The circadian pattern of blood pressure is mainly affected by autonomic nerve activity. Changes that occur throughout the day affect the cardiovascular system, leading to various alterations in cardiac function including diastolic function based on the time and circumstances. Doppler-echocardiography is the most useful tool for routinely measuring diastolic function noninvasively. According to the recommendations for left ventricular (LV) diastolic evaluation using echocardiography, the severity of diastolic dysfunction is assessed using a combination of several indices for the left and right sides of the heart, including E/A, deceleration time, E/e´, tricuspid regurgitation velocity, ventricular and atrial strain pattern, and left atrial (LA) volume.1–3 Evaluation of disease severity is useful for prognostic evaluation in patients with heart failure (HF) with preserved ejection fraction (HFpEF).4 However, there is an important crosstalk between arterial load and diastolic function5: diastolic relaxation is reduced with acute increases in the arterial load.6–8 Unfortunately, none of the noninvasive diastolic indices consider systemic vascular resistance.
Validity Of A Vascular Resistance-integrated Diastolic Index
HFpEF is historically thought to be driven by increased afterload in arterial stiffness and hypertension, with consequential LV diastolic dysfunction as one of the hallmarks of the syndromes. Early studies suggested that in the clinical settling, E/e’ is a marker of reduced LV long-axis early diastolic expansion, that is, one of the markers of LV diastolic function, resulting in showing estimated LA filling pressure.9, 10 The correlation between E/e´ and direct LA pressure or pulmonary capillary wedge pressure is significant in a stable state.11, 12 Among several indices evaluated using Doppler echocardiography, E/e´-related indices, such as (E/e´)/stroke volume (SV), i.e., operant diastolic elastance (Ed), reportedly reflect LV diastolic function.13, 14 The effective arterial elastance (Ea) is calculated as (0.9 × systolic blood pressure)/SV.13 We previously reported on age- and sex-related differences in LV diastolic function relative to arterial elasticity among hypertensive patients with preserved LV ejection fraction (LVEF) and no history of HF.15, 16 We found that the vascular resistance-integrated diastolic index Ed/Ea [calculated as (E/e´)/(0.9 × systolic blood pressure)] was significantly increased in older (aged ≥ 75 years) hypertensive women and was coincident with cardiac structural alterations. Under stable conditions, Ed and Ea are higher in women than in men.17, 18 Since LV diastolic function is strongly affected by arterial load, the elevated Ed in women could be an epiphenomenon because of an associated increase in Ea. In this sense, Ed must be corrected according to Ea, considered together as Ed/Ea, when analyzing sex differences in diastolic function. E values are significantly higher and mean e’ values are significantly lower in HFpEF patients showing high Ed/Ea levels than in those showing low Ed/Ea levels.19 The blood pressure can be significantly influenced by antihypertensive treatment, but the E/e´ ratio would change according to the blood pressure changes, resulting in a subtle change in the Ed/Ea. Although systolic pressure might decrease as SV decreases, there are no significant differences in LV volume and SV between the patients with high and low Ed/Ea under the preserved LVEF before discharge.19 A high ratio of Ed/Ea is not just a sign of reductions in LV volume and filling.
Increased LV filling pressure owing to exercise correlates with changes in the diastolic relaxation rate and arterial afterload.20 The linear slope of the single-beat diastolic pressure-volume relationship is defined as Ed.21, 22 Exercise induces an increase in Ed evaluated invasively20 and noninvasively ([E/e´]/SV).23 In individual patients, the noninvasive index E/e´ cannot be used to reliably track changes in the left-side filling pressure induced by volume change24 or exercise12; however, in previous studies, the changes in the arterial load were not considered during assessment. Arterial load could be assessed using effective arterial elastance (Ea = end-systolic pressure/SV).13, 25 It is essential to measure systolic blood pressure around the examination of echocardiography to measure this index. Advanced age and female sex are associated with increases in arterial and ventricular stiffness even in the absence of cardiovascular disease.13 Exercise also increases Ea; however, according to the study by Borlaug et al.,20 Ed/Ea does not seem to change significantly after stress if it were not for stress-induced ischemia. Changes in Ea, in addition to those in diastolic elastance, are compromised in HFpEF26; these changes are outside of those associated with aging or hypertension.27
We recently reported that the LA volume index and Ed/Ea are high in patients with HFpEF.28 Ed/Ea reflects LA pressure relative to the systemic pressure29 that means the ratio of LV end-diastolic pressure to LV end-systolic pressure, which can change minimally over a day under various circumstances. Thus, the Ed/Ea ratio may reflect the left-sided heart function status, including atrioventricular-arterial interaction, under preserved LVEF conditions.
Utility Of The Vascular Resistance-integrated Diastolic Index For Prognosis In Hfpef
The pathophysiology of HFpEF is complex and includes alterations in cardiac structure and function, systemic and pulmonary vascular abnormalities, and comorbidities.30 HFpEF is detectable in a relevant portion of antiphospholipid syndrome.31 Hospitalizations related to HFpEF are increasing and the growing older population has led to this trend worsening. We recently reported prognostic data in older patients (mean age, 81 years) recruited from the Prospective Multicenter Observational Study of Patients with Heart Failure with Preserved Ejection Fraction registry32 and found that Ed/Ea is a useful prognostic marker in patients with HFpEF.33, 34 Our findings can help determine which single index of LV diastolic function is significantly associated with the prognosis. In particular, in patients with a higher level of N-terminal pro-brain natriuretic peptide (NT-proBNP), a higher Ed/Ea was associated with a poor prognosis.34 In a recent prospective study among patients with established HFpEF, Ea35 and Ed34 failed to predict adverse outcome. As the other indices for arterial stiffness, carotid-femoral pulse wave velocity (cf-PWV) and carotid augmentation index (cAIx) are well established.36–38 However, Huang et al. reported that cf-PWV, but not cAIx, was related to clinical outcomes in patients with HFpEF.38 The patient characteristics were quite different between Huang’s study and our study; the incidences of male gender and coronary artery disease were high and those of hypertension and atrial fibrillation (AF) were low in their study.38, 39 Furthermore, the noticeable aspect is that important clinical variables were not included in their analysis of a multivariable model.38 Although arterial stiffness itself may be of added prognostic value in studies with a longer follow-up,38 cf-PWV was not reported as a predictor for all-cause mortality in patients with HFpEF.36–38 Our proposed index may be easily applicable in hospitals and clinics because the precise measurement of volumetry or strain pattern for cardiac chambers by echocardiography is not needed.
Factors Affecting The Significance Of Ed/ea In The Prognosis Of Patients With Hfpef
Prognostic factors related to LA overload may differ according to the clinical endpoint in older patients with HFpEF.39 Ed/Ea and SV/LA volume (LAV), a relative index for LA volume overload,29 were significant prognostic factors for re-admission in HF.38 However, Ed/Ea, but not SV/LAV, was a significant prognostic factor for all-cause mortality. E/e’ is reported to be a validate predictor for HF re-admission, but not all-cause mortality, in a univariable model after the first acute HFpEF event.40, 41 When E/e’ was used in place of Ed/Ea in our study, E/e’ was not a prognostic factor for re-admission for HF in a multivariable Cox hazard analysis.39 The cutoff point of Ed/Ea for all-cause mortality (nearly 0.130) or re-admission for HF (nearly 0.100), observed in a receiver operating characteristic curve analysis during a short-term period in patients with HFpEF,39 was in accordance with that in hypertensive patients with preserved LVEF without HF15 (mean ± SD value of Ed/Ea, 0.100 ± 0.030, mean age 80 years, unpublished observation). In older patients with HFpEF, the most important events are re-admission for HF in each subject and in socioeconomical point. Recently, average and healthy life expectancy increases further in all-over the world and the evaluation of prognosis for mortality is important more than before within a super-aging society with a declining population like in Japan.
The prognostic significance of Ed/Ea may be valid only over a short-term period in older patients. Ed/Ea assessed before discharge was a significant prognostic factor for all-cause mortality during the first, but not the second, year after discharge.33 In landmark Kaplan–Meier survival (Fig. 1) and multivariable Cox hazard analyses performed using the value of Ed/Ea at 1 year after discharge, Ed/Ea remained a significant prognostic factor during the second year after adjusting for age and sex.19 To strictly evaluate the prognostic risk for all-cause mortality, serial examination for Ed/Ea would be optimal in clinical setting. No differences in LVEF or LV mass index (LVMI) were observed between patients with and without events during the first year among those with a high Ed/Ea before discharge; however, a reduced LVEF and larger LVMI were observed in patients with events during the second year among those with high Ed/Ea at 1 year after discharge.19 The cause of death or the pathophysiology of HFpEF may differ in the first and second years after discharge. The clinical significance of prognostic factors related to hemodynamics in patients with HFpEF may differ based on the follow-up period. The serial measurement of Ed/Ea is needed to assess the occasional prognosis accurately. Under these conditions, one can change the medications as needed while observing the pathophysiology. In this sense, the role of LVEF in prognosis may be the same as that of Ed/Ea.42, 43 A previous study reported that the most recent Kansas City Cardiomyopathy Questionnaire score is most strongly associated with subsequent prognosis in serial health status evaluations of patients with HfpEF44; this corresponds with our findings.
Arterial stiffness may preferentially contribute to abnormal diastolic function during exercise in women with HFpEF compared with men.45 Originally, Ed/Ea was used as an index for the sex-related differences in cardiac function in older hypertensive patients with preserved LVEF15; however, the sex differences in Ed/Ea as a prognostic indicator remained undefined. We recently reported that between-sex difference in the significance of Ed/Ea as a prognostic factor was observed in patients with HFpEF (Table 1).46 Ed/Ea was significant for re-admission for HF during the first year in men, but not in women, in a multivariable analysis.46 P for interaction regarding Ed/Ea for this prognosis between sexes was significant. In contrast, Ed/Ea was significant for all-cause mortality during the first year after discharge in women, but not in men. Although Ed/Ea was a significant prognostic factor for all-cause mortality in HFpEF patients including both sexes39 or each sex (unpublished observation) during the three years after discharge, the most of the differences in the prognosis for all-cause mortality between the patients with high and low Ed/Ea were observed during the first year.33 During the short-term after enrollment, the role of hemodynamic index such as Ed/Ea in a prognostic factor may be different between sexes.
The significance of Ed/Ea as a prognostic risk factor may be reduced in HFpEF patients with AF, because the calculation of Ed/Ea is based on the correct measurement of E/e´, which should be carefully and precisely measured. The incidence of AF did not differ significantly in our patients (38%)46 as compared to the previous report (40%).47 E/e’ is correlated with invasive LV filling pressure and adequate reproducibility even in patients with AF.48 The R-R interval is irregular in AF, and we measured the mean value of E/e´ among several beats in AF patients with an unstable blood pressure. However, changes in E/e´ may parallel those in blood pressure, and the ratio of E/e´ to blood pressure, i.e., Ed/Ea, does not differ largely under stable conditions. In fact, Ed/Ea did not differ significantly between the HFpEF patients with and without AF before discharge.49 Although Ed/Ea may provide lesser important information for evaluating all-cause mortality in HFpEF patients with AF than with sinus rhythm,49 in a larger number of HFpEF patients irrespective with or without AF, Ed/Ea is a significant prognostic factor of all-cause mortality.39 Ed/Ea exhibits the relative and not the absolute value of LA filling pressure and could be representative of the general performance of the left-sided heart under preserved LVEF including the patients with AF. The measurement of systolic blood pressure and E/e’ by echocardiography may be a first line of clinical procedure in HFpEF patients with history of hospitalization irrespective of the incidence of AF in addition to the evaluation of laboratory data such as NT-proBNP to predict their prognoses.
Table 1
Multivariable analysis for re-admission for heart failure and all-cause mortality during the first year after discharge in patients with heart failure and a preserved ejection fraction
Re-admission for heart failure | Men | | Women |
---|
Ratio | 95% CI | P value | | Ratio | 95% CI | P value |
---|
Age | 1.272 | 0.692–2.338 | 0.440 | | 1.061 | 0.602–1.867 | 0.840 |
---|
LAVI | 1.251 | 0.699–2.236 | 0.450 | | 2.094 | 1.121–3.912 | 0.020 |
Ed/Ea | 2.108 | 1.161–3.825 | 0.014 | | 0.965 | 0.558–1.669 | 0.900 |
eGFR | 0.810 | 0.446–1.471 | 0.490 | | 0.694 | 0.376–1.281 | 0.240 |
NT-proBNP | 2.462 | 1.214–4.994 | 0.012 | | 1.643 | 0.900-2.999 | 0.110 |
All-cause mortality | Men | | Women |
Ratio | 95% CI | P value | | Ratio | 95% CI | P value |
Age | 3.656 | 1.632–8.189 | 0.001 | | 6.054 | 2.707–13.54 | < 0.001 |
LAVI | 0.851 | 0.391–1.851 | 0.683 | | 1.212 | 0.585–2.509 | 0.604 |
Ed/Ea | 1.598 | 0.711–3.588 | 0.256 | | 3.582 | 1.540–8.335 | 0.003 |
eGFR | 0.804 | 0.341–1.895 | 0.619 | | 0.957 | 0.451–2.032 | 0.909 |
NT-proBNP | 2.808 | 1.116–7.070 | 0.028 | | 2.878 | 1.275–6.495 | 0.010 |
CI, confidence interval; Ea, arterial elastance; Ed, diastolic elastance; eGFR, estimated glomerular filtration rate; LAV, left atrial volume; LAVI, left atrial volume index; NT-proBNP, N-terminal pro-brain natriuretic peptide; SV, stroke volume
These findings were obtained from the Prospective Multi-centre Observational Study of Patients with Heart Failure with Preserved Ejection Fraction (PURSUIT HFpEF) registry (UMIN ID: UMIN000021831).29, 32 We examined the differences in significant prognostic factors related to LA volume and pressure overload by sex in patients with HFpEF. Cut-off point of the prognostic factors was evaluated using receiver operating characteristic curve analysis. A multivariable analysis using categorical variables determined by each cut-off point (a dichotomous cut-off) was performed under adjusting with comorbidities, including atrial fibrillation, hypertension, diabetes mellitus, dyslipidemia, and history of coronary artery disease (men/women 406/492). The hazard ratio with 95% CI was estimated. The part of the data in this table was cited from Reference 46.