Percent achieved of predicted peak exercise oxygen uptake and kinetics of recovery of oxygen uptake after exercise for risk stratification in chronic heart failure

International Journal of Cardiology 64 (1998) 117–124

Percent achieved of predicted peak exercise oxygen uptake and kinetics of recovery of oxygen uptake after exercise for risk stratification in chronic heart failure a, a a a a Domenico Scrutinio *, Andrea Passantino , Rocco Lagioia , Francesco Napoli , Antonio Ricci , b Paolo Rizzon a

Division of Cardiology, ‘‘ S. Maugeri’’ Foundation, IRCCS, Rehabilitation Institute of Cassano M., Bari, Italy b Institute of Cardiology, University of Bari, Bari, Italy Received 20 November 1997; received in revised form 7 January 1998; accepted 7 January 1998

Abstract To investigate whether percent achieved of predicted peak exercise oxygen uptake (%VO 2 max ) and recovery of oxygen consumption after exercise may provide prognostic information in chronic heart failure (CHF), we prospectively studied 196 patients with mild to moderate CHF. The following variables were examined: age, etiology of CHF, functional class, ejection fraction (EF), peak exercise oxygen uptake normalized for body weight (VO 2 max ), %VO 2 max , time to reach 50% of the peak oxygen uptake after exercise (T 1 / 2 VO 2 max ), presence of nonsustained ventricular tachycardia (NSVT) and inability to take ACE-inhibitors. VO 2 max was the most powerful predictor of cardiac death (P,0.0001). Other independent predictors of death were EF, T 1 / 2 VO 2 max , NSVT and inability to take ACE-inhibitors. The discriminatory accuracy of VO 2 max for cardiac death was not significantly greater than that of %VO 2 max . In conclusion, the determination of %VO 2 max does not enhance risk stratification in CHF whereas the kinetics of oxygen consumption after exercise can provide prognostic information.  1998 Elsevier Science B.V. Keywords: Chronic heart failure; Prognosis; Cardiopulmonary exercise test

1. Introduction The incidence of chronic heart failure (CHF) has more than doubled since 1980 and, despite advances in medical treatment, prognosis still remains generally poor [1–4]. Risk assessment is essential to the management of CHF. Among a variety of prognostic variables, peak exercise oxygen uptake normalized for body weight (VO 2 max ) has been shown to be one of the most powerful predictors of prognosis in mild to advanced CHF [5–11] and a useful guide for *Corresponding author. Tel.: 139 80 7814232; fax: 139 80 7814310; e-mail: [email protected] 0167-5273 / 98 / $19.00  1998 Elsevier Science B.V. All rights reserved. PII S0167-5273( 98 )00019-9

optimal timing of cardiac transplantation [11]. Peak exercise oxygen uptake is however influenced by several factors including age, gender, body weight, motivation and familiarization of the patient with the exercise test [12,13]. Thus, it was suggested that the use of the percent achieved of predicted age-, gender-, and body weight-adjusted peak oxygen uptake (%VO 2 max ) might enhance the prognostic value of cardiopulmonary exercise testing. The recovery of oxygen consumption after exercise, as determined by the time to reach 50% of the peak exercise value (T 1 / 2 VO 2 max ), has been shown to be delayed in patients with CHF in parallel with the severity of the disease [14]. Importantly, it was found

118

D. Scrutinio et al. / International Journal of Cardiology 64 (1998) 117 – 124

to be reproducible and largely unaffected by exercise level [14]. These data suggest the potential of T 1 / 2 VO 2 max ln providing prognostic information. To date, the role of %VO 2 max and T 1 / 2 VO 2 max in assessing prognosis of patients with CHF has been investigated in a few retrospective or small-sized studies [15–17]. In this prospective study we sought: (1) to investigate whether %VO 2 max provides additional prognostic information to the absolute value of VO 2 max ; (2) to assess the prognostic role of T 1 / 2 VO 2 max . 2. Materials and methods

2.1. Study group The population of this prospective study consisted of 196 patients with New York Heart Association (NYHA) class II / III symptoms of CHF and an ejection fraction (EF) #40% as assessed by twodimensional echocardiography; the criterium of .80% endocardial definition must be met for inclusion. Ejection fraction was calculated using the Simpson’s rule from the standard apical view. Patients were excluded from the study if they had any of the following: NYHA IV class symptoms; failure to reach anaerobic threshold; myocardial infarction, coronary by-pass surgery or angioplasty within the previous 6 months; angina pectoris; exercise-induced ischemia or significant ventricular arrhythmias; surgically uncorrected valvular disease or valvular surgery within the previous 6 months; hypertrophic or restrictive cardiomyopathy; uncontrolled hypertension; active myocarditis; significant chronic pulmonary disease; musculoskeletal or symptomatic peripheral vascular disease limiting exercise testing; chest wall deformity. The diagnosis of ischemic heart disease (IHD) was based on a medical history of documented myocardial infarction or on positive coronary angiography (.50% narrowing of one or more major branches), if performed.

continue exercising until dyspnoea or fatigue forced them to stop. Those patients who were not familiar with exercise testing underwent a preliminary exercise test. Oxygen uptake and carbon dioxide output were measured breath-by-breath by an automated system (Sensor Medics system 2900, Anaheim, California). Measurements were taken at rest and every 20 s throughout exercise and recovery. Raw ventilatory and gas exchange parameters were stored on floppy disks. Anaerobic threshold was defined as the point at which the respiratory exchange ratio (VCO 2 / O 2 ) was 1.00 or, subordinately, as the point at which the ventilatory equivalent for O 2 (VE / VO 2 ) was minimal, followed by a progressive increase or by the V-slope method [18]. Peak oxygen uptake was defined as the VO 2 measured at the end of exercise. Percent achieved of predicted peak oxygen uptake was calculated by using a gender-specific equation normalizing peak oxygen uptake for age, weight and height [19]. T 1 / 2 VO 2 max was defined as the elapsed time from the end-of exercise to the point at which VO 2 had fallen to 50% of the end-of exercise value [14]. When this occurred in the middle of two sampling points, T 1 / 2 was set at the second of these points [14].

2.3. Follow-up and outcome Vital status was determined by interviewing patients, their relatives and / or treating physician. No patient was lost to follow-up. Details on the clinical status before death and the circumstances surrounding it were obtained by interviewing medical personnel or family members. The cause of death was prospectively classified as sudden (either witnessed to be instantaneous, unwitnessed but assumed to be instantaneous on the basis of clinical setting, or occurring within 1 h of an abrupt change in previous clinical state); due to pump failure, usually with progressive worsening of heart failure symptoms even if the terminal event was an arrhythmia; caused by another type of cardiovascular event; and noncardiovascular [4].

2.2. Cardiopulmonary exercise testing 2.4. Statistical analysis Informed written consent for exercise testing was obtained from all patients. Treadmill exercise testing with expiratory gas analysis was performed using the modified Bruce protocol. Patients were encouraged to

The primary end-point was cardiac death. In survival analysis, heart transplantation was considered a censored observation (i.e., withdrawn from the

D. Scrutinio et al. / International Journal of Cardiology 64 (1998) 117 – 124

study at the time of transplant). The examined variables were age, etiology of HF, NYHA class, EF, VO 2 max , %VO 2 max , T 1 / 2 VO 2 max , presence of nonsustained ventricular tachycardia (NSVT) on 24 h Holter monitoring defined as six or more consecutive premature complexes at a rate of 100 beats / min or more and inability to take ACE-inhibitors because of contra-indications or poor tolerance. To allow clinically meaningful comparisons between groups, we dichotomized continuous variables when estimating their association with mortality. For each, we sought the dichotomization point that maximized the hazard ratio from a Cox regression model [20] for comparing patients below the cutpoint (expected to be at high risk) with those at or above it (expected to be at low risk); for what concerns T 1 / 2 VO 2 max patients above the cut-point were considered at high risk [21]. The search for dichotomization was restricted in the interval between 10% low and 90% high. The hazard ratio for each dichotomization point within this interval (unadjusted for any covariate) was calculated and the point at which its maximum was reached was identified [21]. Survival curves were constructed for all variables using the Kaplan-Meier method. Univariate analysis was performed by comparing survival curves by log-rank test [22]. Multivariate analysis was performed using the Cox proportional hazards regression model [20] in order to assess the independent contribution of each variable to risk of death. Variables with a P value ,0.1 in univariate analysis were entered into the Cox model. In order to compare the discriminatory accuracies of VO 2 max and %VO 2 max for cardiac death, the Receiver operating characteristics (ROC) curves were constructed by plotting true positive against false positive rates [15,23]. The ROC curves were compared using the Z statistic [23]. Statistical analyses were performed on SPSS for Windows Advanced Statistics 6.1, statistical software. A probability value of less than 0.05 was considered significant. Data are expressed as mean6standard deviation (SD).

3. Results

3.1. Survival analysis Patient baseline characteristics are presented in

119

Table 1 Baseline characteristics of the study population n (%) Gender (M / F) Age (years) Etiology Ischemic DC Idiopatic DC Hypertension Surgically corrected valvular disease NYHA Class II III EF (%) VO 2 max (ml / kg / min) %VO 2 max T 1 / 2 VO 2 max (s) NSVT Inability to take ACEinhibitors Mean follow up (days)

Mean6sd

176 / 20 5869 72 (37) 114 (58) 8 (4) 2 (1)

139 (70) 57 (30) 2766 18.765 61622 152654 29 (15) 30 (15) 5516130

NYHA, New York Heart Association classification; DC, dilated cardiomyopathy; EF, ejection fraction; VO 2 max , oxygen consumption; NSVT, non sustained ventricular tachycardia; T 1 / 2 VO 2 max , half recovery time of peak oxygen consumption. RR (C.I.)5relative risk (95% confidence interval).

Table 1. During a mean follow-up of 5516130 days, 36 patients died, 13 were sudden deaths and 23 due to pump failure. Four patients underwent heart transplant. The actuarial cumulative survival rates at 1 and 2 years were 88% and 78%, respectively. Comparison of the characteristics of surviving and nonsurviving patients is displayed in Table 2. At univariate analysis, all the examined variables but etiology were associated with mortality (Fig. 1). At multivariate analysis, VO 2 max was the most powerful predictor of death (P,0.0001) (Table 3). Patients with VO 2 max , 15 ml / kg / min were 2.17 times as likely to die as those with VO 2 max ,15 ml / kg / min. Four patients in the former and none in the latter subgroup underwent heart transplant. Other independent predictors of death were EF (P,0.01), T 1 / 2 VO 2 max (P,0.05), NSVT (P,0.01) and inability to take ACE-inhibitors (P,0.001) (Table 3). Since nine variables were examined, a P value ,0.0055 would be required to define statistical significance in a rigid multivariate analysis. This level of significance was reached only by VO 2 max (P,0.0001) and inability to take ACEinhibitors (P,0.001). For a further comparison between VO 2 max and %VO 2 max , the discriminatory accuracies of VO 2 max and %VO 2 max for cardiac death were compared by

120

D. Scrutinio et al. / International Journal of Cardiology 64 (1998) 117 – 124

Table 2 Comparison of the characteristics of the surviving and nonsurviving patients

No. patients Age Gender (M / F) NYHA Class II III Etiology Ischemic DC Idiopatic DC EF (%) VO 2 max (ml / kg / min) %VO 2 max T 1 / 2 VO 2 max NSVT Inability to take ACEinhibitors

Survivors

Non Survivors

160 56.9610 142 / 18

36 61.769 34 / 2

119 41

P ,0.01 ns P,0.05

20 16 ns

55 105 2866 19.665 63.3615 146647 18 16

17 19 2466 15.165 50.6617 179675 11 14

,0.01 ,0.01 ,0.01 ,0.01 P,0.01 P,0.01

Abbreviations as in Table 1.

Fig. 1. Survival curves for all patients subdivided by age, etiology of heart failure, NYHA class, EF, VO 2 max , %VO 2 max , T 1 / 2 VO 2 max , presence of NSVT and inability to take ACE-inhibitors (abbreviations as in Table 1).

D. Scrutinio et al. / International Journal of Cardiology 64 (1998) 117 – 124

121

Table 3 Results of multivariate analysis Variable

P

RR (C.I.)

Age Etiology NYHA EF VO 2 max %VO 2 max T 1 / 2 VO 2 max NSVT ACE

ns ns ns ,0.01 ,0.0001 ns ,0.05 ,0.01 ,0.001

– 1.59 (1.09–2.32) 2.17 (1.52–3.12) – 1.49 (1.02–2.14) 1.75 (1.17–2.59) 1.83 (1.28–2.32)

RR (C.I.)5relative risk (95% confidence interval). Other abbreviations as in Table 1.

using ROC analysis. The discriminatory accuracy of VO 2 max (0.74) was not significantly greater than that of %VO 2 max (0.70) (P50.3) (Fig. 2).

Fig. 3. Plot of VO 2 max versus T 1 / 2 VO 2 max (abbreviations as in Table 1).

T 1 / 2 VO 2 max higher in males. The relationship between T 1 / 2 VO 2 max and VO 2 max is shown in Fig. 3.

3.2. Determinants of T1 / 2VO2 max Among the examined variables (age, gender, etiology, NYHA class, EF, VO 2 max and ACE-inhibitor treatment), multivariate analysis identified two significant determinants of T 1 / 2 VO 2 max , they were VO 2 max (P50.0001) and gender (P50.016), being

4. Discussion We investigated the prognostic value of %VO 2 max and T 1 / 2 VO 2 max in mild to moderately symptomatic CHF patients with left ventricular systolic dysfunction. Our results show that %VO 2 max determination does not provide additional prognostic information to the absolute value of VO 2 max , while the rate of recovery of oxygen uptake after exercise is independently associated with prognosis.

4.1. Percent achieved of predicted peak oxygen uptake

Fig. 2. Receiver operating characteristics (ROC) curves for cardiac death for VO 2 max (continuous line) and %VO 2 max (dotted line); for each variable the area under the curve is given along with its standard error. The area under the curve is a measure of the discriminatory accuracy of the variable in predicting an outcome; its maximum value is 1. The two areas were compared by means of Z statistics. (Abbreviations as in Table 1).

Consistent with previous multiple studies [5–11], our data confirm the powerful value of VO 2 max for risk stratification of patients with CHF. The determination of %VO 2 max did not enhance risk stratification compared to the absolute value of VO 2 max normalized for body weight. The relative prognostic role of %VO 2 max has been investigated only in a few recently published studies [15–17]. Stelken et al. [15] studied 181 patients with ischemic or dilated cardiomyopathy. Although both %VO 2 max and VO 2 max were independent predictors of cardiac death or Status 1 transplant priority, %VO 2 max and the cut point ,50% provided prognostic information exceeding that provided by measurement of VO 2 max alone. However, while the overall discriminatory accuracy

122

D. Scrutinio et al. / International Journal of Cardiology 64 (1998) 117 – 124

using ROC analysis for %VO 2 max was significantly greater than that for VO 2 max for cardiac death or Status 1 priority, the difference was not significant for cardiac death alone. Some differences between our study and that of Stelken et al. [15] should be considered. The present study was prospective while that of Stelken et al. was retrospective. Moreover, the proportion of patients with ischemic dilated cardiomyopathy was lower in our study (37% vs 52%). This may be relevant as the magnitude of prognostic separation afforded by %VO 2 max in the study of Stelken et al. [15] was greater in the ischemic than in the nonischemic cardiomyopathy subgroup. In another study by de Groote et al. [17], VO 2 max was not an independent prognostic marker when it was compared with %VO 2 max . Survival data were however analyzed, once again, retrospectively. Moreover, exercise testing was performed on a bicycle thus the results cannot be applied to other protocols. In the study of Di Salvo et al. [16], neither %VO 2 max nor VO 2 max were significant predictors of the combined end-point of death or hospitalization for inotropic or mechanical bridging to cardiac transplantation. Finally, percent of predicted VO 2 was not found to yield better risk stratification than the absolute value in selected patients referred for cardiac transplantation [24]. A major caveat with using %VO 2 max is that there is a lack of uniformity among predicted maximal VO 2 [12,13] and most of these are based on bicycle ergometric tests. Thus, using such values to represent a hypothetical control group in determining %VO 2 max of CHF patients may be questionable and cause inconsistency of results.

4.2. Recovery of oxygen uptake The recovery of oxygen uptake after exercise in CHF was shown to be delayed in parallel with the severity of the disease [14]. Although it is likely that multiple mechanisms are involved in the prolongation of recovery of oxygen uptake in CHF, slowed replenishment of energy stores in peripheral skeletal muscles and C0 2 retention may be the most important [14]. Consistently with previous studies [14,17], we found T 1 / 2 VO 2 max to be closely correlated with VO 2 max but not with EF. These features suggest that the same mechanisms responsible for impaired exer-

cise tolerance may be involved in the delayed recovery of oxygen uptake [17] and the degree of left ventricular dysfunction does not play a role in determining the kinetics of recovery. To date, the role of T 1 / 2 VO 2 max in risk stratifying CHF patients has been poorly investigated. The present study shows that T 1 / 2 VO 2 max is significantly though weakly associated with prognosis. It is noteworthy that the relative risk for cardiac death associated with a more markedly delayed recovery of oxygen consumption was comparable to that with a more severely depressed EF, a known marker of prognosis. Our results are in line with those in the recent study by de Groote et al. [17] who found T 1 / 2 VO 2 max to be of borderline statistical significance in predicting prognosis in patients with moderate exercise intolerance (average VO 2 max of 17.3 ml / kg / min). Though T 1 / 2 VO 2 max significantly predicts prognosis, its role appears to be only marginal compared to VO 2 max . Given this finding, how might T 1 / 2 VO 2 max be used in the evaluation of CHF patients? Since determination of VO 2 max requires maximal symptom-limited exercise by patients who are not used to performing intense activities and are often poorly motivated, VO 2 max can not be reliably determined in a substantial number of patients. T 1 / 2 VO 2 max , which can be derived even when exercise is submaximal [14], may be used to enhance risk stratification of those patients who are not capable of performing a maximal exercise test. This hypothesis should however be investigated in ad-hoc designed studies.

4.3. Other prognostic markers Previous large studies have established that EF is closely correlated with prognosis in CHF [10] as well as that, ACE-inhibitor treatment significantly decreases the risk of death [2–4]. Our results are in line with those studies: low EFs and inability to take ACE-inhibitors were associated with a worse prognosis. The presence of NSVT also resulted as being an independent predictor of death in our cohort of patients; the patients who had NSVT were 1.7 times as likely to die as those without. This finding is consistent with most of the previous studies that have confirmed an association between NSVT and overall cardiac mortality [25]. In the recent study of Doval et al. [26] including patients with severe CHF, NSVT

D. Scrutinio et al. / International Journal of Cardiology 64 (1998) 117 – 124

was an independent marker for increased overall mortality rate and sudden death. The relative risk for death was 1.6, a value very close to that in our study.

4.4. Limitations of the study Ejection fraction was measured by echocardiography; although this technique may slightly differ from others, it is the most widely used. Not all patients underwent coronary angiography; as a consequence, we cannot exclude that a few patients classified as idiopathic dilated cardiomyopathy had coronary artery disease. Coronary angiography is however not indicated in all patients with a clinical diagnosis of idiopathic dilated cardiomyopathy [27]. Moreover, there may be patients with significant coronary atherosclerosis whose myocardial dysfunction appears to be disproportionate to the extent of coronary disease [28]. By definition, such patients should be classified as ischemic; yet, a cardiomyopathic process may exist and have prognostic and therapeutic implications [28]. Oxygen uptake values were averaged every 20 s throughout recovery. Although using shorter sampling intervals would increase the accuracy of measuring T 1 / 2 VO 2 max , it is unlikely that the possibly resulting small differences would influence its relation with prognosis. Other known prognostic variables, such as haemodynamic and neurohormonal parameters, were not evaluated since data were not available for all patients. However, cardiac catheterization is not indicated in all patients with CHF; it should be reserved for patients with severe symptoms for tailoring medical therapy [27,29]. Although elevated plasma concentrations of neurohormones have prognostic value [30], they are seldom used clinically.

5. Conclusions The present study suggests that, in mild to moderate CHF patients, (1) determination of the percent achieved of predicted peak oxygen uptake does not provide additional prognostic information to the absolute value of VO 2 max ; (2) the kinetics of oxygen consumption after exercise as determined by the time to reach 50% of the peak exercise value is significantly, though weakly, associated with prognosis.

123

Acknowledgements The authors would like to thank Mrs. Saba Campanale for technical assistance in collecting and analyzing data.

References [1] Bourassa MG, Gurne´ O, Bangdiwala SI, Ghali JK et al. Natural history and patterns of current practice in heart failure. J Am Coll Cardiol 1993;22(Suppl A):14A–9A. [2] The Consensus Trial Study Group. Effects of enalapril on mortality in severe congestive heart failure: results of the Cooperative North Scandinavian Survival Study (CONSENSUS). New Engl J Med 1987;316:1429–35. [3] The SOLVD Investigators. Effect of enalapril on survival in patients with reduced left ventricular ejection fraction and congestive heart failure. New Engl J Med 1991;325:293–302. [4] Cohn JN, Johnson G, Ziesche S et al. A comparison of enalapril with hydralazine-isosorbide dinitrate in the treatment of chronic congestive heart failure. New Engl J Med 1991;325:303–10. [5] Szlachcic J, Massie BM, Kramer BL, Topic N, Tubau J. Correlates and prognostic implication of exercise capacity in chronic congestive heart failure. Am J Cardiol 1985;55:1037–42. [6] Likoff MJ, Chandler SL, Kay HR. Clinical determinants of mortality in chronic congestive heart failure secondary to idiopathic dilated or to ischemic cardiomyopathy. Am J Cardiol 1987;59:634–8. [7] van den Broek SAJ, van Veldhuisen DJ, de Graeff PA, Landsman MLJ, Hillege H, Lie KI. Comparison between New York Heart Association classification and peak oxygen consumption in the assessment of functional status and prognosis in mild to moderate chronic congestive heart failure secondary to either ischemic or idiopathic dilated cardiomyopathy. Am J Cardiol 1992;70:359–63. [8] Parameshwar J, Keegan J, Sparrow J, Sutton GC, Poole-Wilson PA. Predictors of prognosis in severe chronic heart failure. Am Heart J 1992;123:421–6. [9] Roul G, Moulichon ME, Bareiss P et al. Prognostic value of peak oxygen uptake during exercise in chronic congestive heart failure. Eur Heart J 1994;15:495–502. [10] Cohn JN, Johnson GR, Shabetai R et al. Ejection fraction, peak oxygen consumption, cardiothoracic ratio, ventricular arrhythmias, and plasma norepinephrine as determinants of prognosis in heart failure. Circulation 1993;87:VI5–VI16. [11] Mancini DM, Eisen H, Kussmaul W, Mull R, Edmunds LH, Wilson JR. Value of peak exercise oxygen consumption for optimal timing of cardiac transplantation in ambulatory patients with heart failure. Circulation 1991;83:778–86. [12] Astrand P-O, Rodahl K. Physical work capacity. In: Black NW, Belser JR, editors. Textbook of work physiology. Physiological bases of exercise. New York: McGraw-Hill, 1977:289–330. [13] Fletcher GF, Balady G, Froelicher VF, Hartley LH, Haskell WL, Pollock ML. Exercise standards. A statement for health care professionals from the American Heart Association. Circulation 1995;91:580–615. [14] Cohen-Solal A, Laperche T, Morvan D, Geneves M, Caviezel B, Gourgon R. Prolonged kinetics of recovery of oxygen consumption after maximal graded exercise in patients with chronic heart failure. Analysis with gas exchange measurements and NMR spectroscopy. Circulation 1995;91:2924–32.

124

D. Scrutinio et al. / International Journal of Cardiology 64 (1998) 117 – 124

[15] Stelken AM, Younis LT, Jennison SK et al. Prognostic value of cardiopulmonary exercise testing using percent achieved of predicted peak oxygen uptake for patients with ischemic and dilated cardiomyopathy. J Am Coll Cardiol 1996;27:345–52. [16] Di Salvo TG, Mathier M, Semigran MJ, Dec GW. Preserved right ventricular ejection fraction predicts exercise capacity and survival in advanced heart failure. J Am Coll Cardiol 1995;25:1143–53. [17] de Groote P, Millaire A, Decoulx E, Nugue O, Guimier P, Ducloux G. Kinetics of oxygen consumption during and after exercise in patients with dilated cardiomyopathy. New markers of exercise intolerance with clinical implications. J Am Coll Cardiol 1996;28:168–75. [18] Yeh M, Gardner R, Adams T, Yarowitz F, Crapo R. Anaerobic threshold: problems of determination and validation. J Appl Physiol 1983;55:1178–86. [19] Jones, NL. Clinical exercise testing. Philadelphia: W.B. Saunders, 1986. [20] Cox DR. Regression models and life tables (with discussion). J R Stat Soc (B) 1972;34:187–202. [21] Bigger JT, Fleiss JL, Steinman RC, Rolnitzky LM, Kleiger RE, Rottman JN. Frequency domain measures of heart period variability and mortality after myocardial infarction. Circulation 1992;85:164– 71. [22] Peto R, Pike MC, Armitage P et al. Design and analysis of randomized clinical trials requiring prolonged observations of each patient II. Analysis and examples. Br J Cancer 1977;35:1–39.

[23] Hanley JA, McNeil BJ. A method of comparing the areas under Receiver Operating Characteristic Curves derived from the same cases. Radiology 1983;148:838–43. [24] Aaronson K, Chen T, Mancini DS. Is percent predicted VO 2 a better selection riterion than peak VO 2 for cardiac transplantation? J Heart Lung Transplant 1995;14:981–9. [25] Bigger JT. Why patients with congestive heart failure die: arrhythmias and sudden death. Circulation 1987;75(suppl IV):IV28–35. [26] Doval HC, Nul DR, Grancelli HO et al. Nonsustained ventricular tachycardia in severe heart failure. Independent marker of increased mortality due to sudden death. Circulation 1996;94:3198–203. [27] Dec GW, Fuster V. Idiopathic dilated cardiomyopathy. New Engl J Med 1994;331:1564–75. [28] Abelmann WH, Lorell BH. The challenge of cardiomyopathy. J Am Coll Cardiol 1989;13:1219–39. [29] Stevenson LW, Tillisch YH, Hamiltom MA et al. Importance of hemodynamic response to therapy in predicting survival with ejection fraction ,20% secondary to ischemic or nonischemic dilated cardiomyopathy. Am J Cardiol 1990;66:1348–54. [30] Rector TS, Olivari MT, Levine TB, Francis GS, Cohn JN. Predicting survival for an individual with congestive heart failure using the plasma norepinephrine concentration. Am Heart J 1987;114:148–52.