Why the study?
Do acute increases in systemic arterial pressure influence left ventricular diastolic function and pressures in patients with HFPEF?
Do acute increases in systemic arterial pressure influence left ventricular diastolic function and pressures in patients with HFPEF?
Answers are grounded in this paper and cite their sources.
This editorial highlights that acute increases in blood pressure in HFpEF patients primarily affect end-diastolic stiffness, emphasizing the importance of ventricular-arterial coupling and late systolic loading in HFpEF pathophysiology.
This editorial refers to ‘Impact of acute hypertension transients on diastolic function in patients with heart failure with preserved ejection fraction’ by C. Pérez del Villar et al., pp. 906–914 The syndrome of heart failure with preserved ejection fraction (HFPEF) continues to arouse controversy. Any research that can shed light on mechanisms and thereby offer specific targets for treatment should be welcomed. Patients may develop pulmonary oedema that can be triggered by acute increases in conduit arterial stiffness1 and exacerbated by simultaneous volume-loading.2 In this issue of Cardiovascular Research, Pérez del Villar and colleagues report an invasive clinical investigation in which they assessed how acute increases in systemic arterial pressure influence left ventricular (LV) diastolic function and pressures.3 They suggest that the main mechanism is a direct effect on end-diastolic stiffness rather than delayed relaxation. Before these results are generalized, it is important to consider some details of their study. The work is based on an advanced mathematical analysis of previously acquired and unique data.4 They included 39 patients with HFPEF: 54% were men, 59% had hypertension, and 10% had diabetes; their mean age was 54 years, their mean body mass index 27, and all were in sinus rhythm. Acute increases in blood pressure were induced by handgrip, which is a complex intervention that differs in important respects from dynamic exercise. Effects of tachycardia were assessed by atrial pacing. LV pressure-volume curves were constructed from data obtained with conductance catheters and high-fidelity transducers during acute reductions in preload induced by inflating a balloon in the inferior caval vein. Classically, relaxation and diastolic stiffness have been assessed separately, during isovolumic relaxation and at end-diastole, respectively. The new computational method (‘global optimization fitting’5) attempts to overcome that limitation but the model does not include determinants of diastolic physiology such as atrial function, coronary perfusion, ventricular interdependence, and synchrony. A study in 1980 suggested that increased LV diastolic pressures during handgrip were due to ventricular interaction, since LV end-diastolic stiffness was unchanged.6 It would have been valuable to compare the new estimates of the time constant of early relaxation (tau), and chamber stiffness constants, with those derived by classical methods, since an earlier study in some of the same patients found that these were unchanged after handgrip or pacing4 and their own previous validation study did not include any subjects with HFPEF.5 Also, the Matsubara logistic regression method is poorly sensitive for evaluating effects of systolic pressure-dependent changes in myocardial relaxation on the end-diastolic pressure-volume relationship.7 Their methodology is provocative and innovating but it needs more validation. Pérez del Villar and colleagues estimated the contributions to LV pressure at each time-point during diastole that came separately from elastic recoil, early diastolic relaxation, and diastolic stiffness. They demonstrated that recoil and tau were similar at rest in HFPEF patients to younger control subjects – so stress interventions are necessary for diagnosis. They found that the impact of arterial pressure on diastolic function was reduced when LV volumes were reduced during atrial pacing – so we should avoid fluid retention in patients with HFPEF. Since patients were studied while they remained on their usual medical treatment, some may have been relatively hypovolaemic. The arterial pressure-sensitivity of diastolic stiffness correlated well with collagen volume fraction and cross-linking, which is interesting, although biopsies were taken from the right side of the ventricular septum. A surprising finding was that there were increased restoring forces in HFPEF patients, during acute increases in blood pressure − since a previous study found that the ability to enhance restoring forces or LV suction was impaired in HFPEF because systolic function was reduced during stress.8 The estimated parameters of diastolic function correlated more strongly with peak systolic pressure than with arterial elastance or end-systolic wall stress, although those are better indexes of afterload.9 Handgrip increased arterial elastance (Ea) but did not alter systemic vascular resistance (SVR); it has been reported that SVR and heart rate together account for 98% of the variability of Ea.10 These observations raise questions about the validity of their model or the representativeness of the studied sample. The detailed pathophysiology of ventricular-arterial coupling has been explored in experimental and clinical studies in which arterial pressure was increased by banding or cross-clamping the aorta or by inflating a balloon within it. In open-chest dogs, increasing systolic load during early systole (by timed inflation of an intra-aortic balloon) prolonged the duration of contraction but did not alter the rate of fall of pressure in early diastole, whereas augmenting load in late systole induced an earlier onset of the early diastolic fall in pressure but at a slower rate which delayed filling.11,12 In a small animal model, more LV hypertrophy was provoked by banding the suprarenal aorta than the aortic arch, implying that late systolic loading is more important.13 In another study the only independent predictor of elevated end-diastolic pressure in response to acutely increased afterload was the reduced time available to relax.14 Diastolic filling time is also reduced in patients during acute exacerbations of HFPEF.1 Clinical investigations have confirmed these observations. The magnitude of wave reflection, importantly affecting late systolic loading, predicted incident heart failure in 6124 adults followed for 8.5 years, with a hazard ratio of 2.3; notably, the risk of heart failure was increased in normotensive patients with strong wave reflections more than in patients with systolic hypertension but normal wave reflections.15 Heart failure events are predicted by late systolic backwards-travelling wave reflections.16 Cardiovascular events in hypertensive patients were predicted by higher intensity of late systolic reflections, with a risk ratio of 2.2, but not by early systolic pressure or augmentation index.17 So instead of a simple model (arterial hypertension causes LV hypertrophy which leads to myocardial fibrosis and then heart failure), we can now be more specific, although we still do not know the precise haemodynamic trigger for hypertrophy (Figure 1). The profile of arterial pressure in mid and late systole (Figure 2A) is more important than peak systolic pressure in determining the rate at which pressure falls during isovolumic relaxation (Figure 2B). Pérez del Villar and colleagues report ‘systolic pressure’ as a single variable (‘LVP max’), while the timing of augmented systolic loading is crucial. There is evidence that increased wall stress during early (but not late) systolic ejection is associated with higher early diastolic velocities of mitral annular motion, and with enhanced LV suction.18 To interpret the significance of their findings, therefore, we need to know how handgrip alters the shape of the arterial pressure waveform and specifically if it affects late systolic loading. It may be premature to extrapolate from their data that there is a direct causative link rather than an association between increased peak systolic pressure and LV diastolic stiffness, implying a dynamic effect on the properties of collagen. Haemodynamic mechanisms that lead to increased LV filling pressures in HFPEF. Pathways for which there is established evidence are shown with red arrows. Those for which precise evidence is lacking are shown with dotted black arrows. The association proposed by Pérez del Villar and colleagues is shown with the orange arrow; it remains speculative. Carotid arterial and left ventricular pressure waveforms. (A) These systemic arterial waveforms were recorded by applanation of the common carotid artery, in a healthy subject with a normal pressure and waveform, and in 2 subjects with a similarly increased peak systolic pressure − one of whom has no evidence of systolic augmentation or of increased late systolic pressures, while the other has an increased area under the late systolic pressure curve related to increased wave reflections; it is this subtle latter change that has an adverse impact on LV myocardial relaxation and early filling. The dotted black line marks the end of systole (coinciding with the dicrotic notch). Note that the subject with increased reflections also has a higher pulse pressure. BP = blood pressure. Traces by courtesy of Dr Julio A Chirinos, University of Pennsylvania. (B) Left ventricular (LV) pressure traces drawn (not recorded) to illustrate how arterial waveforms can affect LV diastolic filling. The patterns have been exaggerated for illustrative purposes; note that other patterns of ventricular-arterial coupling are posssible, and that the examples illustrate extremes. Compared with a normal subject (green trace), a subject with increased sympathetic stress or who is exercising (blue trace) may demonstrate increased early systolic pressure and then increased elastic recoil, a more rapid early diastolic fall in LV pressure, and increased suction. A subject with increased early loading would have prolonged ejection with delayed onset and faster rate of LV pressure fall in early diastole (not shown). In a patient with increased late systolic pressure (red trace), the onset of relaxation may start earlier than normal but the rate of fall of LV pressure is slowed, suction is reduced, diastolic pressures are elevated, and atrial filling is increased. Alternative explanations are possible for correlations between arterial pressure, intrinsic LV diastolic chamber stiffness, and acute elevations of end-diastolic pressure during transient increases in blood pressure. Coronary suction and microvascular perfusion are reduced in pathological LV hypertrophy19; ischaemia induced by mismatched oxygen demand and supply during stress could increase stiffness. Isoform shifts and post-translational changes in titin can lead to loss of LV capacitance.20 Causes of arterial stiffening such as oxidant stress and advanced glycation end-products,21 together with endothelial damage,22 may promote myocardial fibrosis. LV hypertrophy can be caused by direct effects of angiotensin II from the kidney. And of course the relative contributions of different mechanisms may vary at different stages in the natural history of the disease, and in different patients. The global optimization methods reported by Pérez del Villar and colleagues could now be applied to chronic animal models of HFPEF and ideally to larger cohorts of more typical HFPEF patients in whom pressure-volume loops were recorded. Alternatively, we need a validated non-invasive method for diagnosing LV diastolic responses to dynamic exercise in patients with suspected HFPEF. New echocardiographic methods for assessing elastic recoil and early diastolic untwisting using speckle tracking, and to measure diastolic stiffness using ultrafast plane-wave imaging, look promising. Whichever diagnostic target and methods become most useful for clinical practice, however, will depend above all on their details – until we understand them better it will remain difficult to identify new specific targets for treatment. Conflict of interest: none declared.
No takes yet. Share an insight, caveat, or question.
Fraser et al. (2017) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: