dt is a predictor of mortality in patients with heart failure

Noninvasively determined radial dP/dt is a predictor of mortality in patients with heart failure Jean-Michel Tartie´re, MD,a Jean-Yves Tabet, MD,b Damien Logeart, MD, PhD,b Lamia Tartie`re-Kesri, MD,b Florence Beauvais, MD,b Christophe Chavelas, MD,b and Alain Cohen Solal, MD, PhDb Paris, France

Background The left ventricular (LV) developed pressure is a marker of contractility, associated with a poor prognosis during systolic heart failure. The maximal first derivative or slope of the radial pulse wave (Rad dP/dt) has been proposed as a marker of LV systolic function. This study sought to assess the prognostic value of the baseline dP/dt of the radial pulse in patients with heart failure. Methods The Rad dP/dt was noninvasively measured by applanation tonometry, and its effect on mortality was analyzed by using multivariate Cox regression models. We studied 310 consecutive patients. Mean follow-up was 327 ± 187 days, and 64 patients died or were transplanted during this period. Results Death or transplantation was associated with New York Heart Association class III or IV, low systolic or mean blood pressure, low LV ejection fraction, and low Rad dP/dt (634.6 ± 373.3 vs 730.2 ± 367.4 mm Hg/s for patients who survived without transplantation, P b .02). A Rad dP/dt b440 mm Hg/s was associated with death or transplantation before and after adjustment for confounding variables (OR [95% CI] 2.19 [1.33-3.58] and 2.88 [1.29-6.38], respectively, P b .01 for both). This relationship was independent of pulse pressure and no significant interaction was found between the Rad dP/dt and the pulse pressure. Conclusion

This study demonstrates, for the first time, that the Rad dP/dt, proposed as a noninvasive peripheral marker of LV systolic function, is an independent predictor of death or transplantation in patients with HF regardless of LV ejection fraction. (Am Heart J 2008;155:758-63.)

The noninvasive assessment of the left ventricular (LV) systolic function and/or contractility has long been an important issue in heart failure. In the cardiovascular area, a large number of noninvasive parameters have been evaluated in term to characterize the LV function, principally by using echocardiography or Doppler methods or pulse analysis.1-14 Some parameters have been associated with prognosis,2-6 especially in the case of patients with LV systolic dysfunction,2,3,5,6 but all are either imprecise or, if accurate, time-consuming and complex to assess, which implies the need for trained operators. Such parameters, for example, the LV ejection fraction (LVEF), are therefore markedly underused in diagnosis,15-19 which may result in underprescribing validated medications and poor prognosis.19 The carotid arterial wave intensity at the first peak has been recently validated as a substitute of LV contractility From the Departments of aPhysiology and bCardiology, Lariboisière Hospital, Assistance Publique, Paris, France. This study was supported by a grant from the French Society of Cardiology. Submitted July 3, 2007; accepted November 12, 2007. Reprint requests: Jean-Michel Tartiére, MD, Lariboisière Hospital – Department of Physiology, Assistance Publique des Hôpitaux de Paris, 2 Rue Ambroise Paré, 75475 Paris Cedex 10, France. E-mail: [email protected] 0002-8703/$ - see front matter © 2008, Published by Mosby, Inc. doi:10.1016/j.ahj.2007.11.030

for LV dP/dt,20-22 a recognized index of LV contractility. The principal component of this parameter is the square of the noninvasive recordable dP/dt of the arterial pulse. Moreover, the oscillometric brachial dP/dt has been related to LVEF.23 The radial pulse wave can be easily recorded by applanation tonometry (SphygmoCor Px Pulse Wave Analysis system; AtCor Medical, West Ryde, Australia) using a portable device for assessing the radial pulse wave and determining the maximal first derivative or slope (Rad dP/dt).24,25 This Rad dP/dt appears to be a valuable and reproducible peripheral marker LV systolic function independently of arterial compliance and wave reflection,26 even if this parameter and its relation to the invasive LV dP/dt remain unproved and debated.27 The aim of this study was to determine the usefulness of the Rad dP/dt regardless of LVEF in the prognostic assessment of a large population of patients with heart failure.

Methods Patients To be eligible for the study, patients had to have a past history of heart failure and be in stable condition at the time of the examination. The diagnosis of heart failure was based on criteria defined by the European Society of Cardiology28 and serial measurements of brain natriuretic peptide.29,30 All patients were referred to a single center of cardiology from May 2002 to

American Heart Journal Volume 155, Number 4

October 2004 and gave their informed consent to the study. The study was approved by the institutional review board of the Beaujon Hospital, Paris, France. At enrollment, data were collected on presence of diabetes mellitus, hypertension, dyslipidemia, tobacco use (current or past), the New York Heart Association (NYHA) classification of congestive heart failure, coronary artery disease, and ongoing medications. All patients were followed up by phone call to patients or general/specialist practitioners. The primary end point was death or heart transplantation.

Tartiére et al 759

Figure 1

Blood pressure measurement Brachial systolic blood pressure (SBP) and diastolic blood pressure (DBP) were measured on the right arm by use of a mercurial sphygmomanometer after 10 min of rest, with the patient lying in the supine position. Three measurements were averaged. Pulse pressure (PP) was defined as SBP minus DBP. The radial pulse wave was calibrated by using the brachial SBP and the DBP, and mean blood pressure (MBP) was directly calculated from the calibrated radial pulse wave.

Example of radial pulse wave shape and maximal radial dP/dt.

Radial dP/dt measurement Radial pulse wave was recorded by applanation tonometry by using a high-fidelity Millar strain gauge transducer (Millar SPT-301 probe; Millar Instruments, Houston, TX) on the right radial artery, insuring a frequency response flat over 100 Hz, which is required for dP/dt assessment. The maximal first derivative (or slope) of the radial pulse wave (Rad dP/dt) was automatically calculated as previously described 24 by use of the SphygmoCor Px Pulse Wave Analysis System (AtCor Medical) (Figure 1) as radial ejection duration. The left ventricular ejection time index (LVETI) was calculated as previously described. 13,14 Two measurements were averaged from a series of waves over a 10-second period. Reproducibility of the Rad dP/dt was assessed and coefficient correlation and mean difference ± 1.96 × SD between 2 repeated measurements were r = 0.99, P b .001 and 11 ± 71 mm Hg/s, P N .126.

Left ventricular ejection fraction measurement The LVEF was measured by echocardiography (n = 190) or radionuclide ventriculography (n = 120). Echocardiographic measurement of LVEF involved use of the Simpson method (n = 66), the Teichholz method (n = 78), or a visual method (n = 46), depending on the presence or absence of segmental wall motion abnormalities and technical limitations. The reproducibility of the echocardiographic method (Simpson and Teichholz) was assessed relative to radionuclide ventriculography, 31 and the reproducibility of the visual method was assessed relative to Simpson method with the following results: Simpson method (n = 13), r = 0.89, P b.001, mean difference ± 1.96 × SD −2.07% ± 13.7%; Teichholz (n = 19), r = 0.88, P b .001, −3.2% ± 14.7%; and visual method (n = 33), r = 0.88, P b .001, 0.4% ± 8.4%. The investigators were blinded to the outcome of the patients when assessing baseline LVEF.

Statistical analysis Categorical data are presented as percentages, and continuous data as means ± SD. Mann-Whitney and Fisher

exact tests were used when indicated. Cox proportional hazards regression models were used to assess the relation of the Rad dP/dt and clinical variables with prognosis. Rad dP/dt and MBP were log-transformed to achieve a normal distribution. Rad dP/dt was evaluated both as a continuous variable (per 1 SD decrease) and as a categorical variable .The cutoff value was determined by using receiver operating characteristic curve (Figure 2). The relation between the Rad dP/dt and death or transplantation was adjusted for age, sex, heart rate (HR), MBP, PP, LVEF, NYHA class III or IV as compared with class I or II, therapy with angiotensin-converting enzyme inhibitor and/or angiotensin II receptor inhibitor (ACE/ARB), hypertension, diabetes, dyslipidemia, tobacco use (current or past), and coronary artery disease. Interactions were tested according to the incidence of combined end point: LVEF × Rad dP/dt, LVEF × LVETI and PP × Rad dP/dt. A robust regression method was used to study the relation between Rad dP/dt and other clinical and hemodynamic parameters. This procedure has been shown to be more robust than classic parametric regression to the marginal violation of normality assumption and the presence of outliers. Covariates were selected by using a manual backward stepwise regression approach. P b .05 was considered significant. All P values were 2-sided. Analyses involved use of NCSS 6.0.21 software (Kaysville, UT).

Results Baseline characteristics Table I shows the baseline characteristics and therapeutics of patients with heart failure by mortality. Patients who died had lower SBP, MBP, LVEF and Rad dP/ dt, and more often had a NYHA classification of III or IV than those who lived. They tended to have low PP and high HR and were older. There was no difference in term of medications between groups.

American Heart Journal April 2008

760 Tartiére et al

Figure 2

Table I. Baseline characteristics of patients with heart failure who were alive without transplantation (survivors) or not at follow-up Survivors

Receiver operator characteristic curve to determine the cutoff value of the Rad dP/dt.

Age, y Female, % NYHA class III or IV, % Hypertension, % Dyslipidemia, % Diabetes, % Tobacco use, % Coronary artery disease, % SBP, mm Hg DBP, mm Hg Pulse pressure, mm Hg MBP, mm Hg HR, beat/min LVETI, milliseconds LVEF, % Rad dP/dt, mm Hg/s Treatments ACE/ARB, % β-blockers, % Diuretics, % Spironolactone, % Digoxin, % Calcium antagonists, % Amiodarone, % Aspirin, %

67.3 ± 13.8 30 33 73 56 32 58 44 124.8 ± 24.0 73.7 ± 12.9 51.1 ± 21.5 90.0 ± 14.5 69.2 ± 13.9 401 ± 28 38.2 ± 16.6 730.2 ± 367.4 87 76 80 31 8 18 24 49

Nonsurvivors 70.1 ± 14.5 28 66 70 60 38 58 47 117.6 ± 26.8 70.7 ± 11.1 46.9 ± 24.7 85.3 ± 13.9 72.2 ± 14.7 402 ± 29 33.4 ± 19.0 634.6 ± 373.3 80 73 81 27 5 19 21 51

P .09 NS b.001 NS NS NS NS NS .02 NS .09 .02 .09 NS .02 .02 NS NS NS NS NS NS NS NS

NS, Nonsignificant.

Rad dP/dt and outcome During a mean follow-up of 327 ± 187 days, 64 patients died or were transplanted: 43 patients died because of cardiovascular disease, 6 patients were transplanted, 5 died because of septic chock, 4 because of cancer, 1 because of digestive hemorrhage, and 5 were unexplained. In an univariate analysis (Table II, Figure 3), death or transplantation was related to age, NYHA class-III or -IV, absence of ACE/ARB therapy, high HR and low MBP or Rad dP/dt, but not with LVETI or PP. Moreover, no significant interaction was found between the LVEF and the Rad dP/dt or the LVETI, and no significant interaction was found between the Rad dP/dt and the PP. After adjustment for confounding variables (Table III), NYHA class III or IV and low Rad dP/dt were the most powerful predictors of outcome but not HR, MBP or PP. For a decrease of 1 SD of the logtransformed Rad dP/dt (1 SD of the untransformed value: 370 mm Hg/s), the crude and adjusted OR [95% CI] were 1.34 [1.04-1.72], P = .02, and 2.50 [1.30-4.80], P = .006, respectively. For a Rad dP/dt threshold of 440 mm Hg/s, the crude and adjusted OR [95% CI] were 2.19 [1.33-3.58], P = .002; and 2.88 [1.29-6.38], P = .009, respectively. The relationship between Rad dP/dt and death or transplantation was not modified by further adjustment for LVETI (P = .01 for the log-transformed Rad dP/dt, and for Rad dP/dt b 440 mm Hg/s) and LVETI was not related to mortality before or after adjustment. When patients with a LVEF assessed by the visual method

Table II. Predictors of death or transplantation in patients with heart failure by univaried Cox proportional hazard regression analysis Crude Predictors

OR

Lower 95% CI

Higher 95% CI

P

Age, 1 SD increase NYHA class III or IV ACE/ARB, current use HR, 1 SD increase Pulse pressure, 1 SD increase Mean BP, 1 SD decrease Rad dP/dt, 1 SD decrease Rad dP/dt b440 mm Hg/s

1.30 3.37 0.59

1.00 2.01 0.32

1.70 5.64 1.08

.049 b.001 .09

1.24 0.85

0.98 0.65

1.57 1.10

.08 NS

1.40

1.09

1.79

.007

1.34

1.04

1.71

.02

2.19

1.33

3.58

.002

were withdrawn of the statistical analysis relation between Rad dP/dt and prognosis was unchanged: crude OR (95%CI), 2.48 (1.48-4.22), P b .001; adjusted OR 2.92 (1.38-6.14), P = .005. Because of a potentially significant colinearity between PP and Rad dP/dt, a separate statistical analysis was performed without PP as a covariate and the relationship between the log transformed Rad dP/dt or the Rad dP/dt b440 mm Hg/s and prognosis was not modified (P = .02

American Heart Journal Volume 155, Number 4

Figure 3

Tartiére et al 761

Table III. Predictors of death or transplantation in multivariate Cox proportional hazard regression analysis Adjusted Predictors Age, 1 SD increase NYHA class III or IV ACE/ARB, current use HR, 1 SD increase Pulse pressure, 1 SD increase Mean BP, 1 SD decrease Rad dP/dt, 1 SD decrease Rad dP/dt b440 mm Hg/s

OR

Lower 95% CI

Higher 95% CI

P

1.38 3.04 0.47 1.24 1.90

0.99 1.72 0.23 0.93 0.90

1.94 5.36 0.95 1.75 3.99

.06 b.001 .03 NS NS

1.22

0.90

1.66

NS

2.50

1.30

4.80

.006

2.88

1.29

6.38

.009

Adjusted for age, NYHA class III or IV, ACE/ARB treatment, HR, MBP, and Rad dP/dt (as a continuous or a categorical variable), sex, tobacco use (past or current), diabetes, dyslipidemia, hypertension, coronary artery disease, pulse pressure, and left ventricular ejection fraction.

Kaplan-Meier curve showing the cumulative incidence of death or transplantation according to the cutoff value of Rad dP/dt b440 mm Hg/s.

Table IV. Determinants of the Rad dP/dt Crude analysis

Determinants

and P = .005 respectively). The estimated cumulative incidence of combined end-point was 44.6% in the low Rad dP/dt group versus 20.4% (P = .002) in the high Rad dP/dt group. Only 6 patients were recommended for heart transplantation; their Rad dP/dt values ranged between 167 and 392 mm Hg/s. The crude OR (95% CI) for the relation between Rad dP/dt b440 mm Hg/s and death or transplantation was 2.04 (1.12-3.72), P = .02, for patients with LVEF b40%, and 2.99 (1.09-8.17), P = .03, for patients with LVEF N 40%. The Rad dP/dt was low (b440 mm Hg/s) in 40% of patients with an LVEF b40% (n = 184) or in 10% with an LVEF ≥ 40%, (n = 126, P b .001). The sensitivity, the specificity, the positive and the negative predictive values of a Rad dP/dt b440 mm Hg/s to predict death or transplantation were 44%, 76%, 33% and 84%, respectively, and the area under the ROC curve of 0.59 ± 0.10 (Figure 2) . Regarding age, NYHA class, HR, PP and MBP areas under ROC curves were: 0.57 ± 0.10, 0.69 ± 0.11, 0.54 ± 0.10, 0.58 ± 0.10, 0.57 ± 0.10 and 0.59 ± 0.11, respectively.

Determinants of the Rad dP/dt The determinants of a high Rad dP/dt (Table IV) were an older age, a female gender , a NYHA class I or II, a past history of hypertension or diabetes, the presence of a coronary artery disease, a high pulse or MBP, a preserved LVEF and a low HR. In a stepwise backward regression, mains determinants of the Rad dP/dt were hemodynamics parameters in addition with age. In this multi-

Coefficient ± 1 SD ⁎

Age 8.4 ± 0.9 Female sex 156.5 ± 32.7 NYHA class III −65.1 ± 29.7 or IV Hypertension 266.0 ± 28.5 Diabetes 154.0 ± 30.3 Coronary artery 82.7 ± 29.1 disease Pulse pressure 14.87 ± 0.2 MBP 11.2 ± 0.9 LVEF 12.2 ± 0.6 HR −6.2 ± 0.9

P

Multivaried analysis Coefficient ± 1 SD ⁎

b.0001 −0.7 ± 0.3 b.0001 – .03 – b.0001 b.0001 .004

– – –

b.0001 15.0 ± 0.3 b.0001 0.8 ± 0.3 b.0001 0.6 ± 0.3 b.0001 1.9 ± 0.3

P .03 – – – – – b.0001 .03 b.05 b.0001

⁎Coefficient ± 1 SD: coefficient of the robust regression ± 1 SD.

varied analysis, the relationship with Rad dP/dt was inverted and less significant for age and also inverted but highly significant for HR.

Discussion In this large study of 310 patients with heart failure, we have shown for the first time a relation between decreased Rad dP/dt and death or transplantation. This relation was independent of the usual clinical parameters, blood pressure (BP) components, and LVEF. Moreover, a Rad dP/dt of 440 mm Hg/s can be used to identify a subgroup of patients with high estimated risk of death or transplantation, 44.6% at the end of our follow-up study. The poor prognosis after discharge related well with results of other studies of unselected community-dwelling patients. 17,18,29,30

762 Tartiére et al

The estimation of the LV dP/dt has been extensively used in cardiology, especially in the study of heart failure. 2,7-9,11 Kolias et al 2 showed the value of a positive LV dP/dt determined from continuous-wave Doppler spectra of the mitral regurgitation jet as a prognostic marker in patients with severe heart failure and a mean LVEF b50% (23% ± 9%). Regarding heart failure with an LVEF within the reference range, most of the studies assessing long-axis myocardial function by use of Doppler imaging have shown results suggesting an overlooked impaired myocardial systolic function, 9,10 but this result remain debated, especially when using more global marker of LV systolic function. 11 Indeed, the association of a low Rad dP/dt (b440 mm Hg/s) and an LVEF N50% was rare in our study (6% (4/64) of patients) but was more frequent (18% (9/50) of patients) for patients with a lower range of LVEF (40%-50%). However, although a low Rad dP/dt was less frequently implied in the occurrence of combined end-point with an LVEF ≥40%, the relation with prognosis was, however, persistent in both groups. This result suggests pathologic phenomenon other than loss of myocardial contractility leading to worsening prognosis being logically unpredictable by the Rad dP/dt assessment. Moreover, because causes of death in patients with heart failure are varied, we note a poor sensitivity and positive predictive value of the Rad dP/dt compared to less restrictive markers of myocardial dysfunction such as B-type natriuretic peptide assay or Doppler mitral flow pattern. 29,30 Left ventricular ejection fraction is a well-recognized prognostic factor in patients with heart failure and LV dysfunction.6 However, epidemiological studies have shown the great underuse (3%-88%) of echocardiography in recognized or suspected heart failure. 15-17,19,32 This situation is probably explained in part by differences between general practitioners and cardiologists in managing heart failure, as well as the low availability of this technique in some geographic sites or its relatively high cost and the lack of trained operators. Moreover, this results in an underuse of validated medical therapy and poorer prognosis. 19 In this situation, noninvasive assessment of the Rad dP/dt may be a relevant tool for the assessment of LV systolic function and may offer a clear predictive value, regardless of the availability of echocardiography, especially for large epidemiological studies. Another critical point regarding this new parameter is its exceptional feasibility and reproducibility, 26 and the rapid learning curve, which offers a potentially large diffusion of the method. This method could be complementary to usual echocardiography and probably automatically assessed during BP measurement, by using devices with high frequency response flat. Moreover, contrary to the dP/dt assessed by Doppler with mitral regurgitation jet, the Rad dP/dt does not need the presence of mitral regurgitation to be measured. Finally, brachial PP was used for calibration of the pulse wave and

American Heart Journal April 2008

was naturally a main determinant of the Rad dP/dt. However, in this study, the predictive value of the Rad dP/ dt was independent of the brachial PP, and PP was not related to prognosis, whatever the statistical model. This study has several important limitations. First, the Rad dP/dt by using the SphygmoCor PWA System (AtCor Medical) has never been compared with the invasive LV dP/dt. However, a recent study has pointed out the absence of beat-to-beat relationship between the Rad dP/ dt and the LV dP/dt by using another applanation tonometer device. 27 In this study, the entire recording of the pulse wave shape was calibrated by one value of mean and diastolic BP for a beat-to-beat relation. To compare both dP/dt measures (radial and LV), the radial pulse should be calibrated beat to beat. In this study, the poor relation shown between methods might reflect their poor reproducibility. 27 Previous studies have used the wave intensity index as a hemodynamic index, defined as (dP/dt) × (dU/dt) at any site of the circulation, where dP/dt and dU/dt are the derivatives of BP and velocity, respectively, with respect to time. 20-22 The magnitude of the first peak of this wave intensity index could be derived as (dP/dt) 2/ρc, where ρ is blood density and c pulse wave velocity. This first peak, by use of the carotid dP/dt, was closely related to the actual LV dP/dt (r = 0.74, P b .001) in patients with suspected coronary disease. 20 Second, the pulse wave morphology is continuously transformed along the arterial tree by the local viscoelastic properties and by the addition of the reflected wave, which leads to a steeper slope of the maximal first derivative.25 This phenomenon leads to a systematic increment of the arterial dP/dt when the distance from the aortic root increases.25 Therefore, if the value of the first derivative of the arterial pulse is clearly amplified during travel in the arterial tree, its ability to characterize the LV contractility is not necessarily affected. However, this systematic error, more or less its SD, associated with some misclassification, should only be in favor of the null hypothesis. Finally, in echocardiographic or Doppler parameters of LV function such as the mitral regurgitation dP/dt, the peak systolic velocity of the mitral annulus was not systematically assessed, which does not allow for a direct comparison between these parameters and the Rad dP/ dt. However, all these parameters have never been validated in a diverse population, and their reproducibility and feasibility are less attractive. Therefore, further studies are needed for direct comparison between the Rad dP/dt and echocardiography, Doppler parameters, or natriuretic peptide or more accurate parameters of exercise limitation than NYHA classification. Another remarkable point of this study is the high proportion of well-treated patients, in terms of the extensive use of ACE/ARB, β-blockers, and spironolactone according to age and LVEF. Treatment with ACE/ARB was the only therapeutic regimen

American Heart Journal Volume 155, Number 4

associated with outcome before or after adjustment for confounding factors. However, because dosage, time of onset of disease, indication, and eventual discontinuation of these medications were not known, caution must be taken in the analysis of drug effects.

Conclusions This study demonstrates, for the first time, that the maximal first derivative of the radial pulse, Rad dP/dt, is a useful tool for assessing the prognosis in heart failure patients. This prognostic value was significant regardless of baseline characteristics and, especially, LVEF. This simple, noninvasive measurement could play a role in the management of patients with heart failure. Further studies should be performed to compare its prognostic value with that of usual echocardiography and Doppler parameters or with biomarkers.

References 1. Lewis T. The factors influencing the prominence of the dicrotic wave. J Physiol 1906;34:414-29. 2. Kolias TJ, Aaronson KD, Armstrong WF. Doppler-derived dP/dt and −dP/dt predict survival in congestive heart failure. J Am Coll Cardiol 2000;36:1594-9. 3. Harjai KJ, Scott L, Vivekananthan K, et al. The Tei index: a new prognostic index for patients with symptomatic heart failure. J Am Soc Echocardiogr 2002;15:864-8. 4. Wang M, Yip GW, Wang AY, et al. Tissue Doppler imaging provides incremental prognostic value in patients with systemic hypertension and left ventricular hypertrophy. J Hypertens 2005;23:183-91. 5. Dujardin KS, Tei C, Yeo TC, et al. Prognostic value of a Doppler index combining systolic and diastolic performance in idiopathic-dilated cardiomyopathy. Am J Cardiol 1998;82:1071-6. 6. Curtis JP, Sokol SI, Wang Y, et al. The association of left ventricular ejection fraction, mortality, and cause of death in stable outpatients with heart failure. J Am Coll Cardiol 2003;42:736-42. 7. Garcia-Lledo A, Moya JL, Balaguer J, et al. Sensitivity of the Doppler rate of pressure rise to changes in the inotropic state: an experimental comparison with invasively obtained dP/dt. Eur J Echocardiogr 2000;1:271-6. 8. Bauer F, Jones M, Shiota T, et al. Left ventricular outflow tract mean systolic acceleration as a surrogate for the slope of the left ventricular end-systolic pressure-volume relationship. J Am Coll Cardiol 2002;40:1320-7. 9. Bruch C, Gradaus R, Gunia S, et al. Doppler tissue analysis of mitral annular velocities: evidence for systolic abnormalities in patients with diastolic heart failure. J Am Soc Echocardiogr 2003;16:1031-6. 10. Yu CM, Lin H, Yang H, et al. Progression of systolic abnormalities in patients with “isolated” diastolic heart failure and diastolic dysfunction. Circulation 2002;105:1195-201. 11. Baicu CF, Zile MR, Aurigemma GP, et al. Left ventricular systolic performance, function, and contractility in patients with diastolic heart failure. Circulation 2005;111:2306-12. 12. Kawaguchi M, Hay I, Fetics B, et al. Combined ventricular systolic and arterial stiffening in patients with heart failure and preserved ejection fraction. Implications for systolic and diastolic reserve limitations. Circulation 2003;107:714-20. 13. Weissler AM, Harris WS, Schoenfeld CD. Systolic time intervals in heart failure in man. Circulation 1968;37:149-59.

Tartiére et al 763

14. Weber T, Auer J, O'Rourke MF, et al. Prolonged mechanical systole and increased arterial wave reflections in diastolic dysfunction. Heart 2006;92:1616-22. 15. Majeed A, Williams J, de Lusignan S, et al. Management of heart failure in primary care after implementation of the National Service Framework for Coronary Heart Disease: a cross-sectional study. Public Health 2005;119:105-11. 16. Houpe D, Peltier M, Cohen-Solal A, et al. Heart failure due to left ventricular systolic dysfunction: treatment at discharge from hospital and at one year. Int J Cardiol 2005;103:286-92. 17. Senni M, Tribouilloy CM, Rodeheffer RJ, et al. Congestive heart failure in the community: trends in incidence and survival in a 10-year period. Arch Intern Med 1999;159:29-34. 18. Senni M, Tribouilloy CM, Rodeheffer RJ, et al. Congestive heart failure in the community: a study of all incident cases in Olmsted County, Minnesota, in 1991. Circulation 1998;98:2282-9. 19. Senni M, Rodeheffer RJ, Tribouilloy CM, et al. Use of echocardiography in the management of congestive heart failure in the community. J Am Coll Cardiol 1999;33:164-70. 20. Ohte N, Narita H, Sugarawa M, et al. Clinical usefulness of carotid arterial wave intensity in assessing left ventricular systolic and early diastolic performance. Heart Vessels 2003;18:107-11. 21. Niki K, Sugarawa M, Chang D, et al. A new non-invasive measurement system for wave intensity: evaluation of carotid arterial wave intensity and reproducibility. Heart Vessels 2002;17:12-21. 22. Niki K, Sugawara M, Uchida K, et al. A noninvasive method of measuring wave intensity, a new hemodynamic index: application to the carotid artery in patients with mitral regurgitation before and after surgery. Heart Vessels 1999;14:263-71. 23. Germano G, Angotti S, Muscolo M, et al. The (dP/dt)max derived from arterial pulse waveforms during 24 h blood pressure oscillometric recording. Blood Press Monit 1998;3:213-6. 24. Dahlquist G, Bjorck A, Anderson N. Numerical methods. Englewood Cliffs, N.J.: Prentice -Hall; 1974 25. Chiu YC, Arand PW, Shroff SG, et al. Determination of pulse wave velocities with computerized algorithms. Am Heart J 1991;121:1460-70. 26. Tartière JM, Logeart D, Beauvais F, et al. Non invasive radial pulse wave assessment for the evaluation of the left ventricular systolic performance in heart failure. Eur J Heart Fail 2007;9:477-83. 27. Sharman JE, Qasem AM, Hanekom L, et al. Radial pressure waveform dP/dt max is a poor indicator of left ventricular systolic function. Eur J Clin Invest 2007;37:276-81. 28. Swedberg K, Cleland J, Dargie H, et al. Guidelines for the diagnosis and treatment of chronic heart failure: executive summary (update 2005): The Task Force for the Diagnosis and Treatment of Chronic Heart Failure of the European Society of Cardiology. Eur Heart J 2005;26:1115-40. 29. Logeart D, Saudubray C, Beyne P, et al. Comparative value of Doppler echocardiography and B-type natriuretic peptide assay in the etiologic diagnosis of acute dyspnea. J Am Coll Cardiol 2002;40:1794-800. 30. Logeart D, Thabut G, Jourdain P, et al. Predischarge B-type natriuretic peptide assay for identifying patients at high risk of re-admission after decompensated heart failure. J Am Coll Cardiol 2004;43:635-41. 31. Tartière JM, Logeart D, Safar ME, et al. Interaction between pulse wave velocity, augmentation index, pulse pressure and left ventricular function in chronic heart failure. J Hum Hypertens 2006;20:213-9. 32. Rutten FH, Grobbee DE, Hoes AW. Differences between general practitioners and cardiologists in diagnosis and management of heart failure: a survey in every-day practice. Eur J Heart Fail 2003;5:337-44.