Article Text

Download PDFPDF

Assessment of aortic stenosis severity: when the gradient does not fit with the valve area
  1. Philippe Pibarot,
  2. Jean G Dumesnil
  1. Québec Heart and Lung Institute, Department of Medicine, Laval University, Québec, Canada
  1. Correspondence to Dr Philippe Pibarot, Institut Universitaire de Cardiologie et de Pneumologie de Québec, 2725 Chemin Sainte-Foy, Québec, Quebec, Canada; philippe.pibarot{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

In this issue of Heart, Minners et al1 (see page 1463) provide an extension of their previous work2 where they retrospectively analysed the data of their Doppler-echocardiography laboratory and reported that there is a discrepancy in the criteria of aortic valve area (AVA; <1.0 cm2) and mean gradient (>40 mm Hg) proposed in the guidelines to define severe aortic stenosis (AS).2 In the present study1 they present data obtained by cardiac catheterisation in a subset of the previous series.2 The main findings of this study are: (1) when using the framework of the current guidelines, inconsistent grading of AS (ie, AVA <1.0 cm2 but gradient ≤40 mm Hg) occurred in 36% of patients with preserved left ventricular (LV) systolic function (LV ejection fraction (LVEF) ≥50%), and this proportion was similar irrespective of the method used to assess stenosis severity (ie, Doppler-echocardiography vs cardiac catheterisation); and (2) the proportion of patients with reduced stroke volume (stroke volume index ≤35 ml/m2) despite apparently normal LV systolic function was substantially higher in the subset of patients with inconsistent grading (52%) than in those with consistent grading (29%).

This latter finding lends further support to the concept that discordance between AVA and gradient is often due to paradoxical low-flow AS, a disease pattern recently described by our group.3 We indeed reported that an important proportion of patients with severe AS may paradoxically have a low flow and thus often a low gradient, despite the presence of normal LVEF.3 When compared with patients with normal LV outflow, patients with paradoxical low flow are characterised by a higher prevalence of women and concomitant hypertension, older age, a higher degree of LV concentric remodelling, impaired LV filling, smaller end-diastolic volume and reduced mid-wall and longitudinal shortening. These patients also have markedly increased global (valvular+arterial) LV haemodynamic load as reflected by increased valvuloarterial impedance (Zva), which is calculated using the formula: Zva=(systolic blood pressure+mean transvalvular gradient)/stroke volume index. Overall, these features suggest that these patients are in fact at a more advanced stage of their disease. The prevalence and typical features of this pattern have subsequently been corroborated by other studies4 5 including a substudy from the SEAS trial conducted in a large (>1800) series of patients. Among the 374 patients found to have severe AS in this study,4 28% had paradoxical low flow and 59% of them had a mean gradient <30 mm Hg compared with 35% and 65%, respectively, in our own cohort.3 6

Previous studies also show that patients with paradoxical low flow have 40–50% lower referral to surgery than those with normal flow,3 5 6 probably due to underestimation of stenosis severity in light of the relatively low gradient. Nonetheless, in this subset of patients, aortic valve replacement (AVR) was associated with a better outcome than medical treatment, even after adjustment for the level of gradient (> or <40 mm Hg) and for differences in baseline risk profile of surgical versus medical patients. Interestingly, AVR was also found to be associated with better survival in patients with a small AVA (<1.0 cm2) and normal flow, including those with low gradient. This latter subset corresponds to the patients with inconsistent grading and normal stroke volume described in the paper by Minners et al.1

In this context and in the absence of supportive outcome data, their actual and previous recommendation2 that the cut-off value of AVA to define severe AS should be lowered to 0.8 cm2 in the guidelines does not appear justified. On the contrary, given that (1) the gradient is much more flow-dependent than the AVA; (2) recent studies, including the present study by Minners et al,1 have shown that a high proportion (30–35%) of patients with AS and preserved LVEF may nonetheless have low stroke volume and transvalvular flow rate3–6; and (3) several studies have reported that an AVA<1.0 cm2 or indexed AVA <0.6 cm2 predicts excess mortality and morbidity irrespective of gradient or peak velocity,3 5–7 we would rather be in favour of the status quo for AVA criteria and more flexibility in the definition/interpretation of gradient criteria. More importantly, however, since both parameters have inherent limitations (namely, gradient is highly flow-dependent whereas AVA is more prone to measurement errors), the application of strict criteria to define severe AS may be misleading and we would rather recommend the use of a more comprehensive approach that integrates other Doppler-echocardiographic findings as well as other diagnostic modalities, if needed.

Practical approach when confronted with discordant findings

There are several potential causes of discordance between AVA (eg, 0.8 cm2) and gradient (eg, 30 mm Hg) in patients with preserved LVEF including (1) measurement errors; (2) small body size; (3) paradoxical low-flow AS; and (4) inconsistent grading related to intrinsic discrepancies in guidelines criteria, and it is important to make the differential diagnosis between these four potential situations given that they have markedly different implications in terms of therapeutic management (figure 1).

Figure 1

Proposed algorithm for differential diagnosis in patients with aortic stenosis and preserved LV ejection fraction (LVEF) presenting with a small aortic valve area (<1.0 cm2) but a low mean gradient (<40 mm Hg). AS, aortic stenosis; AVA, aortic valve area; AVAi, indexed aortic valve area; ∆Pmean, mean transvalvular gradient; BNP, B-type natriuretic peptide; CT, multislice computed tomography; LVEDD, LV end-diastolic diameter; LVEDVi, indexed LV end-diastolic volume; RWTR, relative wall thickness ratio; SVi, stroke volume index; Zva, valvuloarterial impedance.

Measurement errors

The stroke volume and therefore the AVA may be underestimated because of underestimation of LV outflow tract and/or misplacement of pulsed-wave Doppler sample volume. This situation may lead the treating physician to conclude that the patient has a paradoxical low-flow low-gradient severe AS whereas, in fact, he or she has a moderate AS with normal flow. Several methods can be used to corroborate the Doppler-echocardiographic measurements of stroke volume and AVA. For example, in the absence of significant mitral regurgitation, the stroke volume can easily be estimated by multiplying the LVEF by the LV end-diastolic volume obtained by the Teicholz formula. If the stroke volume measured by this independent method is consistent with the stroke volume measured in the LV outflow tract, one can be reassured about the accuracy of the measurement of stroke volume.

Small body size

A patient with a small body surface area may have a relatively low flow and thus a low gradient despite the presence of a small AVA. This situation can easily be ruled out by calculating the indexed AVA and, for this reason, this should be done in every patient. A value >0.6 cm2/m2 indicates the presence of moderate AS (figure 1).

Paradoxical low-flow AS

Once measurement errors and small body size have been ruled out by calculating the indexed AVA and by corroborating the estimation of stroke volume and AVA by several independent methods, the next step is to make the differential diagnosis between paradoxical low-flow AS versus inconsistent grading due to discrepancy in the guidelines criteria. This can be achieved by calculating the stroke volume index and by identifying the typical features of paradoxical low-flow AS (figure 1)—that is, a stroke volume index ≤35 ml/m2, LV geometry (LV end-diastolic internal diameter <47 mm and/or LV end-diastolic volume index <55 ml/m2), a noticeable increase in relative wall thickness ratio (ie, >0.45) and Zva >4.5 mm Hg/ml/m2. To that effect, it should be noted that a reduction in cardiac output may result in lower gradients and a pseudo-normalisation of a previously high blood pressure, and that the calculation of Zva may be particularly helpful in these circumstances since these patients are often at a more advanced stage of their disease. On the other hand, it cannot be excluded that, as in low-flow AS with low LVEF, the AVA might be ‘pseudo-severised’ in some patients—that is, the flow may not be high enough to fully open a valve that is only moderately stenotic. However, as of yet, there is little experience with identifying this phenomenon and the value of exercise stress and/or dobutamine echocardiography in this context remains to be determined. Dobutamine stress echocardiography should, however, be used with caution since it might cause a reduction in preload, stroke volume and blood pressure in these patients who often have a restrictive physiology. Plasma natriuretic peptides might also prove helpful in this context.

The results of the study of Minners et al suggest that, based on catheterisation data, the prevalence of paradoxical low-flow low-gradient AS is about 15%, which is threefold higher than that of ‘classical’ (ie, low LVEF) low-flow low-gradient AS.1 Since patients with paradoxical low-flow low-gradient AS have greater potential for technical pitfalls, measurement errors and ‘pseudo-severisation’ of AVA, it is also relevant to further corroborate stenosis severity with the use of methods independent of flow, especially if the patient is symptomatic and is being considered for surgery. In this regard, multislice CT might prove useful, a valve calcium score >1500–1600 Agatston Units being in favour of a true severe stenosis. On the other hand, in the event of a low calcium score, the echocardiographic data should be reassessed and additional diagnostic tests should be considered.

Inconsistent grading related to discrepancy in guidelines criteria

If the differential diagnosis process (figure 1) rather indicates the presence of normal flow (stroke volume index >35 ml/m2) and inconsistent grading related to intrinsic discrepancy in the guidelines criteria, the patient may have borderline moderate/severe AS.

Nonetheless, regardless of the conclusion of the differential diagnosis process, it should be remembered that the management of the patient will ultimately be dictated by the patient's symptomatic status. Hence, if the patient is asymptomatic, exercise testing may be used to confirm that the patient is truly asymptomatic, especially if the global haemodynamic load is severely increased (ie, Zva >4.5 mm Hg/ml/m2). If, on the other hand, the patient is symptomatic, further corroborating methods including valve calcium quantification by CT, B-type natriuretic peptide or dobutamine stress echocardiography can be used and, if these independent diagnostic modalities suggest the presence of severe stenosis, the patient should be referred to surgery. If, on the other hand, the stenosis is found to be moderate but the global haemodynamic load as measured by the Zva is increased (>3.5), the symptoms could be due to the additive effects of moderate AS and reduced arterial compliance and/or increased vascular resistance. In this situation, treatment of arterial hypertension should be contemplated. Patients who remain symptomatic despite optimal antihypertensive treatment are particularly challenging. Indeed, it may eventually be found that some of these patients might actually benefit from AVR even though they do not meet the currently accepted criteria for AS severity, the rationale being that their total LV haemodynamic load is markedly increased and that any significant decrease might contribute to improving their prognosis and well-being.

Further outcome studies are of course needed to address the several unanswered questions raised by the study of Minners et al.1 However, in these future studies it would be inadequate to include AVR in the end-points, given that this end-point is essentially determined by the treating cardiologist's perception of disease severity which in turn, is highly influenced by the magnitude of the gradient (or peak jet velocity), the latter often being pseudo-normalised in patients with paradoxical low flow. Hence, the most appropriate and robust end-point for future studies would be occurrence of heart failure and cardiovascular mortality regardless of the type of treatment.

In conclusion, the most important message of these findings is that the presence of a moderately increased transvalvular gradient (<40 mm Hg) or velocity (<4 m/s) does not necessarily exclude the presence of a severe stenosis in patients with preserved LVEF, and that particular attention should be paid to patients with discordant AVA–gradient findings, especially if they are symptomatic.


View Abstract


  • Linked articles 181982.

  • Funding PP holds the Canada Research Chair in Valvular Heart Diseases, Canadian Institutes of Health Research (CHIR), Ottawa, Canada. This work is supported by a research grant from CIHR (MOP#57445).

  • Competing interests None.

  • Provenance and peer review Commissioned; externally peer reviewed.

Linked Articles