Article Text

Download PDFPDF

Echocardiographic severity grading in aortic stenosis: no holy grail, only lessons towards patient individualisation
  1. Hector I Michelena1,
  2. Philippe Pibarot2,
  3. Maurice Enriquez-Sarano1
  1. 1Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota, USA
  2. 2Department of Medicine, Institut Universitaire de Cardiologie et de Pneumologie de Québec, Laval University, Québec, Canada
  1. Correspondence to Dr Hector I Michelena, Division of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, MN 55905, USA; michelena.hector{at}

Statistics from

The term ‘severe aortic stenosis (AS)’ carries a hefty prognostic connotation; it should oblige diligent workup, cautious interval follow-up or intervention on the patient.1 Trained as problem solvers, we aspire to develop a ‘theory of everything’ for diagnosis and management of complex illness, and AS has not escaped our efforts. For example, until recently, a patient with normal left ventricular EF unable to generate a mean gradient (MG) >40 mm Hg (or peak velocity >4 m/s) across a calcified and restricted aortic valve was deemed not to harbour severe AS. We now know that this oversimplification may exclude from potentially life-saving intervention a number of ‘paradoxical low-flow’ patients with substantial AS who despite having a normal EF and MG <40 mm Hg, have the same or worse prognosis as patients who generate the time-honoured ‘cut-off’ gradient.2 Thus, flow-dependent echocardiographic parameters (peak velocity and MG), despite having excellent correlation between them1 and proven prognostic value,3 do not always reflect disease severity. Interestingly, the aortic valve area (AVA), a flow-independent parameter, remains abnormally decreased (<1 cm2) in most of these ‘paradoxical low-flow’ patients, serving as a clue for their diagnosis.4

Paradoxical low-flow is partly to blame for the parameter inconsistencies found in echocardiographic severe-AS grading (defined as MG <40 mm Hg and AVA <1 cm2).5 The prevalence of severe-AS grading inconsistencies has now been studied in over 11 000 patients,1 ,5 and one out of three patients exhibits inconsistent severe-AS grading. However, there are also a significant number of patients without paradoxical low-flow (normal flow) and normal EF who also have ‘discordant’ severe AS1 with MG <40 mm Hg and AVA <1 cm2, and the opposite as well: MG >40 mm Hg and AVA >1 cm2 (less-than-severe AS inconsistent grading). We have shown that for patients with normal EF and normal flow, the left ventricular outflow tract diameter (LVOTd) is associated with significant severe-AS grading inconsistencies,1 such that the current definition of severe AS is most consistent for patients with large LVOTd (≥23 mm), but not so for patients with average (2.0–2.2 mm) or small (1.7–1.9 mm) LVOTds, in whom severe AS AVA cut-offs of 0.8–0.9 cm2 would significantly improve consistency.1 ,5 An intuitive simplification of this AVA conundrum is to index the absolute AVA by the body surface area (BSA), a decent attempt at patient individualisation, which should identify small-sized patients with absolute AVA <1 cm2 who have less-than-severe AS (increase diagnostic specificity) and identify large-sized patients with absolute AVA >1 cm2 who do have severe AS (increase diagnostic sensitivity). Supporting this ‘indexing’ concept is the fact that most of the variability in AVA calculation is derived from the LVOTd since it is squared in the continuity equation, and the LVOTd exhibits good correlation with BSA and height.1 An AVA/BSA index (AVAi) cut-off of 0.6 cm2/m2 has been advocated for this purpose. Surprisingly, we found this partition coefficient to increase grading inconsistency even more, by virtue of reassigning patients to the severe AS category, particularly those with average-sized and small-sized LVOTds,1 in whom we found a cut-off of 0.5 cm2/m2 to be theoretically superior as it corresponded to a peak velocity of 4 m/s and decreased grading discrepancies.

Jander et al6 report the influence of AVAi on the prevalence of severe stenosis in a retrospective consecutive tertiary-referral AS cohort, as well as in a low-risk prospective AS cohort. Expectedly, they found that regardless of the population evaluated, the AVAi reassigns numerous patients to the severe category and increases grading discrepancy, and a partition coefficient of 0.5 cm2/m2 decreases inconsistent grading in both populations.6 They must be commended, however, as they also evaluated the clinical significance of being classified as having severe AS by AVAi in the setting of all other echocardiographic parameters being less-than-severe, which represents an important contribution to our current understanding of AS grading, and constitutes yet another lesson on the limitations of oversimplification and generalisation of single parameters that have not been extensively clinically tested. Within the rigorous independent adjudication of outcomes of a prospective randomised trial (the SEAS trial),7 patients classified as having severe AS exclusively by AVAi were mostly young males with higher BSA whose valvular event-free survival was approximately 60% at 5 years, compared with 40% in patients with severe AS according to AVA (p<0.001), suggesting that despite their AVAi <0.6 cm2/m2, their valve disease was indeed milder. Importantly, they also exhibited a trend towards lower cardiovascular death, a critical outcome when evaluating asymptomatic AS. In addition, the authors confirm, in a low-risk population, that the AVAi offers no incremental outcome-predictive accuracy than that derived from absolute AVA.4 The use of this low-risk population7 has the advantage of biologically isolating the effects of the aortic valve stenosis by virtue of the paucity of associated comorbidities (by trial design); however, it is not representative of current patients presenting for evaluation of significant AS who are older and may harbour significant comorbidities. This is an important limitation of the study that renders the applicability of findings to a younger, healthy population with predominantly mild–moderate AS. The low-risk population studied could also explain the relatively low cardiovascular-death predictive accuracy of both AVA and AVAi (area under curve (AUC) 0.57–0.63).

Echocardiographic AS grading is complex since it relies on a few critical measurements that are technically demanding, subject to error and of variable acquisition methodology across institutions (ie, LVOTd, pulsed and continuous wave Doppler assessments). From these few critical measurements, numerous formula-based severity parameters are then derived. Due to their common origin, these parameters are often interconnected and overlap, such that fundamental errors and/or methodological differences are perpetuated across multiple severity parameters. For example, a recent analysis by Bahlmann et al,8 based on the same prospective SEAS trial data, rendered AVAi a more acceptable prognostic parameter8 than it appears in the current study, and the likely explanation for this contrast is the different echocardiographic stroke volume estimation methodologies used in each study. Thus, different accepted measurement methods for the same parameters may alter their final prognostic accuracy. This is a current standardization that must be addressed by the imaging research community.

Notwithstanding, it remains important to recognise that in large patients with large LVOTds, normal EF and normal flow, inconsistently graded AS is usually characterised by a MG >40 mm Hg and AVA >1 cm2 and the existence of a ‘normalised’ AVA (ie, AVAi) is justified to increase echocardiographic sensitivity. Conversely, in smaller patients with small LVOTs and normal EF and normal flow, inconsistently graded AS is usually characterised by a MG <40 mm Hg and AVA <1 cm2 and the existence of a ‘normalised’ AVA is also justified to increase echocardiographic specificity. However, as discussed above, current observations suggest that correcting AVA exclusively by BSA is likely insufficient or inadequate, and generalising a partition coefficient for all is not the answer. Furthermore, AVAi does not appear to provide incremental prognostic information beyond other parameters.

Lessons learned and emergent inquiries

  1. AVAi must not be used as a single severity parameter. The currently recommended AVAi partition coefficient increases echocardiographic grading inconsistency and may misleadingly classify patients with clinically insignificant AS as having severe disease, particularly in low-risk, younger populations.

  2. The AVAi thresholds (0.5 cm2/m2 vs0.6 cm2/m2) should be prospectively evaluated as outcome predictors in significant AS and their value further categorised by subgroups defined by body size, body mass index and LVOTd. Moreover, other anthropometric ‘correction’ variables for AVA such as height should be considered.

  3. Both AVAi and dimensionless index,1 parameters recommended as reconciliatory when severe AS disparities exist, do not reconcile inconsistent grading, and their cut-offs should be re-evaluated and individualised further.

  4. Systematisation and standardisation of methodologic approaches for fundamental echocardiographic aortic valve measurements must be sought.

  5. Integration of echocardiography with emerging independent tools such as CT valvular-calcium-load assessment9 has the potential of rendering a more comprehensive platform from which to individualise AS severity grading.

  6. We must strive to treat patients, not algorithms and cut-offs. Decision-making in AS must ultimately be guided by the combination of comprehensive imaging, which defines severity; patient functional assessment, which defines the individual interaction between severity and physiology, as well as the patients’ wishes.


View Abstract


  • Contributors All authors contributed to this paper.

  • Competing interests None.

  • Provenance and peer review Commissioned; internally peer reviewed.

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.

Linked Articles