Left atrioventricular plane displacement predicts cardiac mortality in patients with chronic atrial fibrillation

International Journal of Cardiology 91 (2003) 1–7 www.elsevier.com / locate / ijcard

Left atrioventricular plane displacement predicts cardiac mortality in patients with chronic atrial fibrillation Erik Rydberg*, Martin Arlbrandt, Petri Gudmundsson, Leif Erhardt, Ronnie Willenheimer Department of Cardiology, Centre of Heart and Lung Diseases, Malmo¨ University Hospital, University of Lund, 205 02 Malmo¨ , Sweden Received 25 June 2002; received in revised form 18 September 2002; accepted 3 November 2002

Abstract Aim: The aim of the present study was to investigate if left atrioventricular plane displacement (AVPD) has a prognostic value in patients with atrial fibrillation. Methods and results: Left AVPD was assessed by two-dimensionally guided M-mode echocardiography in the four- and two-chamber views in 160 consecutive patients with chronic atrial fibrillation, who were followed up with regard to mortality for an average of 45 months. All-cause mortality during follow-up was 49% (n578). AVPD was lower in patients who died compared to those who survived: 6.661.7 versus 7.561.7 mm, P50.0005. In 49 patients (31%), death was due to chronic heart failure or acute myocardial infarction. Among those who died of cardiac events, AVPD was 6.361.6 mm, versus 7.161.8 mm among those who died of other causes, P50.0001. In multiple logistic regression analysis, AVPD (P50.005), age (P50.0005), and a history of chronic heart failure (P50.004) correlated independently with mortality. Conclusion: Left AVPD was clearly decreased in patients with atrial fibrillation. The decrease was most pronounced in patients who died of cardiac events, whereas it did not differ significantly between those who died of non-cardiac causes and those who survived. The discriminative value of left AVPD was limited.  2003 Elsevier Ireland Ltd. All rights reserved. Keywords: AVPD; Mortality; Prognosis; Atrial fibrillation

1. Introduction Echocardiographic assessment of left ventricular function by evaluation of left atrioventricular plane displacement (AVPD) is a simple and reliable method in patients with chronic heart failure, acute myocardial infarction, atrial fibrillation, and stable coronary artery disease [1–10]. Left AVPD is highly reproducible and rapidly performed [1–10]. It demands little of image quality since the atrioventricular plane is highly echogenic and can therefore be performed in virtually all patients [1–10]. Furthermore, left AVPD *Corresponding author. Tel.: 146-40-331-000; fax: 146-40-336-209. E-mail address: [email protected] (E. Rydberg).

can be used to reliably evaluate left ventricular systolic function despite asymmetric contraction [1– 10]. The AVPD method therefore compares favourably with most other echocardiographic methods for the evaluation of left ventricular systolic function. Left AVPD has been shown to correlate well with left ventricular ejection fraction (LVEF) assessed by two-dimensional echocardiography using the area– length method (r50.96), left ventricular wall motion index (r50.91), radionuclide ventriculography (r5 0.82–0.87), and contrast cineangiography (r50.89) [6,7,9,10]. However, left AVPD and LVEF are quite different expressions of LV systolic function. AVPD is considered to be the result of the contraction of mainly the subendocardial, longitudinal fibres and

0167-5273 / 03 / $ – see front matter  2003 Elsevier Ireland Ltd. All rights reserved. doi:10.1016 / S0167-5273(02)00578-8

2

E. Rydberg et al. / International Journal of Cardiology 91 (2003) 1–7

LVEF is predominantly a function of the contraction of subepicardial, circumferential fibres [11,12]. Left AVPD seems to be more sensitive to small changes in LV function than LVEF [2,3,13]. In patients with sinus rhythm, AVPD and LVEF correlate rather well, independent of cardiac disease. Compared to patients with sinus rhythm, patients with atrial fibrillation have a strongly reduced systolic long axis shortening, whereas the short axis shortening shows a rather small decrease [4,14,15]. After electrical cardioversion to sinus rhythm the decreased long axis shortening is improved or even normalised [4,14,15]. In a study by Emilsson, AVPD increased by 57% after electrical cardioversion, whereas LVEF increased by 11%, P,0.001 [14]. Alam et al. showed decreased AVPD in patients with atrial fibrillation and, after electrical cardioversion, mean AVPD increased substantially [4]. Further improvement was shown after 2 weeks and 1 month. AVPD has previously been shown to be strongly related to prognosis in patients with chronic heart failure, coronary artery disease, and after a myocardial infarction [1,8,16,17]. Independent of cardiac disease, AVPD was significantly lower in patients who died, especially in those who died of cardiac events, compared to the survivors and those who died of non-cardiac causes. AVPD has therefore been shown to be a strongly independent and clinically useful prognostic tool in these patients. The aim of the present study was to investigate if left AVPD has a prognostic value in patients with atrial fibrillation.

2. Patients and methods

2.1. Patients We consecutively included 160 patients (age 7569 years, 97 men) with chronic atrial fibrillation of different duration. Concomitant diseases and medication are shown in Table 1. The patients were examined by echocardiography by one sonographer during 1997 at the Department of Cardiology at Malmo¨ University Hospital, at the time of a hospitalisation due to elective electrocardioversion.

Table 1 Concomitant diseases and medication

Congestive heart failure Prior myocardial infarction Prior non-Q-wave myocardial infarction Angina pectoris Hypertension Diabetes Chronic obstructive lung disease Thyroid disease b-Blocker Ca-blocker Long-acting nitrates Diuretics ACE inhibitor AT1RB Digitalis

Dead (n578)

Survivors (n582)

P

71 33

53 25

,0.0001 NS

8 22 26 17

12 31 29 13

NS NS NS NS

14 6 23 14 13 75 36 6 46

16 7 46 17 17 65 37 2 54

NS NS ,0.001 NS NS 0.001 NS NS NS

ACE, angiotensin-converting enzyme; AT1RB, angiotensin type 1receptor blocker.

2.2. Echocardiographic examination Two-dimensional echocardiography and Doppler examinations were performed using a Hewlett-Packard (Andover, MA, USA) Sonos 2000, or 2500 echocardiography system and a 2.5 MHz transducer. Pulsed, continuous and colour-flow Doppler examinations were performed with the same transducer. Parasternal and apical views were obtained with the patient in a left lateral recumbent position. Measurements were acquired during silent respiration or endexpiratory apnea. Left AVPD was determined by two-dimensionally guided M-mode in the four- and two-chamber views, as described previously [1,5,8]. The regional AVPD (mm) was the distance covered by the atrioventricular plane between the position most remote from the apex (corresponding to the onset of contraction) and the location closest to the apex (corresponding to the end of contraction, including any post-ejection shortening), i.e. the full extent of the displacement. AVPD was measured in the septal, lateral, posterior and anterior regions of the left ventricle, and was calculated from an average of four heart cycles at each site during a period of as regular rhythm as possible. The mean of the average displacement in the four regions was calculated and expressed as left AVPD. In our laboratory the mean inter-individual vari-

E. Rydberg et al. / International Journal of Cardiology 91 (2003) 1–7

ability between two investigators (examining each patient immediately after one another) was 4.8% (displacement difference range 0–1.2 mm) in a series of 53 consecutive patients with a mean left AVPD of 7.8 mm (range 3.3–15.5 mm) [1]. The intra-observer variability of the determination of left AVPD was a mean of 2.0% (range 0–6%), corresponding to 0.23 mm (range 0–0.45 mm), in 39 randomly examined patients with a mean left AVPD of 11.2 mm (range 5.6–17.5 mm) [1,18]. The average AVPD in 15 controls with sinus rhythm (mean age 65 years, range 54–77 years, eight (53%) women) was 13.561.1 mm, and a mean displacement of ,10.0 mm is considered abnormal [18]. Cardiac dimensions were measured in the parasternal long axis view in two-dimensional mode [19]. Semi-quantitative grading of valvular regurgitation was based on colour Doppler signal area, continuous Doppler signal density, and continuous Doppler pressure half time (aortic regurgitation only) in the parasternal and apical views. The different degrees were: none (0), trace (0.5), mild (1), mild–moderate (1.5), moderate (2), moderate–severe (2.5), and severe (3).

2.3. Follow-up Patients were followed up with respect to mortality for an average of 4564 months after the echocardiographic examination. All deaths in Malmo¨ are registered in a central database. Since all patients included were residents of Malmo¨ the mortality was reliably assessed from this database. The causes of death were established from information in the medical records and from the death certificates and were divided into cardiac and non-cardiac. Cardiac causes were heart failure and myocardial infarction. The analysis of mortality data was performed by one investigator who was blinded to the results of the determinations of AVPD.

2.4. Statistics To test differences between two groups regarding continuous variables the unpaired t-test was applied. Differences between groups with respect to nominal variables were tested by a chi squared test. Correlation’s between a nominal dependent variable and

3

nominal or continuous independent variables were examined by uni- and multivariate logistic regression analysis. Before being entered into a multivariate model, all variables were tested for internal correlations in univariate linear regression analysis. Any correlation between two variables with an R value of at least 0.3 resulted in removal of the variable with the weakest correlation with the dependent variable in univariate logistic regression analysis. The hazard ratio was estimated to calculate the relative mortality risk. Data are expressed as mean6S.D., and P,0.05 was considered significant.

2.5. Ethics This study was approved by the Ethics Committee of Lund University.

3. Results All-cause mortality during follow-up was 49% (n5 78). Mortality due to chronic heart failure or acute myocardial infarction was 31% (n549) and due to non-cardiac causes 18% (n529). The non-cardiac causes were pneumonia (n58), various malignancies (n57), cerebrovascular insult (n54), sepsis (n52), renal insufficiency (n51), uraemia (n51), intoxication (n51), colitis ischemia (n51), cachexia (n51), senile astenia (n51), drowning (n51), and one unspecified cause (n51). Concomitant diseases and medication in patients who died and in survivors are shown in Table 1. Left AVPD was significantly lower in patients who died compared to the survivors: 6.661.7 vs. 7.561.7 mm, P50.0005. The relation between left AVPD and all-cause mortality by logistic regression analysis is shown in Fig. 1. In patients who died of cardiac causes, left AVPD was significantly lower compared to those who survived: 6.361.6 vs. 7.561.7 mm, P50.0001. The relation between left AVPD and cardiac mortality by logistic regression analysis is shown in Fig. 2. AVPD was also significantly lower in the patients who died of a cardiac event compared to those who died of non-cardiac causes: 6.361.6 vs. 7.161.8 mm, P50.035. No significant difference was found in AVPD between those who died of non-

4

E. Rydberg et al. / International Journal of Cardiology 91 (2003) 1–7

Fig. 1. All-cause mortality risk during 45 months plotted versus atrioventricular plane displacement by logistic regression analysis. Dotted lines show the 95% confidence interval.

cardiac causes and the survivors: 7.161.8 vs. 7.561.7 mm. Patients with a left AVPD below the median, ,7.0 mm (n579), had a 44% (n535) cardiac mortality and 12 patients died of a non-cardiac cause. Those with AVPD at or above 7.0 mm (n581) had a 17% (n514) cardiac mortality and 17 died of non-cardiac causes. Thus, the cardiac mortality hazard ratio for patients with left AVPD less than 7 mm versus those with 7 mm or more was 2.56, P50.0003. The

corresponding hazard ratio for all-cause mortality was 1.55, P50.011. Variables correlating with mortality showing P, 0.05 in univariate logistic regression analysis (Table 2) were tested for inclusion in multivariate logistic regression analysis. Variables tested that did not show P,0.05 in univariate analysis were left ventricular posterior wall diameter (P50.09), interventricular septum diameter (P50.21), mitral regurgitation (P5 0.12), prior myocardial infarction (P50.12), history of angina pectoris (P50.20), diabetes (P50.34), hypertension (P50.79), thyroid disease (P50.85), chronic obstructive lung disease (P50.80), use of calcium blockade (P50.66), long-acting nitrates (P5 0.51), angiotensin-converting-enzyme inhibitor (P5 0.90), angiotensin type 1-receptor blockade (P5 0.15), and digitalis (P50.37). AVPD and LVEF are both measures of left ventricular systolic function, therefore two multiple regression analyses were performed, including either. Besides left AVPD, the variables qualifying for the first multiple regression analysis were age, and history of chronic heart failure (Table 3). In the second multiple regression analysis, the variables, besides LVEF, were the same as in the first analysis Table 2 Variables correlating significantly with mortality in univariate logistic regression analysis

Age Body surface LA LVEDD RVIDD Degree of aortic regurgitation LVEF Left AVPD Congestive heart failure b-Blocker Diuretics

Fig. 2. Cardiac mortality risk during 45 months plotted versus atrioventricular plane displacement by logistic regression analysis. Dotted lines show the 95% confidence interval.

P

OR

95% lower

95% upper

,0.0001 0.016 0.007 0.002 0.03

1.087 0.164 1.130 1.131 1.147

1.043 0.038 1.033 1.047 1.013

1.134 0.711 1.235 1.222 1.297

,0.001 ,0.001 ,0.001

2.625 1.041 1.389

1.527 1.017 1.144

4.525 1.066 1.685

,0.001 ,0.001 0.004

5.550 0.327 6.538

2.259 0.170 1.833

13.635 0.629 23.321

LA, left atrial diameter at end diastole; LVEDD, left ventricular internal diameter at end diastole; RVIDD, right ventricular internal diameter at end diastole; LVEF, left ventricular ejection fraction; AVPD, atrioventricular plane displacement. The 95% confidence interval lower and upper limits are given. The ORs refer to the increased mortality risk per unit of increased age (1 year), increased body surface (1 m 2 ), increased LA (1 mm / m 2 ), increased LVEDD (1 mm / m 2 ), increased RVIDD (1 mm / m 2 ), increased degree of aortic regurgitation (1 unit), decreased LVEF (1%), and decreased AVPD (1 mm).

E. Rydberg et al. / International Journal of Cardiology 91 (2003) 1–7 Table 3 Multivariate logistic regression analysis with regard to all-cause mortality

Age Congestive heart failure AVPD

P

OR

95% lower

95% upper

,0.001

1.080

1.034

1.127

0.004 0.005

3.989 1.368

1.539 1.101

10.337 1.701

AVPD, atrioventricular plane displacement; OR, odds ratio. The 95% confidence interval lower and upper limits are given. The ORs refer to the increased risk of mortality per unit of increased age (1 year) and decreased AVPD (1 mm).

(Table 4). In the first analysis, AVPD, age and a history of chronic heart failure correlated independently with mortality, whereas LVEF, age and a history of chronic heart failure correlated independently with mortality in the second analysis. Among patients with prior myocardial infarction (n558), left AVPD was 6.061.5 mm in those who died (n533) and 7.762.0 mm in the survivors (n5 25), P,0.001. AVPD was also significantly lower among patients who died of a cardiac event, 5.861.4 mm (n525), compared to the survivors, P,0.001. There was no significant difference in AVPD between patients who died of non-cardiac causes, 6.461.7 mm (n58), and survivors (P50.12), or between those who died of cardiac events and patients who died of non-cardiac causes (P50.33).

4. Discussion The Framingham study showed that patients with atrial fibrillation had an 80% higher mortality rate compared to those with sinus rhythm [20]. During an average follow-up of 10 years after the onset of atrial fibrillation, 60% of the men and 45% of the women had died. Of those who died, 70% of the men and Table 4 Multiple logistic regression analysis between survivors with regard to mortality

Age Congestive heart failure LVEF

P

OR

95% lower

95% upper

,0.0001

1.094

1.046

1.144

0.014 0.0009

3.356 1.049

1.274 1.020

8.845 1.078

AVPD, atrioventricular plane displacement; OR, odds ratio. The 95% confidence interval lower and upper limits are given. The ORs refer to the increased risk of mortality per unit of increased age (1 year) and decreased LVEF (1%).

5

91% of the women died of cardiovascular causes, including stroke. A number of other studies have found a similar, approximately two-fold increase in all-cause mortality in patients with atrial fibrillation [21–25]. Despite the fact that atrial fibrillation is rather common in the population and has a poor prognosis, there is no useful prognostic method to identify those at higher risk for an event / death among these patients. We wanted to examine the prognostic value of left AVPD in patients with chronic atrial fibrillation since left AVPD has been shown to have a strong prognostic value in patients with chronic heart failure, acute myocardial infarction and stable coronary artery disease. It seems that the prognostic value of AVPD is independent of LVEF and, therefore, this method may be a useful prognostic tool independent of cause of heart disease. Another reason for using AVPD is that, since most patients with chronic atrial fibrillation are elderly, it is often difficult to obtain an image quality sufficient to reliably evaluate LVEF, whereas AVPD is useful even when the image quality is poor [1]. Furthermore, AVPD is highly reproducible, quickly performed, and reliable despite wall motion asymmetry [26,27]. Echocardiographically assessed left AVPD correlates well with LVEF in patients with sinus rhythm, independent of cardiac disease. However, a study by Emilsson et al. showed that patients with atrial fibrillation had a significantly different relationship between LVEF and AVPD compared to patients with sinus rhythm. Those patients had moderately decreased LVEF and relatively more decreased AVPD [15]. The relationship between LVEF and AVPD (ratio LVEF /AVPD) before cardioversion was 8.461.7 mm compared to 5.860.8 mm after. Thus, after electrical cardioversion to sinus rhythm, left AVPD increased much more than LVEF. Also, Alam et al. showed that, in patients with atrial fibrillation, AVPD increased after cardioversion to sinus rhythm, and further improvement was shown after 2 weeks and 1 month [4]. This is most likely explained by the fact that left ventricular filling is caused by passive left atrial emptying and atrial contraction [28]. In patients with atrial fibrillation the lack of electrical and mechanical activity during atrial systole causes a loss of active atrial contraction, resulting in insufficient atrial emp-

6

E. Rydberg et al. / International Journal of Cardiology 91 (2003) 1–7

tying and impaired left ventricular filling. Therefore, patients with atrial fibrillation usually show a 20– 30% decrease in left ventricular stroke volume and cardiac output [29], as well as a corresponding decrease of left AVPD. In the present study we showed that echocardiographically determined left AVPD was related to mortality also in patients with atrial fibrillation, especially to cardiac mortality. The discriminative value of AVPD with regard to mortality is also supported by the logistic regression analyses. Furthermore, patients with AVPD ,7 mm had an almost three-fold increased cardiac mortality compared to patients with AVPD .7 mm. Left AVPD was more closely related to cardiac mortality, whereas patients who died of non-cardiac causes showed an AVPD similar to that among the survivors. This makes sense since AVPD is not expected to be related to non-cardiac mortality. The findings are in accordance with earlier studies from our group in patients with chronic heart failure, acute myocardial infarction and coronary artery disease. However, the difference in left AVPD between patients who died during follow-up and survivors was rather small. Thus, the method may be less clinically useful in patients with chronic atrial fibrillation than in patients in general with chronic heart failure, acute myocardial infarction, and stable coronary artery disease, where assessment of left AVPD has a strong prognostic value. The reason why there is a smaller difference in AVPD between patients dying and survivors among patients with atrial fibrillation than in patients with sinus rhythm is most likely that all patients with atrial fibrillation have a substantially reduced left AVPD [1,16]. Thus, there is less room for difference between patients dying and survivors, reducing the discriminative value of left AVPD.

5. Conclusion Among patients with chronic atrial fibrillation, echocardiographically determined left AVPD was significantly lower among patients who died than in patients who survived during 45 months of mean follow-up. Left AVPD was clearly decreased in patients who died of cardiac events, whereas it did

not differ significantly between those who died of non-cardiac causes and those who survived. However, the discriminative value of left AVPD was limited. This makes the method somewhat less clinically useful in patients with atrial fibrillation than in patients in general with chronic heart failure, acute myocardial infarction and stable coronary artery disease.

References [1] Willenheimer R, Cline C, Erhardt L, Israelsson B. Left ventricular atrioventricular plane displacement: an echocardiographic technique for rapid assessment of prognosis in heart failure. Heart 1997;78:230–6. [2] Rydberg E, Willenheimer R, Brand B, Erhardt L. Left ventricular diastolic filling is related to the atrioventricular plane displacement in patients with coronary artery disease. Scand Cardiovasc J 2001;35:30–4. [3] Rydberg E, Willenheimer R, Erhardt L. Left atrioventricular plane displacement at rest is reduced in patients with coronary artery disease irrespective of prior myocardial infarction. Int J Cardiol 1999;69:201–7. [4] Alam M, Thorstrand C. Left ventricular function in patients with atrial fibrillation before and after cardioversion. Am J Cardiol 1992;69:694–6. [5] Willenheimer R, Israelsson B, Cline C, Rydberg E, Broms K, Erhardt L. Left atrioventricular plane displacement is related to both systolic and diastolic left ventricular performance in patients with chronic heart failure. Eur Heart J 1999;20:612–8. ¨ [6] Alam M, Hoglund C, Thorstrand C et al. Haemodynamic significance of the atrioventricular plane displacement in patients with coronary artery disease. Eur Heart J 1992;13:194–200. ¨ [7] Alam M, Hoglund C, Thorstrand C et al. Atrioventricular displacement in severe congestive heart failure following dilated cardiomyopathy or myocardial infarction. J Intern Med 1990;228:569– 75. [8] Willenheimer R, Erhardt L, Cline C, Rydberg E, Israelsson B. Prognostic significance of changes in left ventricular systolic function in elderly patients with congestive heart failure. Coron Artery Dis 1997;8:711–7. ¨ [9] Alam M, Hoglund C. Serial echocardiographic studies following thrombolytic treatment in myocardial infarction with special reference to the atrioventricular valve plane displacement. Clin Cardiol 1992;15:30–6. ¨ [10] Alam M, Hoglund C, Thorstrand C. Longitudinal systolic shortening of the left ventricle: an echocardiographic study in subjects with and without preserved global function. Clin Physiol 1992;12:443–52. ¨ S. Cardiac pumping and the function of the ventricular [11] Lundback septum. Acta Physiol Scand 1986;(Suppl 550):1–101. [12] Rushmer RF, editor, Cardiovascular dynamics, 4th ed., Philadelphia: W.B. Saunders, 1976, p. 93. [13] Willenheimer R, Rydberg E, Stagmo M, Gudmundsson P, Ericsson G, Erhardt L. Echocardiographic assessment of left atrioventricular plane displacement as a complement to left ventricular regional wall motion evaluation in the detection of myocardial dysfunction. Int J Card Imaging 2002;18:181–6.

E. Rydberg et al. / International Journal of Cardiology 91 (2003) 1–7 [14] Emilsson K, Wandt B. The relation between ejection fraction and mitral annulus motion before and after direct-current electrical cardioversion. Clin Physiol 2000;20(3):218–24. [15] Emilsson K, Wandt B. The relation between mitral annulus motion and left ventricular ejection fraction in atrial fibrillation. Clin Physiol 2000;20(1):44–9. [16] Brand B, Rydberg E, Ericsson G et al. Prognostication and risk stratification by assessment of left atrioventricular plane displacement in patients with myocardial infarction. Int J Cardiol 2002;83:35–41. [17] Rydberg E, Erhardt L, Brand B, Willenheimer R. Left atrioventricular plane displacement determined by echocardiography: a clinically useful, independent predictor of mortality in patients with stable coronary artery disease. Clin Cardiol (submitted for publication). [18] Willenheimer R. Assessment of left ventricular dysfunction and remodeling by determination of atrioventricular plane displacement and simplified echocardiography. Scand Cardiovasc J 1998;32(Suppl 48). [19] Henry WL, DeMaria A, Gramiak R, King DL, Kisslo JA, Popp RL, Sahn DJ, Schiller NB, Tajik A, Teichholz LE, Weyman AE. Report of the American Society of Echocardiography Nomenclature and Standards in Two-dimensional Echocardiography. Circulation 1980;62:212–7. [20] Kannel WB, Abbott RD, Savage DD, McNamara PM. Epidemiologic features of chronic atrial fibrillation. New Engl J Med 1982;306:1018. [21] Flegel KM, Shipley MJ, Rose G. Risk of stroke in non-rheumatic atrial fibrillation. Lancet i 1987;:526.

7

[22] Onundarson PT, Thorgeirsson G, Jonmundsson E, Sigfrusson N, Hardarson T. Chronic atrial fibrillation—epidemiological features and 14 year follow-up: a case control study. Eur Heart J 1987;8:521. [23] Kitchin AH, Milne JS. Longitudinal survey of ischemic heart disease in randomly selected sample of older population. Br Heart J 1977;39:889. [24] Lake FR, McCall MG, Cullen KJ, Rosman DL, de Klerk NH. Atrial fibrillation and mortality in an elderly population. Aust NZ J Med 1989;19:321. [25] Cameron A, Schwartz MJ, Kronmal RA, Kosinski AS. Prevalence and significance of atrial fibrillation in coronary artery disease (CASS Registry). Am J Cardiol 1988;61:714. [26] Fisher JP, Picard MH, Mikan JS, Fram DB, Fisher JR, Kluger J et al. Quantitation of myocardial dysfunction in ischemic heart disease by echocardiographic endocardial surface mapping: correlation with hemodynamic status. Am Heart J 1995;129:1114–21. [27] Køber L, Torp-Pedersen C, Carlsen J, Vidbaek R, Egeblad H on behalf of the TRACE study group. An echocardiographic method for selecting high risk patients shortly after acute myocardial infarction, for inclusion in multi-centre studies (as used in the TRACE study). Eur Heart J 1994;15:1616–20. ¨ [28] Alam M, Rosenhamer G, Hoglund C. Comparability of echocardiography and chest x-ray following myocardial infarction. J Intern Med 1989;226:171–5. [29] Sanfilippo, Abascal VM, Sheehan M, Oertel LB, Harrigan P, Hughes RA et al. Atrial enlargement as a consequence of atrial fibrillation. A prospective echocardiographic study. Circulation 1990;82:792–7.