Aortic stenosis (AS) is the most frequent valvular heart disease in developed countries,1 with a steady increase in prevalence as the population ages. Progressive degeneration of aortic leaflets,2 age-related and enhanced by common cardiovascular risk factors,3 is the most frequent aetiology. Besides ‘calcific’ AS (the valvular disease of the elderly), the second most frequent aetiology is bicuspid calcific AS, followed by rheumatic AS.
Irrespectively of the aetiology, the natural history of AS is theoretically characterised by a long-standing asymptomatic period, lasting several decades, during which progressive left ventricular (LV) outflow-tract obstruction occurs. During this period, the risk of sudden death is relatively low (less than 1 % per year), which has led to AS being considered as a benign disease, even when severe.4 In patients with asymptomatic severe AS, prophylactic surgery is thus not recommended and current guidelines advocate delaying aortic valve replacement (AVR) until symptoms develop.
However, this ‘wait for symptoms’ strategy requires a careful follow-up, which is not always applicable to all patients, and prompt identification of the onset of symptoms. Indeed, death may occur soon after symptoms develop or, even worse, without any preceding symptoms, and if the waiting period for surgery is too long. Ideally, the surgical decision should be made sufficiently late to outweigh the surgical risk and early enough to avoid irreversible damage of the LV myocardium.5–7 Individual risk stratification could thus help to identify asymptomatic AS and who would be more likely to benefit from early elective surgery (see Table 1), with the goal of reducing mortality and avoid unnecessary intervention. Current guidelines consider surgery as reasonable in asymptomatic patients with reduced (<50 %) left ventricular ejection fraction (LVEF) and in patients who exhibit symptoms during an exercise test.8
Role of Echocardiography
Echocardiography has a key role in the evaluation of patients with asymptomatic AS. It allows the visualisation of valve morphology, the study of valve haemodynamics with quantification of AS severity and the evaluation of LV function and of the interaction between the left ventricle, the valve and the arterial system (see Figure 1).
Among echocardiographic parameters, peak aortic jet velocity (Vmax), valve calcification severity and the rate of haemodynamic progression are able to identify patients at higher risk of symptom development.
In numerous studies, Vmax was described as a useful predictor of outcome.4,9,10 Otto et al. found that a higher Vmax was related to a worse outcome over the whole spectrum of the disease severity. Patients with a Vmax >4 m/s had a two-year event-free survival less than 30 %, while patients with a Vmax <3 m/s had a five-year event-free survival >80 %.9 In a large population with haemodynamically significant AS, Pellikka et al. established a cut-off value of 4.5 m/s to identify patients at higher risk of adverse outcome.4 Lancellotti et al. also supported these findings in a subset of 163 patients with asymptomatic moderate-to-severe AS. A cut-off value of 4.4 m/s identified patients at higher risk of cardiovascular events (symptom onset, death, AVR) during a mean follow-up of 19 months.10 Rosenhek et al. studied the outcome of asymptomatic patients with very severe AS (Vmax >5 m/s) and showed that Vmax is still one of the best predictors of outcome.11 They compared patients with severe AS (Vmax between 4 and 5 m/s) to patients with very severe AS (Vmax >5 m/s), and also patients with Vmax between 5 and 5.5 m/s to patients with Vmax >5.5m/s. They showed that event-free survival in AS gets significantly worse as Vmax increases.11 Interestingly, in this study, aortic valve area (AVA) was not a predictor of outcome, probably because estimating the AVA by a continuity equation in clinical practice still remains challenging and prone to errors.
As pointed out by Rosenhek et al.,12 and confirmed by a study that assessed the degree of aortic calcification by electron-beam tomography,13 the degree of aortic valve calcification has to be taken into account when stratifying risk in asymptomatic AS. Patients with moderate-to-severe calcified valves have a poorer outcome compared with patients with a milder degree of calcification. Although, in that study, Vmax was not a predictor of outcome, a rapid increase in Vmax during follow-up was present in the group of patients who developed symptoms. Interestingly, when looking both at valve calcifications and haemodynamic progression, a group of patients with an impaired outcome was identified: patients with moderate or severe valve calcification and an increase in Vmax ≥0.3 m/s per year.12 This criterion has been included in the European recommendations for AVR (Class IIa).
The Left Ventricle
In AS, there is a chronic and progressive increase in LV afterload. This results in progressive LV remodelling and myocardial hypertrophy.14 LV hypertrophy is an adaptive mechanism, aimed at lowering LV wall stress to a certain level. When excessive and accompanied by interstitial fibrosis it becomes deleterious, leading to LV diastolic dysfunction.
A stiffer, non-compliant left ventricle fills itself under higher pressures and this chronically elevated LV end-diastolic pressure leads to left atrium (LA) enlargement. Elevated LV end-diastolic pressures also limit subendocardial perfusion and determine chronic subendocardial ischaemia with subendocardial fibrosis and subendocardial myocytes dysfunction. As a consequence, one of the first functions to be affected in AS is LV longitudinal function. Looking at LV longitudinal function might be a better way to identify subclinical myocardial dysfunction. At this stage of the disease, LVEF, the most commonly used measurement in echocardiography to appreciate LV performance, is still preserved, but the myocardial performance may not be normal at all.
Studies of LV longitudinal function confirmed that LV longitudinal systolic function is abnormal in AS, even in asymptomatic patients with preserved LVEF. Moreover, symptomatic patients with preserved LVEF had even more severe impairment of LV longitudinal systolic function.15–18 Whether this is a real contractility dysfunction of subendocardial fibres or just an early sign of afterload mismatch (with impairment of subendocardial deformation because of a very high systolic cavity pressure) is hard to determine.
LV longitudinal systolic function was tested as a prognostic marker in asymptomatic patients with AS. Peak systolic mitral annular velocity (s’) assessed by tissue Doppler imaging is a prognostic marker in asymptomatic patients with moderate-to-severe AS. Lancellotti et al. found that an s’ value ≤4.5 cm/s was linked to symptom onset, AVR and cardiac death in patients with asymptomatic severe AS and preserved LVEF.19 Similarly, in this setting, the same group found that a reduced global longitudinal strain ≤-15.9 % was linked to a more than twofold increase in the risk of cardiovascular events.10 Using speckle-tracking analysis to derive LV longitudinal strain, Lafitte et al. found a higher risk of future cardiac events in patients with asymptomatic severe AS and reduced basal longitudinal strain <-13 %.20 Hitherto, only few studies have assessed the short-axis function in AS. After an initial increase in both radial and circumferential strains, the chronic increase in afterload seems to be associated with a significant reduction in short-axis parameters.21–23 No study has evaluated the prognostic power of impaired short-axis function parameters in AS.
Diastolic dysfunction occurs early in AS. It has been shown that diastolic dysfunction and elevation of LV filling pressures could explain exertional dyspnoea in patients with severe AS.24 The E/e' (ratio of the mitral E wave velocity to e' wave velocity at the level of the mitral annulus) ratio correlates well with mean pulmonary capillary wedge pressure measured invasively and proved to be a good echocardiographic estimator of LV filling pressure.16
An E/e’ ratio ≥13 is able to identify patients with a LV end-diastolic pressure >15 mmHg with high sensitivity (93 %) and specificity (88 %).16 Impaired diastolic function is also a predictor of outcome in asymptomatic AS.19 In the study by Lancellotti et al., an E/e’ ratio >13.8 was able to identify a subset of patients at greater risk of future events, implying that the presence of severe LV diastolic dysfunction with elevated LV filling pressures is a marker of worse outcome in asymptomatic severe AS.19 An interesting parameter that estimates the LA contribution to LV filling, the late diastolic mitral annular velocity (a’ ≤9 cm/s) was also able to identify patients at a higher risk of events.19
LA dimension, a marker of chronic diastolic burden, holds prognostic information in patients with AS.10,19 Two studies used the LA-indexed area to show its prognostic information in patients with moderate-to-severe AS. The first study by Lancellotti et al. showed that an LA-indexed area ≥12.4 cm2/m2 was able to identify asymptomatic patients with a higher risk of future cardiac events.19 The second study by the same group, which enrolled 163 patients, confirmed these data.10 A yet simpler measurement of LA dimension, LA diameter, was shown to be able to predict progression of symptoms or all-cause mortality in patients with isolated AS and peak aortic pressure gradient ≥50 mmHg.25
The Valve, the Left Ventricle and the Vascular System
AS can no longer be considered as an isolated valvular disease. The complex changes that take place at the level of the aortic valve are linked to the complex changes that take place at the level of the arterial tree, namely atherosclerosis. In this view, a large number of patients with AS also have stiffer, remodelled and non-compliant arteries and their left ventricle has to face a double load: valvular and arterial. Briand et al.26 recently introduced the valvuloarterial impedance (Zva), an index that estimates global LV afterload. Zva is defined as the ratio between the sum of systolic arterial pressure and mean transaortic pressure gradient, and the indexed stroke volume. It represents the valvular and arterial factors that oppose ventricular ejection. High Zva values are associated with LV myocardial dysfunction.27 Retrospectively, Hachicha et al.28 have found a graded relationship between increased Zva and reduced overall survival in patients with asymptomatic AS. An Zva>5 mmHg/ml/m2 was independently associated with a fourfold increase in the risk of LV dysfunction and a value of >5.5 mmHg/ml/m2 was associated with a 2.5-fold increase in the risk of mortality in patients with symptomatic and asymptomatic AS.29 Lancellotti et al. prospectively followed 163 asymptomatic patients with moderate or severe AS and showed that a high Zva (≥5 mmHg/ml/m2) was a powerful predictor of reduced cardiac event-free survival.10
Role of Exercise Testing
Symptoms in AS are most often triggered by effort. This led to the use of exercise testing as a tool to unmask symptoms in such patients. Otto et al.30 established the safety of exercise testing in AS and encouraged its clinical use. Exercise testing is, indeed, useful to unmask falsely asymptomatic patients and to identify those who may become rapidly symptomatic. Actually, several studies show that approximately 30 % of patients who claim to be asymptomatic in their daily life have an abnormal exercise test.31–33 Exercise testing is also able to give prognostic information and seems to be superior to rest echocardiography in the risk stratification of patients with AS.31–33
The study of Amato et al.32 showed in 66 asymptomatic patients with severe AS (AVA<1.0 cm2) that the best predictor of event-free survival was a normal exercise test, although an AVA<0.7 cm2 was also able to discriminate patients with a poorer outcome. As opposed to the study of Das et al.,33 in which only exercise-induced symptoms were considered abnormal, the study of Amato et al. included ST segment depression, abnormal systolic blood pressure rise during exercise and complex ventricular arrhythmias as criteria to define an abnormal test.32 However, the positive predictive value of ST segment depression and even more that of abnormal blood pressure response remained unclear in that study. The study of Das et al. enrolled 125 patients with an AVA<1.4 cm2 and showed that exercise symptoms were superior to clinical history and resting echocardiography in predicting the outcome of physically active patients younger than 70 years of age. Conversely, in sedentary and older (>70 years) patients, the positive predictive value of exercise testing was rather low,33 reaching the poor predictive accuracy of ST segment changes and abnormal blood pressure response during exercise. These data confirmed the findings of Alborino et al. who reported a positive predictive accuracy as low as 55 % for both changes.31
To elicit symptoms and abnormal blood pressure response during exercise, both the European Society of Cardiology and the American Heart Association/American College of Cardiology recommend exercise testing of patients with severe AS who claim to be asymptomatic. According to the results of exercise testing, indications for AVR in patients with severe asymptomatic AS in the European guidelines are:
- an abnormal exercise test (symptoms during exercise) – Class IC;
- a fall in systolic blood pressure during exercise below baseline value – Class IIa; and
- an abnormal exercise test showing complex ventricular arrhythmias – Class IIb.
The American guidelines recommend AVR in asymptomatic patients with severe AS as a Class IIb indication if there is an abnormal response to exercise (symptom development or asymptomatic hypotension). A matter of concern is that the European survey on heart valve disease reported that exercise testing in asymptomatic patients with AS is rarely used.1
Role of Exercise Echocardiography
Parameters Addressing the Valve
Again, the study of Otto et al.30 remains a cornerstone study regarding the evaluation of patients with asymptomatic AS, introducing the concept that echocardiography can be used immediately after exercise cessation to evaluate the complex interaction between the stenotic valve and the left ventricle. The same group showed later on9 that exercise-induced changes in AVA, cardiac output and blood pressure were univariable predictors of outcome in patients with AS. However, in a multivariate analysis, exercise echocardiographic parameters seemed to be unable to provide additional prognostic information as compared to resting echocardiographic parameters. This might be explained in part by the inclusion of patients with mild AS, which has attenuated the prognostic power of exercise echocardiography. Moreover, imaging was performed after test termination, which might have led to the loss of some of the key information on left ventricle–valve interaction during the test. From then on, the development of a tilting bicycle stress echocardiography table made it possible to monitor cardiac imaging continuously throughout the test.34
Lancellotti et al. showed that an increase in mean transaortic gradient ≥18 mmHg is a powerful predictor of outcome in patients with severe asymptomatic AS. This study was the first one to underline that, in these patients, risk stratification might be improved by exercise echocardiography. The more recent multicentre study of Marechaux et al. reinforced the role of exercise echocardiography in predicting clinical outcome in patients with asymptomatic AS.35 The study enrolled 135 truly asymptomatic patients (an abnormal exercise test was considered an exclusion criteria) with at least moderate AS and preserved LVEF and showed that, at age ≥65 years, the presence of type 1 diabetes, a LV hypertrophy, a mean gradient at rest >35 mmHg and an exercise-induced increase in mean transaortic gradient >20 mmHg were independent predictors of outcome in a multivariable analysis. The combination of a resting gradient >35 mmHg and an exercise-induced rise in mean transaortic pressure gradient >20 mmHg identified a group of patients with a markedly increased risk of events during follow-up. The increase in mean transaortic pressure gradient remained a powerful predictor of the outcome, even in patients with moderate AS. These latter patients should benefit from a closer follow-up than what is recommended classically.
Parameters that Address the Left Ventricle
Another study by Marechaux et al. investigated the LV response during exercise in patients with severe asymptomatic AS and preserved LVEF at rest.36 The LVEF, although an imperfect estimator of LV performance, usually remains at rest within normal limits for years. However, during exercise, owing to the development of LV afterload mismatch, LVEF might decrease, which might help in identifying patients with a worse clinical outcome. This hypothesis was highlighted by Marechaux et al. and confirmed later by Lancellotti et al. A small increase or a decrease in LVEF during exercise represents a determinant of abnormal exercise response and is associated with lower event-free survival.37 Van Pelt et al., more specifically, showed that changes in LV longitudinal function – as assessed by post-exercise changes in s’ by tissue Doppler imaging – were attenuated in patients with AS.38 Donal et al. confirmed these findings by using speckle tracking to assess LV longitudinal function.18 The study showed that, when compared with controls, patients with AS had reduced myocardial longitudinal function and limited longitudinal contractile reserve. These findings were more frequent in patients with an abnormal exercise test. Of note, limited contractile reserve not only reflects the exercise-afterload mismatch, but also impairs coronary flow reserve with or without significant coronary stenosis.
Role of Biomarkers
Plasma brain natriuretic peptide (BNP) levels have been shown to correlate with disease severity in AS, with symptomatic status and with New York Heart Association functional class.39–42 More importantly, different studies have shown that BNP levels are also able to give prognostic information in patients with asymptomatic AS.43–46 The study of Bergler-Klein et al. showed that patients with severe AS and BNP plasma levels <130 pg/ml or N-terminal brain natriuretic peptide (Nt-BNP) levels <80 pmol/l were unlikely to develop symptoms during a nine months follow-up period, having a symptom-free survival close to 90 %.43 However, patients with BNP or Nt-BNP above the mentioned cut-off values had a symptom-free survival rate of less than 50 %. Similarly, the study of Lim et al. showed that high BNP levels are significantly associated with poor outcome in asymptomatic patients with severe AS and that BNP levels are highly accurate when used to separate symptomatic from asymptomatic patients with severe AS.44 The study of Monin et al. showed that the combination of serum BNP levels with female sex and Vmax (Monin’s score) can help discriminate between patients who will experience cardiac events during a two-year follow-up and those who will not.46 For score values less than 11, a low event rate was found (<10 %), whereas score values above 16 were associated with higher event rate (>75 %).
In between (score >11 but <16) there was a steep linear increase in event risk that might suggest a more frequent surveillance may be useful for these patients. Another study19 showed that patients with asymptomatic AS, especially women, who have impaired LV long-axis function with peak s’ wave ≤4.5 cm/s and increased BNP levels ≥61 pg/ml are at increased risk of events. Although very promising as a predictor of outcome in AS, no cut-off value for BNP levels has been accepted to select patients who might benefit from surgery.
Although intensively studied, AS remains one of the valvular diseases that are incompletely defined and understood. The inter-patient variability regarding the progression of the disease, the complexity and variability of LV response to chronically increased afterload continue to puzzle clinicians. This has led to the necessity for individual risk stratification by using a large armamentarium of techniques. The outcome can only be satisfactorily evaluated by favouring an integrative approach, which should take into account the complex interplay between the valve, the ventricle and the arterial system, not only at rest but also during exercise.