Echocardiographic determinants of mortality in patients >67 years of age with chronic heart failure

Echocardiographic determinants of mortality in patients >67 years of age with chronic heart failure

Echocardiographic Determinants of Mortality in Patients >67 Years of Age With Chronic Heart Failure Viorel G. Florea, PhD, DSc, Michael Y. Henein, MD,...

83KB Sizes 0 Downloads 30 Views

Echocardiographic Determinants of Mortality in Patients >67 Years of Age With Chronic Heart Failure Viorel G. Florea, PhD, DSc, Michael Y. Henein, MD, PhD, Mariantonietta Cicoira, MD, Stefan D. Anker, MD, PhD, Wolfram Doehner, MD, Piotr Ponikowski, MD, PhD, Darrel P. Francis, MB, MRCP, Derek G. Gibson, MB, and Andrew J.S. Coats, DM This study sought to assess the prognostic significance of echocardiographic measurements of left and right ventricular dimensions and function in patients >67 years of age with chronic congestive heart failure (CHF). This is a retrospective follow-up of elderly patients who underwent an echocardiography in the tertiary cardiac center. A total of 185 patients (131 men) aged >68 years (mean ⴞ SD 75 ⴞ 5) with CHF were enrolled into the study. After undergoing a detailed echocardiographic examination, all patients were followed-up for a median of 20 months (interquartile range 9 to 36). During the follow-up period 54 patients (29%) died. Left ventricular (LV) M-mode isovolumic relaxation time (IVRT), end-diastolic and end-systolic diameters, fractional shortening and mass, transmitral E:A ratio, and left atrial dimension, as well as New York Heart Association class and the age were found by Cox proportional-hazards uni-

variate analyses to predict the outcome in these patients (all p <0.05). In multivariate analyses including these measurements, LV IVRT (p <0.04), age (p <0.03), and New York Heart Association class (p <0.001) were found to be the independent predictors of outcome. In the Kaplan-Meier analysis, patients with LV IVRT >30 ms had a better prognosis at 3 years (cumulative survival 78% [95% confidence interval 65% to 91%]) than those with LV IVRT <30 ms (survival 52% [95% confidence interval 37% to 68%]). Measurements of LV performance, especially those obtained during diastole, are significantly related to prognosis in patients >67 years of age with CHF. LV M-mode IVRT is among the most important independent predictors of outcome in this population. 䊚2000 by Excerpta Medica, Inc. (Am J Cardiol 2000;86:158 –161)

lthough most cardiac diseases can affect left and right ventricles and left ventricular (LV) failure A may secondarily impair right ventricular (RV) dia-

studied with no exclusions. We studied elderly patients from the age of ⱖ68 years, consistent with the definition of aging thresholds of the population.2,3 The diagnosis of heart failure was based on history, examination, electrocardiogram, chest radiograph, and echocardiographic findings. We took the European Society of Cardiology guidelines definition that the diagnosis of heart failure would be made if both of the following were present: symptoms compatible with a diagnosis of heart failure, mainly exertional breathlessness, and evidence of substantial impairment of LV systolic function or LV filling on Doppler echocardiography.4 Patient characteristics are listed in Table I. They ranged in age from 68 to 87 years (mean 75 ⫾ 5). Twenty-eight patients were in New York Heart Association (NYHA) functional class I, 83 in class II, 51 in class III, and the remaining 23 were in class IV. Heart failure was of ischemic origin in 116 patients and of nonischemic origin in 69 patients. The presence of ischemic heart disease was shown by coronary arteriography or documented myocardial infarction. No patient had significant primary valvular disease, 149 patients were in sinus rhythm, and 52 patients had either left or right bundle branch block. The medical regimens of all the enrolled patients were optimized according to standard practice at the time of enrollment into the study and all were symptomatically stable. Standardized medications included digoxin, an angiotensin-converting enzyme inhibitor, diuretics,

stolic performance,1 there is little information about how RV abnormalities relate to prognosis in elderly patients with chronic congestive heart failure (CHF). This investigation assesses the prognostic value of echocardiographic measurements of LV and RV dimensions and function in patients ⬎67 years of age with CHF, with emphasis on diastolic variables obtained by M-mode and Doppler echocardiography.

METHODS

Patients: The target population for this study was all elderly patients with CHF referred for echocardiographic examination as part of their routine assessment at the Royal Brompton Hospital between 1988 and 1998. A total of 185 consecutive patients were

From the Department of Cardiac Medicine, National Heart and Lung Institute, London; Department of Cardiology, Royal Brompton and Harefield NHS Trust, London, United Kingdom; and Franz-VolhardKlinik, Max-Debru¨ck-Centrum, Charite´, Campus Berlin-Buch, Berlin, Germany. Dr. Florea was supported by a research fellowship from the European Society of Cardiology. Professor Coats is sponsored by the Viscount Royston Trust, London, United Kingdom. Manuscript received October 19, 1999; revised manuscript received and accepted January 19, 2000. Address for reprints: Viorel G. Florea, PhD, DSC, National Heart and Lung Institute, Dovehouse Street, London SW3 6LY, United Kingdom.

158

©2000 by Excerpta Medica, Inc. All rights reserved. The American Journal of Cardiology Vol. 86 July 15, 2000

0002-9149/00/$–see front matter PII S0002-9149(00)00853-5

TABLE I Patient Characteristics (n ⫽ 185) Demographics Age (yrs) Men NYHA class I II III IV Sinus rhythm Left bundle branch block Right bundle branch block Medication Digoxin Angiotensin-converting enzyme inhibitors Diuretics Calcium channel antagonists Nitrates ␤ blockers Aspirin or warfarin

TABLE II Results of Echocardiography at Time of Enrollment (mean ⫾ SD) 75 ⫾ 5 126 (68%) 28 83 51 23 149 37 15

(15%) (45%) (28%) (12%) (81%) (20%) (8%)

44 118 114 39 33 17 120

(24%) (64%) (62%) (21%) (18%) (9%) (65%)

Left ventricle End-diastolic diameter (mm) End-systolic diameter (mm) Mass (g) (n ⫽ 148) Fractional shortening (%) Isovolumic relaxation time (ms) (n ⫽ 144) Peak E diastolic filling velocity (m/s) (n ⫽ 149) Peak A diastolic filling velocity (m/s) (n ⫽ 147) E/A ratio (n ⫽ 147) E-wave deceleration time (ms) (n ⫽ 149) Left atrium Dimension (mm) (n ⫽ 161) Right ventricle Isovolumic relaxation time (ms) (n ⫽ 91) Peak E diastolic filling velocity (m/s) (n ⫽ 99) E-wave deceleration time (ms) (n ⫽ 99)

59 45 400 24 30 0.7 0.5 2.9 61

⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

13 16 200 14 18 0.3 0.3 2.9 21

46 ⫾ 9 27 ⫾ 21 0.4 ⫾ 0.2 60 ⫾ 15

Values are expressed as number (%) or mean ⫾ SD.

calcium antagonists, nitrates, ␤ blockers, and aspirin or warfarin in varying combinations (Table I). Study design: This is a retrospective study of 185 elderly patients with CHF who had undergone a routine Doppler echocardiographic examination. The patients were followed up for a median of 20 months (interquartile range 9 to 36) in the heart failure outpatient clinic at the Royal Brompton Hospital, London. If follow-up did not occur in the hospital, survival information was obtained from the general practitioner or the patient’s local hospital. The study end point was all-cause mortality. Procedures: Simultaneous Doppler and M-mode echocardiograms and phonocardiograms were recorded along with standard lead II of the electrocardiogram with the patient supine and in the left semilateral position. All patients were studied at rest and in quiet respiration. Echocardiograms were recorded using a HewlettPackard Sonos 1500 echocardiograph (Andover, Massachusetts) with a 2.5-MHz transducer. The pattern of LV wall motion was assessed from standard left parasternal and apical views. Systolic and diastolic LV dimensions and septal and posterior wall thicknesses were measured from the M-mode echocardiograms of the LV minor axis obtained with the cursor by the tips of mitral valve leaflets using leading-edge methods. End-diastole was taken as the onset of the Q wave of the simultaneously recorded electrocardiogram, and end-systole as the onset of the aortic component of the second heart sound (A2) on the phonocardiogram. All echocardiographic recordings were obtained at a paper speed of 100 mm/s. LV mass was calculated using the Penn convention: LV mass ⫽ 1.04 䡠 {(EDD ⫹ PWTD ⫹ VSTD)3 – (EDD)3} –13.6 g, where EDD ⫽ end-diastolic dimension; PWTD ⫽ posterior wall thickness in diastole, and VSTD ⫽ ventricular septal thickness in diastole. LV fractional shortening was estimated as the percent decrease in dimension during ejection with respect to end-diastolic dimension. LV M-mode iso-

volumic relaxation time (IVRT) was measured as the time interval from the A2 to the onset of mitral cusp separation on the mitral echogram.5 RV IVRT was measured as the time interval between P2 (the pulmonary component of the second heart sound) and onset of Doppler tricuspid forward flow. From the transmitral and transtricuspid pulsed-Doppler traces, peak early (E) and late (A) diastolic filling velocities were measured and the E/A ratio calculated. Mitral and tricuspid E-wave deceleration times were measured from the peaks of the E waves to their end. Phonocardiograms were recorded from the right or left sternal edge, using the Cambridge Instrument Company microphone (Cambridge, United Kingdom), in the position where A2 was most obvious. The identity of A2 itself was checked against aortic valve closure artifact on pulsed Doppler, and its timing was taken as that of the onset of the first high-frequency component. Statistical analysis: Univariate and multivariate Cox proportional-hazards regression analysis was used to identify variables predictive of outcomes. Survival curves were constructed using the Kaplan-Meier product limit survival curve method when patients were dichotomized by the median value of LV IVRT. For all tests, a p value ⬍0.05 was considered statistically significant. Descriptive values are expressed as mean ⫾ SD. Statistical analysis was performed using a standard statistical program package (StatView, version 4.5, Abacus Concepts Inc., Berkeley, California).

RESULTS Table II summarizes the main echocardiographic measurements for all 185 patients with CHF at the time of enrollment. As shown, in the patient population as a whole, the left ventricle was modestly dilated (as indicated by end-diastolic diameter of 59 ⫾ 13 mm and end-systolic diameter of 45 ⫾ 16 mm), LV mass was increased (400 ⫾ 200 g), and LV systolic function was slightly impaired, indicated by mean LV fractional shortening of 24 ⫾ 14%. At the same time, 42 patients had a normal LV cavity size (i.e., the

CONGESTIVE HEART FAILURE/PROGNOSIS IN ELDERLY WITH HEART FAILURE

159

wave velocity, E/A ratio and E-wave deceleration time, and left atrial dimension, as well as the age and NYHA class were found by Cox proDirection Associated With portional-hazards univariate analyses Variable Chi-Square p Value Increased Mortality to be significant predictors of the Left ventricle outcome in these patients (all p End-diastolic diameter 16.0 0.0001 Increased ⬍0.05). In multivariate analysis (TaEnd-systolic diameter 17.6 0.0001 Increased ble IV) including the age, NYHA Mass 7.6 0.0057 Increased Fractional shortening 10.0 0.0015 Decreased class, left atrial dimension, and 6 LV Isovolumic relaxation time 14.4 0.0001 Decreased echocardiographic measurements— Peak E diastolic filling velocity 0.5 0.8267 Increased namely end-diastolic and end-sysPeak A diastolic filling velocity 6.1 0.0134 Decreased tolic dimensions, mass and fractional E/A ratio 6.6 0.0105 Increased shortening, M-mode IVRT and E/A E-wave deceleration time 1.0 0.3246 Decreased Left atrium ratio—the NYHA class (p ⬍0.001), Dimension 7.1 0.0077 Increased age (p ⬍0.03), and the LV M-mode Right ventricle IVRT (p ⬍0.04) were found to be the Isovolumic relaxation time 5.5 0.0189 Increased most important independent predicPeak E diastolic filling velocity 0.5 0.4698 Decreased E-wave deceleration time 3.5 0.0597 Decreased tors of outcome. Age 13.2 0.0003 Increased Technically satisfactory measureNYHA class 45.9 0.0001 Increased ments of the LV M-mode IVRT were available in 144 patients; it ranged from 0 to 80 ms. Its duration was related neither to the presence of a complete left or TABLE IV Results of Multivariate Cox Proportional-Hazards right bundle branch block nor to the presence of a Regression Analysis of Echocardiographic Variables, Age, QRS complex on the electrocardiogram (r ⫽ ⫺0.03, and NYHA Class in 185 Elderly Patients With Chronic Heart Failure p ⫽ 0.75). Figure 1 shows survival Kaplan-Meier curves when the patients studied were dichotomized Variable Chi-Square p Value by the median value of LV IVRT of 30 ms. Patients LV end-diastolic diameter 0.9 0.3424 with LV IVRT ⬎30 ms (n ⫽ 65) had a better LV end-systolic diameter 1.5 0.2177 prognosis at 3 years (cumulative survival 78% [95% LV mass 1.5 0.2285 confidence interval 65% to 91%]) than those with LV fractional shortening 3.6 0.0577 LV isovolumic relaxation time 4.3 0.0374 LV IVRT ⱕ30 ms (n ⫽ 79, cumulative survival LV E/A ratio 0.6 0.8150 52% [95% confidence interval 37% to 68%]). TABLE III Results of Univariate Cox Proportional-Hazards Regression Analysis of Echocardiographic Variables, Age, and NYHA Class in 185 Elderly Patients With Chronic Heart Failure

Left atrial dimension Age NYHA class

0.2 5.0 11.3

0.6613 0.0257 0.0008

end-diastolic diameter was ⬍55 mm) and normal systolic function (i.e., fractional shortening ⬎29%). Measurements of left- and right-sided filling and relaxation velocities in the entire group of patients indicated moderate LV and RV diastolic dysfunction in terms of a restrictive pattern: increased peak Ewave velocities (0.7 ⫾ 0.3 for mitral and 0.4 ⫾ 0.2 for tricuspid flow), and short left and right E-wave deceleration time (61 ⫾ 21 and 60 ⫾ 15 ms, respectively). The left atrium in the patient group as a whole was slightly enlarged (46 ⫾ 9 mm) compared with the upper limit of the normal range of 42 mm in the elderly control population.6,7 During the median 20-month follow-up (interquartile range 9 to 36 months), 54 patients (29%) died. Of the 42 patients with normal LV cavity size and normal systolic function, 4 (10%) died. Of the remaining 143 patients, 50 (35%) died. The results of univariate Cox proportional-hazards regression analysis of studied echocardiographic variables, age, and NYHA class are presented in Table III. The values of LV end-diastolic and end-systolic dimensions, mass and fractional shortening, LV and RV IVRT, transmitral peak A160 THE AMERICAN JOURNAL OF CARDIOLOGY姞

VOL. 86

DISCUSSION Our main observation was to demonstrate that in addition to age and NYHA class, the most discriminating measurement in determining prognosis in older patients with CHF was LV M-mode IVRT. Other diastolic measurements, such as A-wave velocity, Ewave deceleration time, and E/A ratio had prognostic significance, but in the proportional-hazards regression model, these variables were no longer significant once the effects of IVRT had been allowed. Indeed, the same applied to end-diastolic and end-systolic dimensions, with only fractional shortening achieving borderline significance. It is well established that measurements of LV diastolic function are age related.8,9 In particular, in normal patients, there is progressive prolongation of IVRT,10 and the proportion of the stroke volume entering the ventricle during early diastole decreases as that during atrial systole increases. Like many other diastolic measurements, IVRT is sensitive both to disturbances of diastolic function and to ventricular filling pressure. Abnormal prolongation of IVRT is evidence of abnormal relaxation, whereas values lower than normal are consistently seen when LV filling pressure is increased. The mean value of IVRT of 30 ms in the present study, compared with a normal of 70 to 80 ms10 is thus clear evidence that left atrial JULY 15, 2000

agreement as to whether this combination does, in fact, constitute a discrete clinical entity. Our results would suggest that this combination would have prognostic significance only when IVRT is short (i.e., in patients with increased filling pressures, and a dominant E wave on transmitral Doppler). Such patients thus conform to the clinical diagnosis of restrictive cardiomyopathy. By contrast, patients with normal cavity size, with IVRT prolonged, E-wave suppressed, and A-wave accentuated abnormally for age (i.e., those to whom the diagnosis of “diastolic heart FIGURE 1. Kaplan-Meier curves for all patients and for those with IVRT >30 failure” is usually applied) would seem to and <30 ms. have a particularly good prognosis. However, our results do not indicate whether pressure was raised throughout the group, in line with this combination can be the direct cause of significant the clinical diagnosis of heart failure. The limited limitation of exercise tolerance. prognosis of the patients’ value of IVRT below the median value of 30 ms is thus likely to be related to a significantly raised ventricular filling pressure persist- 1. Henein MY, O’Sullivan CA, Coats AJS, Gibson DG. Angiotensin-converting inhibitors revert abnormal right ventricular filling in patients with reing despite standard medical treatment. The duration enzyme strictive left ventricular disease. J Am Coll Cardiol 1998;32:1187–1193. of IVRT has also been shown to have a powerful 2. Coats AJS. Is preventive medicine responsible for the increasing prevalence of predictive effect of subsequent diastolic events, so that heart failure? Lancet 1998;352(suppl I):39 – 41. McCreadie C, Tinker A. Abuse of elderly people in the domestic setting: a UK when it is short, E-wave velocity is likely to be in- 3. perspective. Age Ageing 1993;22:65– 69. 11,12 creased, and E/A ratio high. Our results are thus in 4. Task Force on Heart Failure of the European Society of Cardiology. Guidelines accord with previous results demonstrating that a re- for the diagnosis of heart failure. Eur Heart J 1995;16:741–751. Mattheos M, Shapiro E, Oldershaw PJ, Sacchetti R, Gibson DG. Non-invasive strictive ventricular filling pattern persisting with 5. assessment of change in left ventricular relaxation by combined phono-, echo-, medical treatment is also a marker of poor progno- and mechanocardiography. Br Heart J 1982;47:253–260. Aguirre FV, Pearson AC, Lewen MK, McCluskey M, Labovitz AJ. Usefulness sis13–15; when directly compared with these measure- 6. of Doppler echocardiography in the diagnosis of congestive heart failure. Am J ments based on filling pattern, a short IVRT appears Cardiol 1989;63:1098 –1102. 7. Appleton CP, Hatle LK, Popp RL. Relation of transmitral flow velocity more discriminating. to left ventricular diastolic function: new insights from a combined Our measurements of IVRT were based on the patterns hemodynamic and Doppler echocardiographic study. J Am Coll Cardiol 1988; interval from the onset of A2 on the phonocardiogram 12:426 – 440. and confirmed by its coincidence with the aortic clo- 8. Wei JY. Age and the cardiovascular system. N Engl J Med 1992;327:1735– sure artifact on the Doppler, and with mitral cusp 1739. 9. Tresch DD, McGough MF. Heart failure with normal systolic function: a separation on the M-mode echocardiogram. This def- common disorder in older people. J Am Geriatr Soc 1995;43:1035–1042. 10. Gibson D. Diastolic heart failure. In: Poole-Wilson PA, Colucci WS, Massie inition is in line with the original one by Wiggers.16 BM, Chatterjee K, Coats AJS, eds. Heart Failure. New York, NY: Churchill Importantly, there are significant discrepancies in tim- Livingstone, 1997:339 –342. ing between the timing of mitral cusp separation on 11. Lee CH, Hogan JC, Gibson DG. Diastolic disease in left ventricular hypercomparison of M mode and Doppler echocardiography for the assessment the M-mode and that of the onset of transmitral flow trophy: of rapid ventricular filling. Br Heart J 1991;65:194 –200. 17 from the Doppler. In normal subjects, the onset of 12. Brecker SJD, Lee CH, Gibson DG. Relation of left ventricular isovolumic transmitral flow at the level of the tips of the mitral relaxation time and incoordination to transmitral Doppler filling patterns. Br Heart J 1992;68:567–573. cusps follows cusp separation by 20 to 30 ms, but in 13. Pinamoti B, Di-Lenarda A, Sinagra G, Camerini F. Restrictive left ventricular patients with dilated cardiomyopathy and functional filling pattern in dilated cardiomyopathy assessed by Doppler echocardiography: echocardiographic and hemodynamic correlations and prognostic implimitral regurgitation, the discrepancy may amount to clinical, cations. Heart muscle Disease Study Group. J Am Coll Cardiol 1993;22:808 – ⬎100 ms.18 Because it is now recognized that the 815. timing of the onset of flow is sensitive to the level of 14. Xie GY, Berk MR, Smith MD, Gurley JC, DeMaria AN. Prognostic value of transmitral flow patterns in patients with congestive heart failure. J Am the sample volume within the ventricle,19 we elected Doppler Coll Cardiol 1994;24:132–139. to use the M-mode definition. It does not necessarily 15. Lipsitz LA, Byrnes N, Hossain M, Douglas P, Waksmonski CA. Restrictive follow that estimates of IVRT based on Doppler will left ventricular filling patterns in very old patients with congestive heart failure: correlates and prognostic significance. J Am Geriatr Soc 1996;44:634 – demonstrate the same discriminating value as those clinical 637. based on M mode. 16. Wiggers CJ. Studies on the consecutive phases of the cardiac cycle. I. The Clinical significance: M-mode IVRT thus provides a duration of the consecutive phases of the cardiac cycle and the criteria for their precise determination. Am J Physiol 1921;56:415– 438. simple measurement that is closely related to progno- 17. Lee CH, Vancheri F, Josen MS, Gibson DG. Discrepancies in the measuresis in elderly patients, more discriminating than mea- ment of isovolumic relaxation time: a study comparing M mode and Doppler Br Heart J 1990;64:214 –218. surements based on diastolic inflow patterns or acti- echocardiography. 18. Ng KSK, Gibson DG. Impairment of diastolic function by shortened filling vation disturbances. It is also independent of ventric- period in severe left ventricular disease. Br Heart J 1989;62:246 –252. ular dimensions or shortening fraction, and so our 19. Brun P, Tribouilloy C, Duval AM, Iserin L, Meguira A, Pelle G, DuboisJL. Left ventricular flow propagation during early filling is related to wall results apply to patients with the combination of heart Rande relaxation: a color M-mode Doppler analysis. J Am Coll Cardiol 1992;20:420 – failure and normal cavity size. There is no general 432. CONGESTIVE HEART FAILURE/PROGNOSIS IN ELDERLY WITH HEART FAILURE

161