Epidemiologic studies have consistently shown that a meaningful percentage of patients with heart failure (HF) has a preserved left ventricular (LV) systolic function or, more properly, a normal LV ejection fraction (EF). This symposium has thoroughly assessed this problem with respect to its epidemiologic, pathophysiologic, clinical, and diagnostic aspects.
Epidemiology
Dr. Michele Senni, (Bergamo, Italy) has nicely described the main epidemiologic features of HF with preserved systolic function. There is a wide variation in the prevalence of this condition. However, on average, these patients encompass 50-55% of all the patients hospitalised for HF. Compared to the patients with HF and a low LVEF, they are more likely to be elderly, to be females, to have a history of hypertension and less likely to have had a previous myocardial infarction. With regards to prognosis, they have a similar rehospitalisation rate and a similar or, more often, slightly lower mortality.
Definition and mechanisms
Prof. Gilles W. De Keulenaer (University of Antwerp, Belgium) has first discussed the different definition proposed to characterise these patients. They include the definition of “diastolic HF” which may then be considered as definite, probable or possible depending on whether an abnormality of LV diastolic function is shown or is only hypothesised based on the finding of a normal LVEF. Based on the difficulties in showing an abnormality of LV diastolic function, European guidelines have now preferred the more conservative definition of HF with preserved (or normal) LVEF.
As elegantly pointed out by Gilles W. De Keulenaer, systolic and diastolic function generally coexist and there is a continuum from normal LV function to LV with preserved LVEF, though with other abnormalities of systolic and diastolic function, to HF with low LVEF. The main pathophysiological mechanisms of HF with normal LVEF are three: inappropriate tachycardia, slow systolic relaxation, decreased LV compliance. Slow systolic relaxation may be caused by excessive pre-load or afterload, impaired inactivation at the cardiac myocyte level, non-uniformity of load and inactivation. The episodes of HF with normal LVEF occurring during a hypertensive crisis are typical examples. The excessive increase in afterload causes impaired relaxation and increased LV and aortic stiffness generally contribute.
Cinical profiles
Prof. Michel Komajda (Paris, France) has nicely showed the clinical characteristics of HF with normal LVEF, as shown by surveys and clinical trials. These patients are older, more likely to be females, to have a history of arterial hypertension and/or to suffer of atrial fibrillation. Patients with HF and normal LVEF are less likely to be treated with neurohormonal antagonists, above all ACE inhibitors, and with digitalis, and more likely to be treated with calcium antagonists. The rationale for the undertreatment of these patients resides in the lower evidence of benefits as less randomised trials have been performed in these patients. These data have been consistently shown by many surveys including the EuroHeart Failure, ADHERE, an Italian survey, and one survey performed in two centers in Switzerland and Finland.
Patients enrolled in trials are generally different from those of the clinical practice. Accordingly, some differences can be found between the patients included in CHARM-Preserved compared to those of surveys. Patients of the ongoing I-PRESERVE trial seem less different from those of surveys.
Diagnostic problems
Despite the ability of Prof. Alan Fraser (Cardiff, UK) in describing the signs of LV diastolic dysfunction at Doppler-echocardiography, the task of accurately diagnosing this condition remains difficult. First, estimation of LVEF with the “eyeball method” may cause mistakes with major inaccuracies above all when we are assessing patients with a LVEF around 40-50%. LVEF should therefore be measured by biplane echocardiography. LVEF alone is an inadequate measurement of LV function and should be always attended by a measurement of LV volumes, namely end-systolic volume.
Patients with HF and normal LV EF generally have other abnormalities of LV systolic function like an impairment of LV long-axis shortening.
The ratio of early to late transmitral flow velocities (E/A ratio) has a U-shaped relationship with LV diastolic function with its highest values in a normal left ventricle and with severe LV diastolic dysfunction. The reverse is true with respect to the early diastolic transmitral flow velocity and isovolumic relaxation time as they have an inverted U-shape relationship with LV diastolic function with their lowest values in normal conditions and with advanced LV diastolic dysfunction. These measurements are therefore largely sub-optimal for the assessment of LV diastolic function. Their diagnostic accuracy may be increased by the use of the Valsalva manoeuvre.
Newer indices allow a better assessment of LV diastolic function. Perhaps the most useful is the ratio of the mitral E velocity to the velocity of mitral annular motion during early diastole recorded by tissue Doppler echocardiography (E/Ea). This parameter is highly related to LV filling pressure. Another important parameter is the difference in duration during atrial contraction of antegrade flow through the mitral valve compared with retrograde flow into the pulmonary veins. Left atrial volume also allows an accurate estimate of LV filling pressure.