possible cardio-embolic source. ECG, therefore, had a sensitivity of 68%, specificity of 61%, positive predictive value of 0.17 and negative predictive value of 0.94 for detecting possible cardiac sources of embolus. Conclusions: Although TTE is a relatively low-yield investigation for the detection of cardiac source of embolus, 12lead ECG is not sufficiently sensitive to identify the ∼10% of patients in whom echo will demonstrate a diagnostic abnormality. (Heart Lung and Circulation 2005;14:242–244) © 2005 Australasian Society of Cardiac and Thoracic Surgeons and the Cardiac Society of Australia and New Zealand. Published by Elsevier Inc. All rights reserved. Keywords. Cardiac source of embolus; Echocardiography; ECG; Screening
Introduction
I
nvestigation for cardiac source of embolus (CSE) is one of the commonest reasons for referral of hospital inpatients for transthoracic echocardiography (TTE), a relatively expensive procedure with a relatively lowdiagnostic yield in this setting.1 More than 30 crosssectional studies have evaluated the yield of TTE or TEE (transesophageal echo), or both, in detecting intracardiac masses in patients with stroke.2 In consecutive patients, the yield of echocardiography for the detection of intracardiac masses ranged from 0 to 21%. Pooled data from these studies suggest an overall yield of 4% for TTE and 11% for TEE.2 The yield of CSE by TTE is even lower, possibly less than 2%, in patients without clinical evidence of cardiac disease by history, physical examination, electrocardiography or ∗ Corresponding author. Tel.: +61 2 9993 4566; fax: +61 2 9993 4502. E-mail address: [email protected] (C.K. Chow).
chest radiography, compared with patients who do have clinical evidence of cardiac disease (less than ∼15%).2,3 The Canadian Task Force on Preventive Health Care concluded from a recent review that there is fair evidence to recommend echocardiography in patients with stroke and clinical evidence of cardiac disease by history, physical examination, electrocardiography or chest radiography. However, there was insufficient evidence to recommend for or against TEE in patients with normal results of TTE. There was also insufficient evidence to recommend for or against routine echocardiography in patients (including young patients) without clinical cardiac disease. The utility of the 12-lead ECG alone in screening patients referred for echo investigation of CSE has not been examined. We hypothesized that a normal 12-lead ECG, showing sinus rhythm and no q-waves, would identify a subset of patients with an extremely low yield for identifiable CSE on echo, thus potentially obviating the need for this more expensive and labour-intensive diagnostic test.
© 2005 Australasian Society of Cardiac and Thoracic Surgeons and the Cardiac Society of Australia and New Zealand. Published by Elsevier Inc. All rights reserved.
1443-9506/04/$30.00 doi:10.1016/j.hlc.2005.06.010
Methods In two periods between 1999 and 2002, we reviewed 400 consecutive inpatients referred for TTE investigation of possible cardiac source of embolus, for whom both complete TTE and ECG data were available. In a 4-month period between April and July 2000, we were unable to collect ECG data for patients, as traces were not systematically stored over a short period when the computer systems ‘‘changed over’’. Hence, all patients investigated in this period were excluded. Referrals were made at the discretion of the patient’s physician. Patients were identified retrospectively from a computerised database. Demographic information, patient echo reports and printouts of patient’s ECG were all obtained from matching this list with other computerised database information. All patients had standard TTE, and images were acquired in the parasternal long and short axis, in apical two-, three-, and four-chamber views, and in subcostal and suprasternal views. TTEs, were classified as ‘‘positive’’ for potential cardiac sources of embolus if there was intracardiac thrombus, echo-density on a valve, atrial mass or patent foramen ovale identified. The 12-lead ECGs were classified as abnormal if they had q-waves, left bundle branch block or were not in sinus rhythm. All tests were read by experienced senior cardiology staff.
Statistics Descriptive data are expressed as mean values and standard deviations. Sensitivity, specificity and predictive values were calculated and difference in proportions between groups was assessed using chi-squared statistics.
Results The mean age (S.D.) of patients was 64 (15) years and 252/400 (63%) were male. Potential cardiac source of embolus was identified in 41/400 (10%) of the subjects. The diagnostic yield for CSE was higher for those with abnormal ECG compared with normal ECG (17% versus 6%, p < 0.001). Of the 232/400 patients with a normal ECG, 200 had a normal TTE (86%). Of the 32 with normal ECG and abnormal TTE, echo found a possible embolic source in 13 [left ventricular thrombus (6), patent foramen ovale (PFO) (6), echo-density on valve (1)]. The remaining 19 had regional wall abnormalities (16), and dilated cardiomyopathy (DCM) (3). Out of 400 patients, 168 had abnormal ECG. Of these 83 were in atrial fibrillation or atrial flutter. Of those 168 with an abnormal ECG, echo was normal in 73 and abnormal in 95, of whom 28 patients had TTE that found a possible embolic source [intracardiac thrombus (22), PFO (4), atrial mass (1), echo-density on valve (1)]. ECG, therefore, had a sensitivity of 68%, specificity of 61%, positive predictive value of 0.17 and negative predic-
Chow et al. ECG screening for cardiac source of embolism
243
tive value of 0.94 for detecting possible cardiac sources of embolus on transthoracic echo.
Discussion This study is the first, to our knowledge, to examine systematically the role of the 12-lead ECG as a screening test in patients requiring investigation for a potential cardiac source for embolus. This has been identified as a topic with an insufficient evidence base for clinical guidelines/recommendations.2 We have found that a normal ECG is insufficiently sensitive to obviate the need for comprehensive TTE examination in this patient group. In our study of 400 inpatients referred for investigation of cardiac source of emboli, 41/400 or approximately 10% of patients had a suspected cardiac source of emboli identified on TTE. This is consistent with other similar studies concerning the yield of TTE in patients referred for CSE investigation.1,5 While an abnormal 12-lead ECG was able to identify 28 of the 41 patients (68%) with suggested CSE on echo, 13 of 41 such patients had entirely normal ECGs. Hence, although TTE is a relatively low-yield investigation for the detection of cardiac source of embolus present in ∼10% of all patients referred, TTE was still abnormal in 13 or 232 (5%) patients in whom ECG was unremarkable. There are certain cardiac lesions, which although rare, can cause systemic embolism but without affecting the ECG. These include atrial myxoma, present in one in 400 in our series and slightly more commonly valvular vegetations and patent foramen ovale without associated abnormalities. Surprisingly, we also found six patients with LV thrombus and no ECG evidence of past q-wave infarction. ECG itself is recognised as specific, but not particularly sensitive, for echo-detected segmental wall motion abnormalities6 (Fig. 1). These ‘‘ECG-silent’’ potential cardio-embolic sources are clearly amenable to medical and/or surgical treatment, to minimise risk of recurrent embolic events. Thus, it is clear from this study of consecutive inpatients, that such ‘‘ECG-silent’’ but treatable lesions are sufficiently common to preclude the confident use of a normal ECG to
Figure 1. Left atrial myxoma. Transthoracic apical four-chamber view showing an atrial myxoma identified on TTE in a patient with a normal ECG.
ORIGINAL ARTICLE
Heart Lung and Circulation 2005;14:242–244
ORIGINAL ARTICLE
244
Chow et al. ECG screening for cardiac source of embolism
exclude CSE, even though it identifies a low-yield (only 6% positive TTE) group. As particular lesions such as PFO and left atrial masses may not be picked up as well by TTE,7 TOE may be a more sensitive diagnostic tool in patients referred for investigation of cardiac source of embolism. This test does, however, carry some risks4 and a greater cost. Hence, many centres will still not use it as a first line investigation for patients referred for assessment of possible cardiac source of embolism.
Study Limitations We recognise the sample population has some level of selection bias, as the majority of subjects were inpatients and referred by neurologists and cardiologists familiar with their medical conditions and some familiar with the current guidelines. However, this does represent a ‘‘real world’’ series of inpatients, referred for clinically realistic reasons. Our data collection was predominantly retrospective, nonetheless complete and consecutive.
Conclusions Selection of patients for echocardiographic investigation of cardiac source of embolism remains a challenging clinical decision. This is a more difficult decision for patients without clinically evident cardiac disease, as these investigations are likely to be of low yield.8 Although an abnormal ECG is associated with a higher yield of CSE found on echo (17% versus 5% in those with normal ECG), we have found that screening out patients with a normal ECG would miss approximately 30% (13/41) of patients with possible CSE. Thus, ECG alone does not appear to be sensitive enough
Heart Lung and Circulation 2005;14:242–244
to identify the approximately 10% of patients in our group, whose echocardiogram demonstrated a diagnostic abnormality for cardiac source of embolus. There are no competing interests or conflicts of interest from any of the authors.
References 1. Sansoy V, Abbott RD, Jayaweera AR, Kaul S. Low yield of transthoracic echocardiography for cardiac source of embolism. Am J Cardiol 1995;75:166–9. 2. Kapral MK, Silver FL, with the Canadian Task Force on Preventive Health Care. Preventive Health Care, 1999 update:2. Echocardiography for the detection of a cardiac source of embolus in patients with stroke. Can Med Assoc J 1999;161(8):989–96. 3. Stollberger C, Brainin M, Abzieher F, Slany J. Embolic stroke and transoesophageal echo; can clinical parameters predict the diagnostic yield? J Neurol 1995;242:437–42. 4. Daniel VW, Erbel R, Kasper W, Visser CA, Engberding R, Sutherland GR, et al. Safety of transoesophageal echocardiography. A multicenter survey of 10419 examinations. Circulation 1991;83:817–21. 5. Leung DY, Black IW, Cranney GB, Walsh WF, Grimm RA, Stewart WJ, et al. Selection of patients for transoesophageal echocardiography after stroke and systemic embolic events: role of transthoracic echocardiography. Stroke 1995;26(10):1820–4. 6. Kontos MC, Arrowood JA, Paulsen WH, Nixon JV. Early echocardiography can predict cardiac events in emergency department patients with chest pain. Ann Emerg Med 1998;31(5):550–7. 7. Cujec B, Polasek P, Voll C, Shuaib A. Transoesophageal echocardiography in the detection of potential cardiac source of embolism in stroke patients. Stroke 1991;22:727–33. 8. Beattie JR, Cohen DJ, Manning WJ, Douglas PS. Role of routine transthoracic echocardiography in evaluation and management of stroke. J Int Med 1998;243(4):281–91.
Original Article
The Role of the 12-Lead ECG as a Screening Test for Cardiac Source of Embolus Clara Kayei Chow, MBBS, FRACP a,c,∗ , Dominic Mou, MBBS, FRACP a , Anushka Patel, MBBS, FRACP, PhD a,c and David S. Celermajer, MBBS, FRACP, PhD a,b a
Department of Cardiology, Royal Prince Alfred Hospital, Missenden Road, Camperdown, Sydney, NSW 2050, Australia b Department of Medicine, University of Sydney, Australia c The George Institute for International Health, Sydney, Australia Available online 26 August 2005
Background: Investigation for cardiac source of embolus (CSE) is one of the commonest referrals for transthoracic echocardiography (TTE) of hospital inpatients, but has a relatively low-diagnostic yield. We sort to investigate whether 12-lead ECG might be useful in screening patients to obviate the need for TTE, in a subset of patients referred for echocardiographic investigation of cardiac source of embolus. Methods: We collected ECG and echo data for 400 consecutively referred inpatients for TTE investigation of possible cardiac source of embolus. We analysed this data for evidence of cardiac source of embolism on TTE in patients with a normal or abnormal ECG. Results: 41/400 (10%) subjects had possible CSE identified on TTE. Diagnostic yield for CSE was higher for those with abnormal compared with normal ECG (17% versus 6%, p < 0.001). Of 232/400 (58%) patients with a normal ECG, 200 had a normal TTE (86%). Of the 32 with normal ECG and abnormal TTE, echo found a possible embolic source in 13. Of those 168 (42%) with an abnormal ECG, TTE was normal in 73 and abnormal in 95, of whom 28 patients had an echo that identified a possible cardio-embolic source. ECG, therefore, had a sensitivity of 68%, specificity of 61%, positive predictive value of 0.17 and negative predictive value of 0.94 for detecting possible cardiac sources of embolus. Conclusions: Although TTE is a relatively low-yield investigation for the detection of cardiac source of embolus, 12lead ECG is not sufficiently sensitive to identify the ∼10% of patients in whom echo will demonstrate a diagnostic abnormality. (Heart Lung and Circulation 2005;14:242–244) © 2005 Australasian Society of Cardiac and Thoracic Surgeons and the Cardiac Society of Australia and New Zealand. Published by Elsevier Inc. All rights reserved. Keywords. Cardiac source of embolus; Echocardiography; ECG; Screening
Introduction
I
nvestigation for cardiac source of embolus (CSE) is one of the commonest reasons for referral of hospital inpatients for transthoracic echocardiography (TTE), a relatively expensive procedure with a relatively lowdiagnostic yield in this setting.1 More than 30 crosssectional studies have evaluated the yield of TTE or TEE (transesophageal echo), or both, in detecting intracardiac masses in patients with stroke.2 In consecutive patients, the yield of echocardiography for the detection of intracardiac masses ranged from 0 to 21%. Pooled data from these studies suggest an overall yield of 4% for TTE and 11% for TEE.2 The yield of CSE by TTE is even lower, possibly less than 2%, in patients without clinical evidence of cardiac disease by history, physical examination, electrocardiography or ∗ Corresponding author. Tel.: +61 2 9993 4566; fax: +61 2 9993 4502. E-mail address: [email protected] (C.K. Chow).
chest radiography, compared with patients who do have clinical evidence of cardiac disease (less than ∼15%).2,3 The Canadian Task Force on Preventive Health Care concluded from a recent review that there is fair evidence to recommend echocardiography in patients with stroke and clinical evidence of cardiac disease by history, physical examination, electrocardiography or chest radiography. However, there was insufficient evidence to recommend for or against TEE in patients with normal results of TTE. There was also insufficient evidence to recommend for or against routine echocardiography in patients (including young patients) without clinical cardiac disease. The utility of the 12-lead ECG alone in screening patients referred for echo investigation of CSE has not been examined. We hypothesized that a normal 12-lead ECG, showing sinus rhythm and no q-waves, would identify a subset of patients with an extremely low yield for identifiable CSE on echo, thus potentially obviating the need for this more expensive and labour-intensive diagnostic test.
© 2005 Australasian Society of Cardiac and Thoracic Surgeons and the Cardiac Society of Australia and New Zealand. Published by Elsevier Inc. All rights reserved.
1443-9506/04/$30.00 doi:10.1016/j.hlc.2005.06.010
Methods In two periods between 1999 and 2002, we reviewed 400 consecutive inpatients referred for TTE investigation of possible cardiac source of embolus, for whom both complete TTE and ECG data were available. In a 4-month period between April and July 2000, we were unable to collect ECG data for patients, as traces were not systematically stored over a short period when the computer systems ‘‘changed over’’. Hence, all patients investigated in this period were excluded. Referrals were made at the discretion of the patient’s physician. Patients were identified retrospectively from a computerised database. Demographic information, patient echo reports and printouts of patient’s ECG were all obtained from matching this list with other computerised database information. All patients had standard TTE, and images were acquired in the parasternal long and short axis, in apical two-, three-, and four-chamber views, and in subcostal and suprasternal views. TTEs, were classified as ‘‘positive’’ for potential cardiac sources of embolus if there was intracardiac thrombus, echo-density on a valve, atrial mass or patent foramen ovale identified. The 12-lead ECGs were classified as abnormal if they had q-waves, left bundle branch block or were not in sinus rhythm. All tests were read by experienced senior cardiology staff.
Statistics Descriptive data are expressed as mean values and standard deviations. Sensitivity, specificity and predictive values were calculated and difference in proportions between groups was assessed using chi-squared statistics.
Results The mean age (S.D.) of patients was 64 (15) years and 252/400 (63%) were male. Potential cardiac source of embolus was identified in 41/400 (10%) of the subjects. The diagnostic yield for CSE was higher for those with abnormal ECG compared with normal ECG (17% versus 6%, p < 0.001). Of the 232/400 patients with a normal ECG, 200 had a normal TTE (86%). Of the 32 with normal ECG and abnormal TTE, echo found a possible embolic source in 13 [left ventricular thrombus (6), patent foramen ovale (PFO) (6), echo-density on valve (1)]. The remaining 19 had regional wall abnormalities (16), and dilated cardiomyopathy (DCM) (3). Out of 400 patients, 168 had abnormal ECG. Of these 83 were in atrial fibrillation or atrial flutter. Of those 168 with an abnormal ECG, echo was normal in 73 and abnormal in 95, of whom 28 patients had TTE that found a possible embolic source [intracardiac thrombus (22), PFO (4), atrial mass (1), echo-density on valve (1)]. ECG, therefore, had a sensitivity of 68%, specificity of 61%, positive predictive value of 0.17 and negative predic-
Chow et al. ECG screening for cardiac source of embolism
243
tive value of 0.94 for detecting possible cardiac sources of embolus on transthoracic echo.
Discussion This study is the first, to our knowledge, to examine systematically the role of the 12-lead ECG as a screening test in patients requiring investigation for a potential cardiac source for embolus. This has been identified as a topic with an insufficient evidence base for clinical guidelines/recommendations.2 We have found that a normal ECG is insufficiently sensitive to obviate the need for comprehensive TTE examination in this patient group. In our study of 400 inpatients referred for investigation of cardiac source of emboli, 41/400 or approximately 10% of patients had a suspected cardiac source of emboli identified on TTE. This is consistent with other similar studies concerning the yield of TTE in patients referred for CSE investigation.1,5 While an abnormal 12-lead ECG was able to identify 28 of the 41 patients (68%) with suggested CSE on echo, 13 of 41 such patients had entirely normal ECGs. Hence, although TTE is a relatively low-yield investigation for the detection of cardiac source of embolus present in ∼10% of all patients referred, TTE was still abnormal in 13 or 232 (5%) patients in whom ECG was unremarkable. There are certain cardiac lesions, which although rare, can cause systemic embolism but without affecting the ECG. These include atrial myxoma, present in one in 400 in our series and slightly more commonly valvular vegetations and patent foramen ovale without associated abnormalities. Surprisingly, we also found six patients with LV thrombus and no ECG evidence of past q-wave infarction. ECG itself is recognised as specific, but not particularly sensitive, for echo-detected segmental wall motion abnormalities6 (Fig. 1). These ‘‘ECG-silent’’ potential cardio-embolic sources are clearly amenable to medical and/or surgical treatment, to minimise risk of recurrent embolic events. Thus, it is clear from this study of consecutive inpatients, that such ‘‘ECG-silent’’ but treatable lesions are sufficiently common to preclude the confident use of a normal ECG to
Figure 1. Left atrial myxoma. Transthoracic apical four-chamber view showing an atrial myxoma identified on TTE in a patient with a normal ECG.
ORIGINAL ARTICLE
Heart Lung and Circulation 2005;14:242–244
ORIGINAL ARTICLE
244
Chow et al. ECG screening for cardiac source of embolism
exclude CSE, even though it identifies a low-yield (only 6% positive TTE) group. As particular lesions such as PFO and left atrial masses may not be picked up as well by TTE,7 TOE may be a more sensitive diagnostic tool in patients referred for investigation of cardiac source of embolism. This test does, however, carry some risks4 and a greater cost. Hence, many centres will still not use it as a first line investigation for patients referred for assessment of possible cardiac source of embolism.
Study Limitations We recognise the sample population has some level of selection bias, as the majority of subjects were inpatients and referred by neurologists and cardiologists familiar with their medical conditions and some familiar with the current guidelines. However, this does represent a ‘‘real world’’ series of inpatients, referred for clinically realistic reasons. Our data collection was predominantly retrospective, nonetheless complete and consecutive.
Conclusions Selection of patients for echocardiographic investigation of cardiac source of embolism remains a challenging clinical decision. This is a more difficult decision for patients without clinically evident cardiac disease, as these investigations are likely to be of low yield.8 Although an abnormal ECG is associated with a higher yield of CSE found on echo (17% versus 5% in those with normal ECG), we have found that screening out patients with a normal ECG would miss approximately 30% (13/41) of patients with possible CSE. Thus, ECG alone does not appear to be sensitive enough
Heart Lung and Circulation 2005;14:242–244
to identify the approximately 10% of patients in our group, whose echocardiogram demonstrated a diagnostic abnormality for cardiac source of embolus. There are no competing interests or conflicts of interest from any of the authors.
References 1. Sansoy V, Abbott RD, Jayaweera AR, Kaul S. Low yield of transthoracic echocardiography for cardiac source of embolism. Am J Cardiol 1995;75:166–9. 2. Kapral MK, Silver FL, with the Canadian Task Force on Preventive Health Care. Preventive Health Care, 1999 update:2. Echocardiography for the detection of a cardiac source of embolus in patients with stroke. Can Med Assoc J 1999;161(8):989–96. 3. Stollberger C, Brainin M, Abzieher F, Slany J. Embolic stroke and transoesophageal echo; can clinical parameters predict the diagnostic yield? J Neurol 1995;242:437–42. 4. Daniel VW, Erbel R, Kasper W, Visser CA, Engberding R, Sutherland GR, et al. Safety of transoesophageal echocardiography. A multicenter survey of 10419 examinations. Circulation 1991;83:817–21. 5. Leung DY, Black IW, Cranney GB, Walsh WF, Grimm RA, Stewart WJ, et al. Selection of patients for transoesophageal echocardiography after stroke and systemic embolic events: role of transthoracic echocardiography. Stroke 1995;26(10):1820–4. 6. Kontos MC, Arrowood JA, Paulsen WH, Nixon JV. Early echocardiography can predict cardiac events in emergency department patients with chest pain. Ann Emerg Med 1998;31(5):550–7. 7. Cujec B, Polasek P, Voll C, Shuaib A. Transoesophageal echocardiography in the detection of potential cardiac source of embolism in stroke patients. Stroke 1991;22:727–33. 8. Beattie JR, Cohen DJ, Manning WJ, Douglas PS. Role of routine transthoracic echocardiography in evaluation and management of stroke. J Int Med 1998;243(4):281–91.