Author + information
- Received September 18, 2011
- Revision received December 5, 2011
- Accepted December 20, 2011
- Published online October 2, 2012.
- Marie-Annick Clavel, DVM, MS,
- Jean G. Dumesnil, MD,
- Romain Capoulade, MS,
- Patrick Mathieu, MD,
- Mario Sénéchal, MD and
- Philippe Pibarot, DVM, PhD⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Philippe Pibarot, Institut Universitaire de Cardiologie et de Pneumologie de Québec, 2725 Chemin Sainte-Foy, Québec G1V-4G5, Canada
Objectives The aim of this case match study was to compare the outcome of patients with paradoxical low-flow (left ventricular ejection fraction [LVEF] ≥50% but stroke volume index <35 ml/m2), low-gradient (mean gradient [MG] <40 mm Hg), a priori severe (aortic valve area [AVA] ≤1.0 cm2) aortic stenosis (AS) (PLG-SAS group) with that of patients with a severe AS (AVA ≤1.0 cm2) and consistent high-gradient (MG ≥40 mm Hg) (HG-SAS group) and with that of patients with a moderate AS (AVA >1.0 cm2 and MG <40 mm Hg) (MAS group).
Background In patients with preserved LVEF, a discordance between the AVA (in the severe range) and the gradient (in the moderate range) raises uncertainty with regard to the actual severity of the stenosis and thus the therapeutic management of the patient.
Methods In a prospective cohort of AS patients with LVEF ≥50%, we identified 187 patients in the PLG-SAS group. These patients were retrospectively matched: 1) according to the gradient, with 187 patients with MAS; and 2) according to the AVA, with 187 patients with HG-SAS.
Results Patients with PLG-SAS had reduced overall survival (1-year: 89 ± 2%; 5-year: 64 ± 4%) compared with patients with HG-SAS (1-year: 96 ± 1%; 5-year: 82 ± 3%) or MAS (1-year: 96 ± 1%; 5-year: 81 ± 3%). After adjustment for other risk factors, patients with PLG-SAS had a 1.71-fold increase in overall mortality and a 2.09-fold increase in cardiovascular mortality compared with the 2 other groups. Aortic valve replacement was significantly associated with improved survival in the HG-SAS group (hazard ratio: 0.18; p = 0.001) and in the PLG-SAS group (hazard ratio: 0.50; p = 0.04) but not in the MAS group.
Conclusions Prognosis of patients with paradoxical low-flow, low-gradient severe AS was definitely worse than those with high-gradient severe AS or those with moderate AS. The finding of a low gradient cannot exclude the presence of a severe stenosis in a patient with a small AVA and preserved LVEF and should mandatorily prompt further investigation.
According to American College of Cardiology/American Heart Association and European Society of Cardiology guidelines for the management of patients with valvular heart disease, only patients having severe aortic stenosis (AS) associated with either symptoms and/or left ventricular ejection fraction (LVEF) <50% or undergoing coronary artery bypass graft surgery or other cardiac surgery have a class I indication for aortic valve replacement (AVR) (1,2). Severe AS is generally defined as an aortic valve area (AVA) ≤1.0 cm2 and a mean transvalvular gradient ≥40 mm Hg. However, the clinician is often confronted with patients having discordant findings (e.g., an AVA = 0.8 cm2 consistent with the presence of a severe AS but a mean gradient [MG] = 30 mm Hg rather indicating the presence of a moderate AS). This situation raises uncertainty with regard to the actual severity of the stenosis as well as the potential indication of AVR if the patient is symptomatic.
We reported that this discordance might be related to the presence of a severe stenosis with concomitant “paradoxical” low flow (i.e., reduced stroke volume and thus transvalvular flow rate despite preserved LVEF) (3,4). The transvalvular pressure gradient is inversely related to the square of AVA and directly related to the square of flow. Hence, a patient with severe AS might nonetheless present with a low gradient if his or her left ventricular (LV) output is reduced, such as is often the case in low-LVEF, low-flow AS, and paradoxical low-flow AS. Paradoxical low-flow, low-gradient AS is found in approximately 10% to 25% of the population with severe AS, and the results of previous studies suggest that this entity often reflects a more advanced stage of the disease (3–7). However, there are very limited data on the outcome of patients with paradoxical low-flow, low-gradient AS and on the potential clinical benefits of AVR in these patients.
Other investigators also reported that the discordant AVA-gradient findings might be due to inconsistency in the guidelines criteria used to grade stenosis severity (8,9). Indeed, when fitting together AVA and gradient data in patients with normal transvalvular flow rates, it seems that the AVA cutoff value of 1.0 cm2 proposed in the guidelines to define severe stenosis corresponds to a value of gradient that is lower (30 to 35 mm Hg) than the 40 mm Hg guidelines criteria (8). And finally, discordance between AVA and gradient might be related to the presence of small body size or errors in the Doppler-echocardiographic measurements. In a recent substudy of the SEAS (Simvastatin and Ezetimibe in Aortic Stenosis) trial, Jander et al. (9) reported that patients with low-gradient (i.e., <40 mm Hg) “severe” (i.e., AVA ≤1.0 cm2) AS and normal LVEF (≥50%) have similar outcome when compared with that in patients with moderate AS.
Therefore the objective of this case match study was to compare the outcome of patients with paradoxical low-flow, low-gradient, a priori severe aortic stenosis (PLG-SAS group) with that of patients with a severe AS and consistent high-gradient severe aortic stenosis (HG-SAS group) and with that of patients with a moderate aortic stenosis (MAS group).
In a cohort of 1,589 consecutive AS patients with at least moderate AS and preserved LVEF (≥50%), we retrospectively identified 223 patients (14%) with paradoxical low-flow (stroke volume index <35 ml/m2), low-gradient (MG <40 mm Hg), a priori severe (AVA ≤1.0 cm2 and indexed AVA ≤0.6 cm2·m−2) AS (Fig. 1). Among this subset, 187 patients (PLG-SAS group) were matched: 1) according to the gradient (±3 mm Hg), with 187 patients with moderate AS defined as an AVA >1.0 cm2 and an indexed AVA >0.6 cm2·m−2 and representing the MAS group; and 2) according to the AVA (±0.05 cm2), with 187 patients with a severe AS documented by an AVA ≤1.0 cm2, an indexed AVA ≤0.6 cm2·m−2, and an MG ≥40 mm Hg representing the HG-SAS group (Fig. 1).
Clinical data included age, sex, documented diagnosis of hypertension, hypercholesterolemia, diabetes, obesity (body mass index ≥30 kg·m−2), and coronary artery disease (3).
Patients were classified as having no cardiac symptoms, mild symptoms (dyspnea [New York Heart Association functional class II], angina [Canadian Cardiovascular Society class I or II], and/or fatigue), or moderate/severe symptoms (dyspnea [New York Heart Association functional class III or IV], angina [Canadian Cardiovascular Society class III or IV], and/or syncope).
The Doppler-echocardiographic measurements included the LV dimensions, the LVEF calculated by the Simpson method, the peak and mean transvalvular pressure gradients obtained with the use of the modified Bernoulli equation, and the AVA obtained with the use of the standard continuity equation. We paid particular attention to the search for the highest peak transvalvular velocity with the use of multi-window continuous-wave Doppler interrogation. The Doppler-echocardiographic measurement of LV outflow tract stroke volume was corroborated by the 2-dimensional volumetric or Teicholz method. As a measure of global LV hemodynamic load, we calculated the valvulo-arterial impedance: Zva = (SAP + MG)/SVI, where SAP is the systolic arterial pressure, MG the mean transvalvular pressure gradient, and SVI the stroke volume indexed to body surface area (10).
The primary endpoints for this study were overall and cardiovascular mortality, regardless of whether or not there was AVR. Hence, this includes the deaths occurring in patients who did not undergo AVR as well as those occurring after operation in patients who underwent AVR. The outcome data were retrospectively obtained from the charts or death certificates of patients. The secondary endpoints were AVR and the combined endpoint of AVR and overall death. Figure 1 shows the reasons for therapeutic decision making as reported in the patient's chart.
Results are expressed as mean ± SD or percentages. For continuous variables, differences between groups were analyzed with the use of 1-way analysis of variance for repeated measures followed by the Tukey's post hoc test for inter-group comparisons. The Cochran-Mantel-Haenszel test was used to compare categorical variables as appropriate.
Cumulative survival was estimated with the stratified Kaplan-Meier method and compared between groups with a log-rank test. The effect of the clinical and Doppler-echocardiographic variables on survival and event-free survival was assessed with the use of stratified Cox proportional hazard models. The impact of AVR during follow-up was tested with AVR as a time-dependent covariate in the stratified Cox proportional hazards model for overall and cardiovascular survival. Age, sex and clinically relevant variables with a p value <0.05 on univariable analysis were incorporated into the multivariable models.
A propensity score representing the probability of having AVR as opposed to conservative therapy was calculated for each patient with a logistic regression analysis that identified variables independently associated with the type of procedure. The calculated propensity score was then incorporated into subsequent multivariable regression models. Data were analyzed by SPSS for Windows (version 19.0, SPSS, Inc., Chicago, Illinois).
Compared with patients in HG-SAS and MAS groups, patients with PLG-SAS were older and had a higher proportion of women; higher prevalence of coronary artery disease and hypertension; higher valvulo-arterial impedance; faster heart rate; and lower body surface area, LVEF, LV end-diastolic volume, stroke volume, and transvalvular flow rate (Table 1). They also had higher prevalence of diabetes compared with the HG-SAS group. The prevalence of moderate/severe symptoms was similar in the PLG-SAS (56%) and HG-SAS (58%) groups but lower in the MAS group (32%).
During a mean follow-up of 4.2 ± 2.4 years, there were 307 AVRs (Fig. 1) and 127 deaths, of which 82 were cardiovascular-related. Among the patients who underwent AVR, the proportion of concomitant coronary artery bypass graft surgery was similar in the 3 groups: 39% in the HG-SAS group, 53% in the PLG-SAS group, and 43% in the MAS group (p = 0.13). Combined (AVR or death) event-free survival rates at 1- and 5-year follow-up were 63 ± 4% and 24 ± 4%, respectively, for PLG-SAS group versus 30 ± 3% and 9 ± 3% for the HG-SAS group and 85 ± 3% and 41 ± 4% for the MAS group (p < 0.0001) (Fig. 2A). Patients with PLG-SAS had lower incidence of AVR (1-year: 29 ± 3%; 5-year: 55 ± 5%) compared with HG-SAS patients (1-year: 69 ± 3%; 5-year: 85 ± 3%) but higher incidence compared with MAS patients (1-year: 12 ± 2%; 5-year: 47 ± 5%) (Fig. 2B). Patients with PLG-SAS had reduced overall survival (1-year: 89 ± 2%; 5-year: 64 ± 4%) (Fig. 2C) and cardiovascular survival (1-year: 91 ± 2%; 5-year: 74 ± 4%) (Fig. 2D) compared with patients with HG-SAS (overall: 1-year: 96 ± 1%; 5-year: 82 ± 3%; cardiovascular: 1-year: 97 ± 1%; 5-year: 85 ± 3%) or MAS (overall: 1-year: 96 ± 1%; 5-year: 81 ± 3%; cardiovascular: 1-year: 98 ± 1%; 5-year: 91 ± 2%).
Predictors of outcomes
In multivariable analysis, the factors associated with increased risk of combined events (AVR or death) were: higher MG (p < 0.0001), coronary artery disease (p = 0.001), moderate/severe symptoms (p = 0.02), obesity (p = 0.002), female sex (p = 0.04), and higher valvulo-arterial impedance (p = 0.004) (Table 2).
The factors independently associated with increased overall mortality were older age (p < 0.0001), conservative treatment (p < 0.0001), lower LVEF (p = 0.03), and being in the PLG-SAS group (p = 0.02) (Table 3). After adjustment for age, sex, symptomatic status, coronary artery disease, diabetes, type of treatment, and LVEF, patients with PLG-SAS had a 1.88-fold increase in overall mortality (Table 3) and 2.87-fold increase in cardiovascular mortality (Table 4) compared with patients with MAS. When compared with HG-SAS and MAS patients pooled together, PLG-SAS patients had a 1.71-fold (95% confidence interval [CI]: 1.07 to 2.71; p = 0.02) and 2.09-fold (95% CI: 1.16 to 3.78; p = 0.01) increase in overall and cardiovascular mortality, respectively.
Effect of type of treatment
Aortic valve replacement was associated with significantly better survival in the 3 groups (Fig. 3). The magnitude of the protective effect of AVR seemed to be more important in the HG-SAS and PLG-SAS groups compared with the MAS group. Age, LVEF, body surface area, peak aortic jet velocity, AVA, coronary artery disease, and presence of moderate/severe symptoms were independently associated with the type of treatment and were used for the calculation of the propensity score (Hosmer-Lemeshow Statistic: p = 0.997). After adjusting for age and propensity score, AVR was associated with improved survival in the whole cohort (hazard ratio [HR]: 0.35; 95% CI: 0.21 to 0.56, p < 0.0001) as well as in the HG-SAS group (HR: 0.18; 95% CI: 0.07 to 0.48, p = 0.001) and PLG-SAS group (HR: 0. 50; 95% CI: 0.25 to 0.99, p = 0.04), and there was also a trend for an association between AVR and improved survival in MAS group (HR: 0.44; 95% CI: 0.17 to 1.12, p = 0.09).
The main findings of this study are that patients with PLG-SAS have worse prognosis compared with patients with MAS as well as those with HG-SAS, and the outcome of these patients with PLG-SAS is improved by AVR. Compared with patients with MAS or HG-SAS, patients with PLG-SAS had a higher prevalence of female sex and hypertension and had smaller LV cavities and markedly higher valvulo-arterial impedance. These typical features are consistent with the presence of paradoxical low-flow severe AS, a clinical entity that we initially described in 2007 (3). Several other studies have subsequently confirmed that paradoxical low-flow AS often corresponds to a more advanced stage of the disease as reflected by more severe intrinsic myocardial damage and the appearance of a restrictive physiology, which in turn contributes to the low-flow state in these patients despite the presence of a preserved LVEF (5–7,11,12). Hence, the worse outcome observed in patients with PLG-SAS is likely due to the combination of more advanced age, more frequent comorbidities such as hypertension, worse intrinsic myocardial damage, and lesser referral to surgery. With regard to this latter aspect, it should be emphasized that the presence of a low gradient in a patient with preserved LVEF might lead the treating physician to conclude that the stenosis is not severe even if the AVA is <1.0 cm2. Hence, among 2 patients having a small AVA, a preserved LVEF, and moderate cardiac symptoms, it is likely that the one with a high gradient (e.g., 50 mm Hg) would promptly be referred to surgery, whereas there would be much questioning with regard to the other patient with a moderate gradient (e.g., 30 mm Hg). Current practice guidelines indeed put a lot of emphasis on the gradient (or the peak aortic jet velocity) in the decision to refer a patient to AVR. Accordingly, in the present study, patients with small AVA and low gradient (i.e., with PLG-SAS) were less often referred to surgery compared with those with small AVA and high-gradient (i.e., HG-SAS), although they had similar AVA and prevalence of moderate/severe symptoms (Table 1, Fig. 1). However, the patients with PLG-SAS had significantly reduced overall and cardiovascular survival even after adjustment for the differences in the baseline risk factors. Furthermore, the performance of AVR was highly protective in this subset of patients even after adjustment for propensity score. Hence, the PLG-SAS entity is often misdiagnosed, which leads to underestimation of stenosis severity and symptoms and therefore under-use or inappropriate delay of AVR.
A previous paradigm was that normal LVEF implies normal stroke volume and transvalvular flow rate. However, it is now well-established that an important proportion of patients with cardiovascular diseases might have a reduced LV pump function despite preserved LVEF. This phenomenon is related to pronounced LV concentric remodeling with reduced cavity size, increased myocardial fibrosis, impaired diastolic filling, increased afterload, and altered myocardial systolic function unrevealed by LVEF. This low-flow state condition might considerably complicate the assessment of stenosis severity in patients with AS, because the main stenotic index used for clinical decision making, namely the gradient, is directly related to the squared function of transvalvular flow rate. Hence, even a modest decrease in flow rate could yield to important reduction in gradient and thus to underestimation of stenosis severity. In the present study, patients with PLG-SAS had a much lower transvalvular flow rate when compared with the 2 other groups. The transvalvular gradient and peak aortic jet velocity have a high specificity to identify severe AS. However, because patients with severe AS often have a reduced stroke volume and transvalvular flow rate irrespective of LVEF, these parameters might be quite lower than expected in an important proportion of patients, thus resulting in relatively low sensitivity. These findings raise the importance of considering other Doppler-echocardiographic parameters and eventually other diagnostic tests when confronted with discordance between AVA (in the severe range) and gradient (in the moderate range).
In the present study, we elected to include both asymptomatic and symptomatic patients, because this approach better reflects the clinical spectrum of the disease. Moreover, the most challenging patient subset from both a diagnostic and a therapeutic standpoint is precisely the one with discordant echocardiographic findings and symptoms. In this subset, it is crucial to confirm stenosis severity to select the most appropriate treatment: AVR versus conservative therapy. The presence and severity of symptoms is difficult to assess and interpret in elderly patients, which represent the majority of the AS population today. In the present study, moderate-to-severe symptoms were powerful predictors of the composite of AVR or death but were not significantly associated with overall mortality. Thus these findings emphasize the point that the main parameters (i.e., MG, peak jet velocity, symptoms, and LVEF) on which therapeutic decisions are predominantly based in current practice have important limitations and that some patients might nonetheless have severe AS despite the presence of low gradient and preserved LVEF.
The primary goal of therapeutic management in AS is to improve longevity and quality of life. Accordingly, we selected overall and cardiovascular mortality as the primary endpoints for this study. Previous studies have generally used the composite of AVR or death as the primary and often the sole endpoint, and when such is the case, this composite endpoint is in very large part driven by AVR (9). However, the limitation of this endpoint is that it is essentially determined by the perception of disease severity by the clinician, which is in turn highly influenced by the magnitude of the gradient (or peak aortic jet velocity) and by the presence of symptoms. Hence, it is not surprising that these parameters are often found to be powerful predictors of this composite endpoint (i.e., they are the reasons why the cardiologist refers the patient to AVR). The findings of the present study underline the importance of not only reporting the composite of AVR or death but also the overall and cardiovascular mortality, regardless of the type of treatment. As a matter of fact, patients with PLG-SAS had lower incidence of the composite of AVR or death when compared with HG-SAS (Fig. 2A), but the reverse was observed when analyzing overall and cardiovascular mortality (Figs. 2C and 2D). Indeed, survival was markedly lower in the PLG-SAS group, whereas it was similar in the MAS and HG-SAS groups. And this difference persisted after adjustment for other risk factors.
At first glance, the results of the present study are discordant with those of the recent study by Jander et al. (9) reporting that patients with low-gradient severe (i.e., small AVA) AS and preserved LVEF have an outcome similar to that in patients with moderate AS. However, the results of Jander et al. do not correspond to the PLG-SAS entity that we specifically examined in this study. Indeed, the low-gradient severe AS patients reported by Jander et al. (9) do not exhibit the features typically observed in patients with paradoxical low-flow AS, and as recently discussed (13), the finding of low-gradient severe AS in these cases could rather be due to 1 or more of the following: 1) small body size; 2) measurement errors; and 3) inconsistent grading due to intrinsic discrepancies in guidelines criteria. This interpretation is further comforted by the results of a previous substudy from the SEAS trial, whereby, in the same cohort of patients, Cramariuc et al. (11) identified 100 low-flow patients with the stroke volume measured by volumetric method rather than the 223 reported by Jander et al. (9), and these patients also exhibited the restrictive features usually associated with paradoxical low-flow AS.
The findings of this study emphasize the importance of not systematically denying surgery to a symptomatic patient with small AVA and low gradient. Indeed, the presence of these discordant findings in a patient with preserved LVEF should not be automatically equated with a moderate AS but rather be conducive to a more comprehensive Doppler-echocardiographic evaluation and potentially to the performance of other diagnostic modalities if the results of the echocardiographic evaluation remain inconclusive with regard to stenosis severity. Hence, the first steps when confronted with discordance between AVA and gradient should be: 1) to rule out potential errors in the measurement of stroke volume, AVA, and/or gradients as emphasized in the recent American Society of Echocardiography/European Association of Echocardiography recommendations (14); 2) to rule out the confounding effect of small body size by calculating the indexed AVA, a value >0.6 cm2/m2 being indicative of the presence of moderate AS; and 3) to establish whether the global hemodynamic load is severely increased (i.e., valvulo-arterial impedance >4.5 mm Hg·ml−1·m2), as is often the case with PLF-SAS (15). The next step could then be to assess the symptomatic status of the patient and, if the patient claims to be asymptomatic, to perform an exercise test. Corroborating methods including dobutamine stress echocardiography, valve calcium quantification by computed tomography, and dosage of plasma B-type natriuretic peptide can also be used to confirm the presence of a severe stenosis as well as to determine the indication for surgery.
Owing to their particular LV geometry/function pattern (i.e., small LV cavity with pronounced concentric remodeling and restrictive physiology, and often small LV outflow tract and aortic annulus), patients with PLG-SAS might have increased risk of perioperative mortality/morbidity (Fig. 3) and prosthesis-patient mismatch after surgical AVR. Thus further studies are needed to determine the potential value of transcatheter aortic valve implantation in this particular subset of patients.
Of note, patients with moderate AS also had a relatively high event rate. This finding is in agreement with several recent studies (16,17) suggesting that moderate AS is not a benign disease. Hence, closer follow-up should be considered in these patients with moderate AS, and the 2-year interval proposed in the guidelines is probably too long.
The main limitation of this study is its retrospective design. The baseline data were prospectively collected in consecutive patients with AS referred to the echocardiographic laboratory. However, the outcome data were retrospectively obtained from the charts or death certificates of patients. Some variables previously reported as being risk factor or risk of disease progression (i.e., LV longitudinal strain, left atrial volume, B-type natriuretic peptide, and valve calcification) were not measured in this study, thus limiting the characterization of the different study groups.
Prognosis of PLG-SAS patients was worse than those with high-gradient severe AS and those with moderate AS. Furthermore, the outcome of these patients with PLG-SAS was improved by surgical treatment. Hence, the finding of a low gradient should not exclude the presence of a severe stenosis in a patient with a small AVA and preserved LVEF. When confronted with this situation, a more comprehensive Doppler-echocardiographic evaluation as well as other diagnostic tests should be used to corroborate the stenosis severity and guide therapeutic management.
The authors thank Jocelyn Beauchemin and Dominique Labrèche for the data collection and their technical assistance.
This work was supported by a grant (MOP 57745) from the Canadian Institutes of Health Research (CIHR), Ottawa, Canada. Dr. Pibarot holds the Canada Research Chair in Valvular Heart Diseases, CIHR. Dr. Mathieu is a research scholar from the Fonds de Recherches en Santé du Québec, Montreal, Canada. Dr. Clavel holds a Vanier Canada Graduate Scholarship, CIHR. All other authors have reported they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic stenosis
- aortic valve area
- aortic valve replacement
- confidence interval
- HG-SAS group
- patients with high-gradient severe aortic stenosis
- hazard ratio
- left ventricular
- left ventricular ejection fraction
- MAS group
- patients with moderate aortic stenosis
- mean gradient
- PLG-SAS group
- patients with paradoxical low-flow, low-gradient, a priori severe aortic stenosis
- Received September 18, 2011.
- Revision received December 5, 2011.
- Accepted December 20, 2011.
- American College of Cardiology Foundation
- Bonow R.O.,
- Carabello B.A.,
- Kanu C.,
- et al.
- Vahanian A.,
- Baumgartner H.,
- Bax J.,
- et al.
- Hachicha Z.,
- Dumesnil J.G.,
- Bogaty P.,
- Pibarot P.
- Dumesnil J.G.,
- Pibarot P.,
- Carabello B.
- Herrmann S.,
- Stork S.,
- Niemann M.,
- et al.
- Minners J.,
- Allgeier M.,
- Gohlke-Baerwolf C.,
- Kienzle R.P.,
- Neumann F.J.,
- Jander N.
- Jander N.,
- Minners J.,
- Holme I.,
- et al.
- Briand M.,
- Dumesnil J.G.,
- Kadem L.,
- et al.
- Dumesnil J.G.,
- Pibarot P.
- Hachicha Z.,
- Dumesnil J.G.,
- Pibarot P.
- Rosenhek R.,
- Klaar U.,
- Schemper M.,
- et al.
- Chan K.L.,
- Teo K.,
- Dumesnil J.G.,
- Ni A.,
- Tam J.