In the developed world, aortic stenosis (AS) represents the most prevalent valvular heart disease. The final stage of AS is transformation of the aortic valve in a severely restricted, thickened, calcific valve; however, the initiating process is less likely to be a degenerative one, but rather similar to atherosclerotic plaque formation. The process is age-dependent: more than 25% of individuals above 65 years of age have aortic sclerosis (i.e. valve thickening without significant leaflets motion reduction), but this prevalence reaches 48% in those over 85 years of age. A minority of these sclerotic valves progress to haemodynamic stenosis, which is noted in up to 7% of individuals over 65 years of age. While even severe AS may be well tolerated for long periods of time, once it becomes symptomatic the prognosis without surgery is dismal, with a yearly mortality rate of 25%, up to three-quarters of patients expected to die within three years and a median survival of about one year if cardiac failure is present.1,2 Peri-operative mortality for isolated aortic valve replacement (AVR) is higher than for uncomplicated coronary artery bypass graft (CABG) – 3–5% below 70 years of age and 5–15% in the older population3 – but results are excellent and AS patients who undergo successful AVR have a similar life expectancy to a matched population free of aortic valve disease. Current guidelines mandate AVR for patients with severe AS and any of the following symptoms: left ventricular (LV) systolic dysfunction (not related to other causes), need for CABG or aortic surgery or abnormal response to exercise.3 Of note, the presence of symptoms, which is by definition subjective and, occasionally, difficult to interpret in the presence of other pathologies, has to be supplemented by objective evidence of AS severity and/or its effect on the LV.
Historically, cardiac catheterisation defined AS in terms of directly measured transvalvular gradient and aortic valve area (AVA) calculated by the Gorlin formula. However, crossing the aortic valve and performing a full right and left study increases the duration and the risk of the procedure and provides only haemodynamic but little morphological and functional information, and is not suitable for serial follow-up. For all practical purposes, echocardiography has displaced heart catheterisation for the work-up of AS and today is the main diagnostic tool in assessing AS patients. Beyond gradient and area values, echocardiography provides a comprehensive assessment of the aortic valve and aortic root morphology, which is of interest when planning the surgery, and of coexistent cardiac pathologies (see Table 1).
It is also instrumental in assessing special subgroups of patients with low gradients or decreased LV contractility, or those who may be considered for AVR even if asymptomatic. The central role of echocardiography in the management of AS patients is acknowledged by the use of echocardiographic indices to define AS severity and indications for surgery3,4 (see Table 2). The main echocardiographic indices used to assess AS severity are summarised in Table 2 and discussed below.
Flow Velocities and Gradients
Both the peak flow velocity (Vmax) and the mean gradient are obtained by Doppler interrogation of the aortic flow. As such, good alignment of the Doppler line and the flow direction (<20º) is required for accurate and reproducible results. As gradients are calculated as ΔP=4(V22–V12) (full Bernoulli equation), Vmax is used rather than peak gradient to minimise the effect of small variations of velocity readings on the final result; this is less of a problem for the mean gradient that is obtained by integrating all the instantaneous gradients generated during ejection and is not a simple computation of mean velocity. For valvular jet velocities (V2) >3m/second and subvalvular velocities (V1) <1.5m/second, the latter can be ignored (simplified Bernoulli equation: ΔP=4V2), otherwise, both the proximal and the distal velocities have to be used (full Bernoulli equation). Furthermore, the cut-off values mentioned in Table 2 are valid if the LV ejection fraction (LVEF) is normal and there is no severe regurgitation across the valve.
Sources of Error for Gradient Calculations
The true gradient can be either under- or overestimated due to technical flaws in the execution of the study.
- Execution flaws resulting in gradient underestimation:
- misalignment of the Doppler line with the main flow direction; and
- missing the transducer position/window providing the optimal signal.
- Execution flaws resulting in gradient overestimation:
- contamination or confusion with a different, higher-velocity systolic flow (mitral regurgitation [MR], LV outflow tract [LVOT] obstruction); and
- inclusion of a beat following a long diastole in measurements.
Aortic Valve Area
AVA is usually calculated using the continuity equation: AVA = VTIV1xCSAV1/VTIV2, where VTI is the traced velocity–time integral at LV outflow tract (V1) and aortic valvular (V2) levels, and CSAV1 is the cross-sectional level area determined by measuring the LVOT mid-systolic diameter. All modern echocardiographic machines have incorporated analysis software to calculate AVA from the traced VTIs and the measured LVOT diameter. An alternative method to obtain the AVA is by direct planimetry of the valve orifice in parasternal short-axis view, using either transthoracic or transoesophageal echocardiography (TEE) (see Figure 1).
Sources of Error for Aortic Valve Area Calculations
The continuity equation method is subject to gradient error calculations (see above) and inaccurate LV outflow tract measurements, for which inter- and intra-observer variability may reach 8%.4 This dimension, being squared in the continuity equation formula, can result in significant overor underestimation of valve area. The direct planimetry of the aortic valve orifice requires good-quality images, occasionally available only with TEE. Even with TEE, aortic valve direct planimetry may be inaccurate with a heavily calcified valve and is therefore considered to be an acceptable alternative when Doppler measurements are unreliable, but it is not a primary method to assess AVA.
Dimensionless Velocity Index
The ratio between the subvalvular (LVOT level) and the valvular peak velocities is a ‘stripped-down’ version of the continuity equation that ignores the LVOT diameter and thus is not subject to errors related to its measurements. A dimensionless velocity index (DVI) <0.25 is indicative of severe AS, with a valvular area of 25% of the expected normal valve area for the patient’s body size. DVI does not provide a valve area but rather confirms or weakens the qualitative diagnosis of severe AS. The velocity index is also useful to differentiate between high valvular gradients due to truly severe AS and mild AS with increased velocities due to high-flow conditions such as sepsis or hyperthyroidism, when the DVI remains >0.3 as both the valvular and subvalvular velocities are high.
Real Dilemmas in the Diagnosis of Aortic Stenosis
The limitations and possible errors described above are generally well known and, although occasionally confusing, their avoidance by appropriate technique and awareness is expected from a good echo study. Beyond these ‘technical’ pitfalls, however, the diagnosis of AS can raise real dilemmas when understanding the true haemodynamic and clinical meaning of echocardiographic results and conclusions is challenging. The echocardiographic quandaries one may encounter in the diagnosis of AS relate mainly to contradictory results and lack of concordance between accepted echocardiographic indices of AS severity or between echocardiographic and catheterisation results (see Table 3). The challenge of appropriate quantification of AS in patients with decreased LV function is well recognised. However, even patients with good LV contractility exhibit a mixture of haemodynamic patterns. In their review of 3,483 echocardiographic studies of patients with various degrees of AS and normal EF, Minners et al. found in 30% of cases a lack of concordance between the different criteria of AS severity. Moreover, Vmax and gradient were in the range of severe AS in 40–45% of patients, while by AVA 69–76% of them were diagnosed as having severe AS.5
Aortic Stenosis Severe by Gradient but Mild to Moderate by Aortic Valve Area
Occasionally, high (>4m/second) peak velocities and elevated mean gradients (>30–40mmHg) are found in patients whose AVA by continuity equation is only in the mild to moderate range (>1cm2). Frequently, the reasons for this discordance are related to execution errors of the echocardiographic study:
- incorrect positioning of the pulsed-wave Doppler sample volume too close to the aortic valve, so that the LVOT signal is contaminated by the high-velocity valvular flow; and
- erroneous measurement resulting in overestimation of the LVOT diameter.
Real discrepancies, i.e. high velocities and gradients in the absence of significant AS, can occur in patients with a high cardiac output state, such as sepsis, hyperthyroidism, anaemia or with AV fistulas. Awareness of the patient’s clinical condition and a DVI >0.25 should clarify this condition.
Aortic Stenosis Mild to Moderate by Gradient but Severe by Aortic Valve Area
Faced with this discrepancy, the first step is to establish the contractility of the left ventricle and dichotomise these low-gradient severe AS patients into those with reduced or normal EF.
Patients with Reduced Left Ventricular Contractility – Low-flow–Low-gradient Aortic Stenosis
LV systolic dysfunction can be present in AS patients as a result of either concomitant pathology (coronary artery disease or cardiomyopathy) or of long-standing severe AS. Low-flow–lowgradient AS (LF/LGAS) is defined as a combination of AVA <1cm2, mean gradient <40mmHg and LVEF <40%, and is described in 5–10% of patients with AS.6,7 Although not very frequent, LF/LGAS poses the echocardiologist with two challenges:
- to decide whether the AS is truly severe and the LG reflects low transvalvular flow or is mild to moderate only and the low area reflects the inability of a hypokinetic ventricle to fully open a mildly restricted valve; and
- to risk-stratify the patient in terms of of peri-operative risk and possible benefit of AVR.
Establishing whether a given patient with LF/LGAS has adequate LV contractile reserve (CR) is critical in answering both questions. The accepted approach is to perform low-dose dobutamine stress echo study (DSE) and to quantify the inotropic response and the changes in AVA and transvalvular gradient. Usual doses are 5–20mcg/kg/minute, and although the dose-response to dobutamine is unpredictable and the inotropic response does not necessarily parallel the chronotropic and blood pressure response,8 an increase in heart rate is generally taken as proof of dopaminergic stimulation sufficient to elicit an inotropic response. CR is considered to be present if the dobutamine infusion results in ≥20% increase in cardiac output.6,9 The possible response patterns to dobutamine in patients with LF/LGAS are summarised in Table 4. Patients with true severe AS and evidence of CR have a clear indication for AVR. Those without CR have a poor operative mortality and generally are not candidates for AVR, but recent data suggest that their outcome with medical management is so poor that, at least in selected cases, they should be considered for intervention.10
Patients with Normal Left Ventricular Contractility and Low-gradient Severe Aortic Stenosis
Conventional teaching postulates that a well contracting ventricle should generate high peak velocities and mean gradients in the presence of severe AS, otherwise the diagnosis is doubtful. However, in clinical practice we are not infrequently challenged by the reality of patients with severe AS by both valve appearance and calculated valve area (AVA <1cm2) and who present with a mean gradient in the mild to moderate range (<40mmHg) despite normal LV contractility.
Obviously, technical errors have to be excluded, but this possible presentation of severe AS has recently been increasingly recognised11–14 and has been described in up to 42% of patients with severe AS and normal LV contractility.5 Importantly, these patients do not seem to have a better prognosis than their ‘high-gradient’ counterparts.12–14 Possible explanations for this haemodynamic pattern include:12–14
- relatively low stroke volume, which is not suggested by an apparently normal EF – this could be related to small LV cavity (small-sized patients, severely hypertrophic ventricles) or occult LV systolic dysfunction (elderly patients, LV hypertrophy); and
- higher systemic vascular resistance and LV afterload.
The importance of recognising this not uncommon haemodynamic pattern cannot be overemphasised, since AVR, when appropriate, should not be denied to these patients due to a possibly misleading ‘not severe enough’ gradient.
Invasive confirmation of gradients and AVA by Gorlin formula should not be routinely sought as part of the work-up of patients with AS.3,4 Although there is generally good agreement between echocardiography and direct measurements,15,16 discrepancies are noted. Occasionally, catheterisation gradients may be higher than Doppler gradients due to possible underestimation of the true gradient by echocardiography (see ‘Sources of Error for Gradient Calculation’ above). A more puzzling and, in fact, more frequent scenario occurs when the catheterisation gradient is significantly lower than the echocardiographic one. As seen in Figure 2, while catheterisation records the virtual ‘peak-to-peak’ gradient, Doppler interrogation records the true, instantaneous transvalvular gradient at the time of peak LV systolic pressure, which is uniformly higher due to the relative morphologies and timings of the aortic and ventricular pressure waveforms. The more severe the AS, the higher this discrepancy may be, due to a more delayed aortic pressure peak. In addition, the pressure difference between the left ventricle and the ascending aorta may diminish distal to the valve, a haemodynamic phenomenon described as ‘pressure recovery’.17 This pattern is more likely to occur in patients with small aortic roots and may be responsible for marked discrepancies between echocardiographic and invasive gradients in 10–14% of patients.17,18
A combination of these two mechanisms explains the reality of vexing disagreements between catheterisation and echocardiographic assessment of AS patients. There are no rules to decide which method is ‘right’, but careful inspection of valve appearance by echocardiography and integration of all clinical data should provide the answer. If the echocardiographic data seem ‘solid’, the diagnosis of severe AS should not be discarded because the catheter-derived gradient is in the mild to moderate range (see Figure 3). Even better, the dilemma of a catheterisation–echocardiography discrepancy can be avoided ‘by not crossing the aortic valve with a catheter’.1
Echocardiography is the first-line diagnostic tool in the assessment of patients with AS. Cut-off values define severity criteria used to decide appropriateness of intervention. A large minority of patients does not fulfil all accepted criteria and may present with perplexing haemodynamic patterns and echocardiographic results. Awareness of the sources of possible errors and of less typical echocardiographic results is essential for the correct management of AS patients whose echocardiographic studies are, apparently, confounding.