The natural history of DHF begins with normal myocardium, then hypertrophied myocardial cells producing abnormal relaxation, and then further progression resulting in interstitial fibrosis and increased collagen and finally development of mild compliance abnormalities. Calcineurin activation and angiotensin II contribute to the formation of myocardial hypertrophy and fibrosis.

Calcium plays a critical role in the natural history. When calcium interacts with troponin C, it results in a transfiguration of the tropomycin-troponin complex and opens binding sites. Calcium attaches to troponin C, binding sites are activated, contraction, relaxation, and then filling occurs. In the normal cardiomyocyte the calcium transient goes through the cycle and then disappears. Abnormal calcium homeostasis results in abnormal relaxation.

Diastolic function is a complex sequence of multiple interrelated events: relaxation, suction, erectile coronary effect, viscoelastic forces, pericardial restraint, ventricular interaction, atrial contraction, chamber stiffness and myocardial stress/strain relations. Diastolic dysfunction is an abnormality of one or more of these events. In simplified terms, the major determinants of left ventricular (LV) filling are elastic recoil (suction); myocardial relaxation (Tau); LV chamber compliance, left atrial pressure and heart rate.

Systemic hypertension is the major cause of DHF and is seen in 70% of patients with this dysfunction. Other conditions that contribute to DHF, alone or in combination with hypertension, are left ventricular hypertrophy, ischemic heart disease (IHD), diabetic disease, and valvular heart disease. Less common contributors are restrictive cardiomyopathy, constrictive pericarditis, hypertrophic cardiomyopathy, infiltrative disorders such as amyloid heart disease, storage disorders, and obstructive sleep apnea. Obesity, older age and female gender predispose to development of DHF.

DHF is clinically indistinguishable from SHF based on history, physical examination, ECG and chest x-ray. Even elevation of BNP, now often used for diagnosing CHF, does not distinguish between the two forms of dysfunction. At least one study has shown that the most marked elevation of BNP occurs in patients with both.

To diagnose DHF, abnormal relaxation, filling or stiffness must be demonstrated, in addition to clinical symptoms and the presence of normal elevated systolic function. According to the Framingham Heart Study, a definitive diagnosis requires the presence of heart failure with a normal ejection fraction noted within 72 hours of admission and, importantly, a cardiac catheterization to establish high filling pressures. Probable or possible DHF is diagnosed in the presence of these features and the absence of a cardiac catheterization. However, the requirement of a cardiac catheterization renders this definition impractical.

The ACC/AHA guidelines state that the diagnostic evaluation for heart failure should include parameters to determine the type and severity of cardiac dysfunction, determine prognosis, and guide treatment. The guidelines recommend Doppler to assess systolic, but more importantly diastolic dysfunction.

Pulse Doppler techniques provide the ability to study the LV during diastole in health and disease and hence provide a non-invasive method to study DHF. Parameters that can be evaluated with Doppler are listed in Figure 1. Mitral inflow velocities (Figure 2), pulmonary venous flow and Doppler tissue imaging are the three most important parameters. Three patterns of diastolic cardiac dysfunction have been established: normal, abnormal relaxation, and restrictive pattern.

Grade One DHF: mild diastolic dysfunction, abnormal relaxation, earliest abnormality. Grade Two: intermediate degree of diastolic dysfunction. Grade Three: advanced diastolic dysfunction with high filling pressures, restrictive pattern. Grade Four: advanced heart failure that remains constrictive despite treatment; poor prognosis. The filling patterns correlate with filling pressures. Grade One diastolic dysfunction nearly always has normal resting mean left atrial pressure. In contrast, Grade Three diastolic dysfunction pattern always implies markedly elevated mean left atrial pressure (Figure 3).

Doppler tissue imaging is a reliable, easy and reproducible methodology. Advancing degrees of diastolic dysfunction are associated with changes in the mitral E and A velocities. But, these are discordant changes. On tissue Doppler at the mitral level, the E prime velocity remains reduced with advancing diastolic dysfunction. As the disease progresses and filling pressures increase, the mitral E and the annulus E velocities increase, thus increasing their ratio. Thus, the ratio of mitral E to annulus E velocity is informative.

The E to E prime ratio and the mean left atrial pressure is increased in patients with elevated diastolic pressures, according to data from the Mayo Clinic, and in other heart diseases, except constrictive pericarditis. In this setting, as the ratio increases, the mean left atrial pressure is decreased, resulting in a paradox in which high pressures correlate with a low ratio since the E prime is well preserved.

A reduction in the systolic forward flow causes a uniform elevation in the mean left atrial pressure, that is, high diastolic forward flow. A shortened deceleration time for the diastolic pulmonary venous velocity is a good indicator of high left atrial pressure, or wedge pressure.

Substantial progress has been made in the estimation and more accurate prediction of pressures. A tight, linear correlation between the mitral deceleration time and pulmonary vein diastolic deceleration time has been shown in patients with myocardial infarction. Supporting new data shows that deceleration of the pulmonary venous diastolic flow is more accurate than the pulmonary artery occlusion pressure in predicting left atrial pressure, and is more accurate than catheter-derived data using the occlusion technique.

Work from Japan in the experimental dog model of infarction shows that as LV compliance is reduced, atrial contraction occurs as the LV pressure rapidly increases and exceeds the left atrial pressure, resulting in a very short duration of forward flow. Clinically, this method has been shown to accurately predict end diastolic pressure.

The severity of diastolic dysfunction and the level of the mean left pressure can be understood by evaluating: the elevation of the left atrial pressure, left ventricular end diastolic pressure, mitral natural history distribution of velocities, reduced forward flow and E prime velocity as measured by tissue Doppler, pulmonary vein measures, and diastolic dominance and slow flow, as noted in the color flow propagation, including diastolic deceleration time (Figure 4).

Left atrial volume mirrors diastolic dysfunction of the left ventricle. In hypertrophic cardiomyopathy the left atrium is markedly enlarged. Left atrium size is a quick indicator of the degree of diastolic dysfunction in the absence of intrinsic mitral valve disease.

Studies at the cellular level in a single myocyte with calcium overload showed no difference in the resting state between the wild-type and genetic knockout mice. Beta adrenergic stimulation with isoproterenol showed no diastolic overloading in the wild-type normal myocyte. In contrast, in the diseased myocyte, beta adrenergic stimulation caused progressively increased diastolic loading, that is, diastolic dysfunction is unmasked.

Relaxation abnormalities can now be investigated in the cardiac catheterization and echocardiographic laboratories, based on data obtained at the cellular level. At Mayo Clinic in the echocardiography laboratory, patients with baseline normal diastolic function and only mild abnormalities are subjected to a bicycle stress test. It is no longer sufficient to perform only a Doppler examination in the resting state. The next step of a stress test must be taken to non-invasively unmask abnormalities.

Outcomes can be stratified based on pulmonary artery wedge pressure, as noted by Warner Stevenson at Brigham & Women's Hospital in Boston. One-year survival was excellent if the initial wedge pressure was little elevated. Survival was intermediate if the initial elevated wedge pressure was reduced to less than 16 mm Hg with treatment. Survival was poor if the pressure remained elevated despite optimal medical treatment.

Since the Doppler provides insight into pulmonary artery capillary wedge pressure, or mean left atrial pressure, it is intuitive that the grade of diastolic dysfunction will also inform about prognosis. Studies at Mayo Clinic showed the worst survival in patients with cardiac amyloidosis and severe diastolic dysfunction. Dilated cardiomyopathy also has a poor prognosis. In patients with dilated cardiomyopathy with an identical ejection fraction, the survival was worse for those with deceleration, which was markedly shortened, that is, advanced diastolic dysfunction.

Simple mitral deceleration time can predict the outcome for patients with ischemic cardiomyopathy. A markedly shortened deceleration time indicates very few viable segments and a high level of scarring. Post-operatively, the least improvement in ejection fraction was predictive and seen in patients with markedly shortened deceleration time. The data is very similar to dobutamine echo prediction. Profound insight into the management and outcome of ischemic cardiomyopathy can be gained using diastolic parameters, such as deceleration time. In patients with acute myocardial infarction, restricted ventricular filling is prognostic. Outcomes worsen as the degree of diastolic dysfunction worsens.

A multicenter study of Doppler tissue imaging and color flow propagation velocities concluded that advanced diastolic dysfunction is associated with a worse prognosis. A Doppler tissue imaging ratio greater than 10 indicated a worse prognosis.

Italian investigators showed in patients with severe diastolic dysfunction with shortened deceleration time that outcomes were worse in the patients in whom the deceleration time remained shortened despite optimal medical treatment. Outcomes (survival, transplantation, repeat hospitalization) improved in patients whose deceleration time increased with treatment.

Unlike the situation for SHF, no evidence-based treatments are available for DHF, since there is no published clinical trial evidence for treatment. However, multiple trials are underway and results will become available to provide the needed data. Until then, the emphasis must be preventive measures, including optimal treatment of hypertension, elimination of ischemia via interventions, catheter-based or surgery, as with ischemic cardiomyopathy; regression of hypertrophy, and reduction of calcium overload and fibrosis.

The effective treatment of hypertension reduces CHF incidence by 50%. However, the control of hypertension is poor worldwide. In the United States, only 27% of patients with hypertension are controlled to levels less than 140/90mm Hg. In England, only 6% of patients are controlled. The AT1 blocker losartan reduces the development of heart failure, and improves the symptoms of patients with heart failure.

Doppler profiles are useful to guide treatment. In Grade One diastolic dysfunction, diuresis need not be increased. The emphasis is to maintain synchrony, reduce the heart rate to allow more time for diastolic filling and maximize neurohumoral antagonism. Grade One patients have the best prognosis. In contrast, in Grade Three or Grade Four diastolic dysfunction, there is marked elevation of filling pressures and diuresis is needed under very controlled conditions. Maximum neurohumoral antagonism is needed and the patients must be followed very closely. These patients have a poor prognosis, and if appropriate early consideration for transplantation is needed.

Investigators from Boston showed that in the adult rat heart adenoviral gene transfer of sarcoplasmic reticulum Ca²⁺-ATPase (AdSERCA2a) restored diastolic dysfunction. With age, LV end diastolic pressure increases. Gene therapy with AdSERCA2a resulted in significantly lower end diastolic pressures that were closer to normal. Also with age, relaxation slows, but is returned to baseline levels with gene therapy.

A comparison of patients with classic hypertrophic cardiomyopathy (gene positive, LVH present) and gene-positive relatives without LVH (normal wall thickness) showed that even in the presence of normal morphology the latter group had abnormal Doppler tissue imaging. Doppler tissue imaging detects abnormalities in the actin myosin interaction at the molecular level. The finding of a non-invasive measure of LV function using Doppler tissue imaging that potentially predicts hypertrophic gene positivity before the development of hypertrophy, if substantiated by other investigators, will truly revolutionize the understanding and management of hypertrophic cardiomyopathy.

A 10% incidence of atrial fibrillation (AF) was shown in a study of 840 patients aged 75 to 100 years in Olmstead County, with a mean follow-up of four years. The development of AF could be predicted based on the size of the left atrium. A low incidence of AF was seen in patients with a low left atrial volume index, while there was a high incidence of AF in patients with a significantly elevated left atrial volume index. The left atrial digitation correlated with the Doppler finding. In patients with a normal pattern there was no incidence of AF at 5 years. For patients with Grade One or Grade Two dysfunction and abnormal relaxation, the risk of AF was intermediate. The highest incidence of AF was seen in the patients with Grade Three dysfunction and a restrictive pattern. This now provides a non-invasive marker to predict the patients at highest risk of developing AF. Diastolic dysfunction appears to be an important precursor of non-valvular AF in the elderly population, with an independent, graded relationship between the severity of diastolic dysfunction and the risk of non-valvular AF.