• aortic stenosis

Both, European and North American guidelines for the treatment of valvular heart disease1–3 define severe aortic stenosis (AS) primarily by transvalvular velocity, mean gradient and valve area. In agreement, cut-offs have been set at 4 m/s for peak velocity, 40 mm Hg for mean gradient and 1.0 cm² for aortic valve area. These cut-offs have remained unchanged now for many years and have been used in clinical practice as well as in research studies of AS.

Assessment of AS severity requires quite a number of considerations. Transvalvular pressure gradients are well known to be flow dependent with the clinically most important consequence that patients with severe AS may present with velocities below 4 m/s and mean gradients less than 40 mm Hg when transvalvular flow is reduced. This highlights the importance of estimating the valve area when evaluating AS and is the reason why AS by definition can be classified severe if only the valve area is smaller than 1.0 cm², while velocity and gradient remain below the recommended cut-offs (ie, ‘low gradient severe AS’).1–3 Although valve area may from a theoretical point of view indeed be the ideal measure to quantify AS, its assessment remains operator dependent and less robust than gradient estimates in clinical practice. Thus, current guidelines recommend that valve area alone with absolute cut-off points should not be relied on for clinical decision-making but should always be considered in combination with flow rate, pressure gradients, ventricular function, size and wall thickness, degree of valve calcification and blood pressure, as well as functional status.1 3

The subgroup of ‘low-gradient AS’ and the decision whether AS is indeed severe in these patients has become particularly challenging in clinical practice. This prompted a detailed recommendation paper by the European Association of Cardiovascular Imaging joint with the American Society of Echocardiography3 as well as more detailed recommendations for the diagnostic workup of AS in the recently published guidelines on valvular heart disease of the European Society of Cardiology.1 While severe AS is considered unlikely if cardiac output (more precisely, transvalvular flow) is normal and there is a mean pressure gradient <40 mm Hg (ie, ‘normal flow—low-gradient AS’), precise further evaluation is recommended when a low flow status (ie, stroke volume index <35 mL/m²) is present.1 3 In patients with poor systolic left ventricular function low dose dobutamine echocardiography is recommended as it may help to distinguish truly severe AS from pseudo-severe AS (functionally small valve area due to low opening forces). Patients with a valve area <1.0 cm² and mean gradient <40 mm Hg at low flow rates despite preserved left ventricular ejection fraction remain the most challenging subgroup in terms of making the right diagnosis. It has been demonstrated that these patients may indeed also have only moderate AS4 5 and that there may frequently be reasons other than an underlying severe AS for this combination of measurements.3 (1) Doppler measurements tend to underestimate flow resulting in eventual underestimation of valve area and erroneous assumption of ‘low flow conditions’. (2) Small body size may be present. (3) The cut-offs for gradients are not entirely consistent as it has been demonstrated that the aortic valve area has to be closer to 0.8 cm² than 1.0 cm² to generate a mean gradient of 40 mm Hg.6 Thus, diagnosis of severe AS in this setting requires careful exclusion of these other reasons for such echo findings, before making the decision to intervene. Evaluation of the degree of calcification by CT has gained increasing importance in this setting. It has not only been demonstrated to be related to AS severity in this context but also to outcome and may therefore help in decision-making. An integrated approach including clinical data, semiquantitative and quantitative imaging data is recommended in this difficult subgroup.3

In summary, the complexity of grading the severity of AS has been well recognised, and current clinical practice guidelines recommend therefore not to rely on single numbers in particular of valve area measurements alone.1 3 Nevertheless, it has been generally accepted that research studies of AS use the inclusion criterion ‘peak velocity ≥4 m/s or mean gradient ≥40 mm Hg or valve area <1.0 cm², to select patients with severe AS. Retrospective analyses searching large echo databases and using this definition to identify patients with severe AS are very common.

In this context, the study by Steven Bradley and colleagues7 published in the current issue of Heart deserves particular attention. In a retrospective analysis of a huge database including transthoracic echo reports of 213 174 patients, they determined sensitivity, specificity, positive predictive value, negative predictive value and accuracy of measures for severe AS using the interpretation of the echocardiographic reader that eventually guided clinical care as the reference standard. Relative to the documented interpretation, aortic valve area and even more the indexed area as a measure of severe AS had an excellent sensitivity but an alarmingly low positive predictive value of only 37.5%. As one might expect, peak velocity and mean gradient were highly specific but had a poor sensitivity of <70%. Remarkably, a measure incorporating the easily obtainable parameters of peak velocity, mean gradient and dimensionless index (either by velocity time integral (VTI) or peak velocity ratio) achieved a balance of sensitivity (92%) and specificity (99%) and an accuracy of 99%. In addition, using this measure, the proportion of patients whose echocardiogram could be assessed for AS severity was 80% as compared with 53% by documented reader interpretation alone.

These findings may have implications for both, research and clinical practice.

Inadequate definition of severe AS may indeed be the major reason for the divergent results of low-gradient AS studies where outcome ranges from similar to moderate AS to worse than high-gradient AS and valve replacement improves or even seems to worsen outcome.4 5 Looking at the differences in study populations raises indeed the suspicion that—although all studies used ‘guideline definitions of severe AS’—the percentage of patients with indeed severe AS in these low gradient cohorts may have markedly differed. The authors’ point that more care may be required with regard to patient selection in AS studies appears indeed to be well taken.

The fact that in the present study, nearly two of every three patients categorised as having severe AS on the basis of valve area <1.0 cm² were categorised as non-severe AS by the interpreting clinician may also question the recommendation of this measure as a first-line criterion in clinical practice at least when using the standard continuity equation. One of the major limitations of this approach is the left ventricular outflow tract (LVOT) measurement. While velocity measurements have a very low intraobserver and interobserver variability, it is markedly higher for the LVOT diameter which is then even squared for the calculation.3 Even more important is the assumption of a circular shape of the LVOT. Multislice CT (MSCT) studies have now very well confirmed that the LVOT is in general, elliptical, rather than circular resulting in underestimation of LVOT cross-sectional area and in consequence underestimation of stroke volume and eventually of valve area (on average approximately 0.2 cm² in one study). Recent data suggest that more precise measurements of the valve area using MSCT (or 3D echocardiography) LVOT area may be of particular importance in low-gradient AS to avoid misclassification of AS.8 If it holds true—as suggested by the present study—that the dimensionless index which is much easier to obtain performs much better than an echocardiographically estimated valve area in identifying severe AS, this measure may gain more importance in the future. Currently, it is recommended as an alternate measure of AS severity, only. While current recommendations acknowledge the advantage of avoiding LVOT measurements with their major limitations, they also raise concerns about neglecting the variability of LVOT size with its impact on actual transvalvular flow when using the dimensionless index.3 The fact that this measure performed equally well in man and women as well as with normal and reduced ejection fraction, normal and low flow conditions in the presented paper may question these concerns for its use in clinical practice.

However, before guidelines and recommendations can be modified more research seems to be required. The major limitation of the current study is the reference standard. It remains unclear how exactly the echocardiographic readers came to their conclusion that AS grade was non-severe when guideline criteria assigned severe AS and how correct they were in their judgement. The most recent recommendations for the assessment of AS were not available at the study period and could therefore not be the reference.

In any case, the study should stimulate further research. It would be interesting to see whether AS severity as determined by the currently recommended approach can also be accurately predicted by the—in comparison relatively simple to obtain—combined measure of peak velocity, mean gradient and dimensionless index. If this were demonstrated, the algorithm of an integrated approach could possibly be simplified. More importantly, further research should evaluate how dimensionless index is related to outcome of AS and to the impact of intervention on the outcome.

In conclusion, Bradley et al 7 raise important concerns in their paper on the use of routine captured echocardiographic data in the diagnosis of severe AS. The most recent guideline and recommendation documents1 3 emphasise that a valve area <1.0 cm² calculated by standard continuity equation only suggests severe AS but requires careful further evaluation to decide whether AS is indeed severe when peak velocity and mean gradient do at the same time not reach the recommended cut-offs. Although the currently recommended algorithm for the diagnosis of severe low gradient AS may be complex, any simplification with more emphasis on the dimensionless index would require further validation. However, inclusion criteria identifying severe AS in research projects—particularly in retrospective analyses—must indeed be carefully revisited. Identification of severe AS by an echocardiographically reported valve area of <1.0 cm² alone should not be acceptable.