Low-Gradient Aortic Valve Stenosis

Myocardial Fibrosis and Its Influence on Function and Outcome

Study population

In total, 130 consecutive patients with moderate to severe AS (referred from general practice or specialist outpatient clinics) were screened, and 86 patients were included in the final dataset. Patients were included if they had either isolated severe AS with symptoms on exertion and an aortic valve area (AVA) <1.0 cm² or moderate AS defined as an AVA ≤1.5 and ≥1.0 cm². The reasons for exclusion were a history of myocardial infarction as well as significant coronary artery disease (degree of vessel stenosis ≥50%), uncured cancer, and other moderate or severe valvular dysfunction. Five patients were excluded solely on the basis of heart catheterization showing a vessel stenosis ≥50%. Some of the data have already been analyzed previously by studying the impact of fibrosis in high-gradient AS (8). All patients with severe AS were referred for surgery.

Study protocol

In the context of pre-operative management, all patients with severe AS underwent cardiac heart catheterization and cardiac magnetic resonance imaging (cMRI) with late-enhancement (LE) imaging for replacement fibrosis within 2 weeks. In patients with a reduced EF (n = 6), low-dose dobutamine stress echocardiography was performed to confirm true severe AS. Within the next 3 weeks, AVR was performed in all patients with severe AS, and 2 endomyocardial biopsy samples were taken from the endocardium of the basal LV septum (for interstitial fibrosis). Nine months after the baseline assessment, a clinical follow-up examination including conventional echocardiography and strain-rate imaging was performed in all patients. In addition, in the group of patients who had undergone AVR, follow-up cMRI was performed. The study was reviewed and approved by the local ethics committee of University Hospital Würzburg and was conducted according to the principles outlined in the Declaration of Helsinki. Informed consent was obtained from all patients before any investigation.

Heart catheterization

All patients with severe AS underwent standard heart catheterization before enrollment in the study. LV pressure was recorded using fluid-filled pigtail catheters after retrograde passage of the aortic valve. Using a venous approach, right heart catheterization was done with a Swan-Ganz catheter.

Standard echocardiography

Transthoracic echocardiography was performed using the Vivid 7 System (GE Vingmed Ultrasound, Horten, Norway) with a 3.5-MHz transducer. A standard echocardiographic study for systolic and diastolic parameters was done. Using established parameters for the assessment of AS severity, patients were characterized and grouped by AVA, transvalvular mean gradients and LVEF (9,10).

All patients were categorized into 1 of 4 groups: group 1, moderate AS (n = 17), AVA ≥1.0 cm²; group 2, severe AS/high gradient (n = 49), AVA <1.0 cm² + gradient ≥40 mm Hg; group 3, severe AS/low-gradient/preserved EF (n = 11), AVA <1.0 cm² + gradient <40 mm Hg + EF ≥50%; group 4, severe AS/low-gradient/reduced EF (n = 9), AVA <1.0 cm² + gradient <40 mm Hg + EF <50%.

LV mass was computed by the Devereux modified cube formula and normalized for body surface area to obtain the LV mass index. LV hypertrophy was defined as LV mass index >115 g/m² in men and >95 g/m² in women (11). In addition, LV wall stress (with the modified La Place formula, relative wall thickness (twice the diastolic posterior wall thickness divided by LV end-diastolic diameter), mid-wall fractional shortening, valvuloarterial impedance (Z_va = SAP + MG/SVI, where SAP is the systolic arterial pressure, MG is the mean gradient, and SVI is the stroke volume index), stroke volume, systemic vascular and valvular resistance, and arterial compliance were assessed (1). Valvular calcium score was assessed in the parasternal short axis and scaled as described by Rosenhek et al. (12).

Systolic mitral ring displacement was measured at the septal and lateral sides using M-mode echocardiography in an apical 4-chamber view. For the final dataset, the average of these 2 measurements was used. The reproducibility of this marker was examined by triple (intraobserver) and double measurement (interobserver) of 10 different patients, yielding coefficients of variation of 5.5% and 6.3% for intraobserver and interobserver reproducibility, respectively.

Strain rate imaging

Tissue Doppler imaging was done immediately after conventional echocardiography and post-processed with dedicated software (Echopac, GE Vingmed Ultrasound). Radial function was assessed by peak systolic strain rate (related to regional contractility) scanning the posterior wall from parasternal views. For longitudinal function, this parameter was extracted from the mid-ventricular segment of the septum and lateral wall using an apical 4-chamber view.

Cardiac magnetic resonance imaging

In patients with severe AS and without contraindications (n = 46), cine cMRI and LE imaging were performed using a 1.5-T magnetic resonance scanner to determine myocardial replacement fibrosis. To detect LE, short-axis images were acquired 10 to 15 min after injection of gadopentetate dimeglumine (0.2 mmol/kg of body weight, Magnevist, Schering, Berlin, Germany) using an inversion recovery sequence (field of view, 240 × 320 mm²; matrix, 165 × 256). Each of the 17 LV segments was assessed for the presence or absence of LE (i.e., replacement fibrosis) by an experienced radiologist who was blinded to patient outcome, degree of fibrosis on endomyocardial biopsy samples, and time sequence of scans.

Grading of endomyocardial biopsy samples

Biopsy samples were taken during AVR from the basal septum. The percentage of the area of fibrosis was calculated by an experienced pathologist and used as the index of the degree of fibrosis (8). The percentage of area of fibrosis in the section was obtained by dividing the sum of the fibrotic areas of the section by the total tissue area, as described by Tanaka et al. (13). In addition, fiber diameters were measured (8).

Data analysis

Data are presented as mean ± SD or median (quartiles) or frequency (percentage), as appropriate. The chi-square test, Fisher exact test, or Mann-Whitney U test was used to compare categories, as appropriate. Statistical testing was done for purely descriptive reasons; hence, no adjustment for multiple testing was implemented. Correlations were computed using Spearman's correlation coefficient. p value for trend was computed using analysis of variance after checking the assumption of homogeneity of variances. Two sets of logistic regression analyses were run to investigate 2 dependent variables: fibrosis (yes vs. no) and no improvement in New York Heart Association (NYHA) functional class or death (yes vs. no). Fibrosis was considered present if the fibrosis score was >2%. Functional improvement was accepted if a patient had improved at least 1 NYHA functional class at follow-up. For both sets, the diagnostic/prognostic performance of biomarkers (procollagen type III amino-terminal peptide [PIIINP], N-terminal pro–B-type natriuretic peptide [NT–proBNP], and mitral ring displacement) was qualitatively compared and the odds ratio (OR) with its 95% confidence interval (CI), Wald index, and the area under the receiver-operator characteristic curve (AUROC) (with its SE) were reported. NT-proBNP entered analyses as quartiles, and a linear trend across categories was reported. These analyses were performed with fixed adjustment for sex and age. Further, in a multivariable logistic regression approach, independent correlates/predictors were sought by backward selection using the likelihood ratio criterion (p_in = 0.05, p_out = 0.1). Because of the small sample size, only the following variables were tested: EF, stroke volume, strain rate, mitral ring displacement, AVA, NT-proBNP, PIIINP, and LE. All tests were performed 2-sided. SPSS version 17.0.1 (SPSS Inc., Chicago, Illinois) was used.

Baseline echocardiography

Tables 2 and 3⇓⇓ show that AVA, valvular resistance, and LV hypertrophy were similar among groups with severe AS, but qualitatively worse compared with moderate AS patients. An inverse relationship with fibrosis (both interstitial and replacement fibrosis) was found for stroke volume, longitudinal strain rate, and mitral ring displacement. Valvuloarterial impedance, wall stress, and E/E′ increased across groups 1 through 4. Mitral ring displacement was closely inversely correlated with the tissue fibrosis score (measured in percentage per micrometer) in biopsy samples and with valvuloarterial impedance (Fig. 1) but not with EF (r = −0.25, p = 0.05). Furthermore, a significant but weak correlation was found between valvuloarterial impedance and tissue fibrosis score in biopsy samples (r = 0.29, p < 0.05). Of note, mitral ring displacement accurately distinguished between groups 1 and 3 using a cutoff value of 9 mm (Fig. 2). By contrast, mean gradient or EF showed overlapping values. Figure 3 demonstrates an example of a patient with severe AS/low-gradient/preserved EF, and a characteristic diagnostic pattern.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Correlations Between Myocardial Fibrosis and Functional Parameters

(A) Inverse correlation between biopsy-based assessment of myocardial fibrosis (expressed as the percentage of area of fibrosis within the respective section) and mitral ring displacement assessed by M-mode echocardiography in the 4 groups (r = −0.79, p < 0.0001). (B) Inverse correlation between valvuloarterial impedance and mitral ring displacement in the 4 groups (r = −0.6; p < 0.001). AS = aortic stenosis; EF = ejection fraction.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Scatterplot of Mean Transvalvular Gradient (mm Hg) Versus Mitral Ring Displacement (mm)

Note that a clear differentiation between moderate AS and severe AS with low gradients is possible by the assessment of mitral ring displacement. Abbreviations as in Figure 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3

Patient With a Typical Pattern of Low-Gradient Aortic Stenosis

A 78-year-old woman from group 3 (ejection fraction >50%) with septal and posterior wall of 13 mm, left ventricular end-diastolic diameter of 40 mm, ejection fraction 60%, aortic valve area = 0.9 cm², mean gradient = 29 mm Hg. Mitral ring displacement is 6 mm (normal value >9 mm). Typical basal septal and lateral late-enhancement on cardiac magnetic resonance imaging. Advanced myocardial fibrosis score is 5.0% and myocyte diameter is 13 μm. AV maxPG = peak aortic gradient (mm Hg); AV meanPG = mean aortic gradient (mm Hg); AV Vmax = peak aortic velocity (m/s); AV Vmean = mean aortic velocity (m/s).

Endomyocardial biopsy samples and markers of collagen metabolism

Myocardial biopsy samples showed a significantly higher amount of interstitial fibrosis in groups 3 and 4 compared with group 2 (both p < 0.001). Similar findings were recorded as well for levels of NT-proBNP and PIIINP, a collagen III degradation product (Table 3). The tissue fibrosis score was closely correlated with PIIINP and NT-proBNP (r = 0.63 and r = 0.69, respectively; both p < 0.01).

Baseline cMRI

Baseline cMRI was not feasible in 23 subjects due to claustrophobia (n = 18) or implanted devices (n = 5). The distribution of LE-positive segments in the 3 groups is shown in Table 3. In patients in groups 3 and 4, 100% had 1 or more LE-positive segment. By contrast, in group 2, 47% (n = 23) had no LE-positive segment. LE was mainly observed in the subendocardial layers of the basal segments. Mitral ring displacement decreased gradually with the increasing number of LE-positive segments on cMRI (p < 0.001 for trend).

Follow-up

All patients with severe AS underwent AVR (n = 49 bioprosthesis; n = 20 mechanical valve). Six patients had cardiac death within the first 30 days (group 2: n = 1 [2%]; group 3: n = 2 [18%]; group 4: n = 3 [33%]). In patients with cMRI at follow-up (n = 40), no appreciable change in LE-positive segments 9 months after AVR in all groups occurred. Echocardiographic follow-up data are shown in Table 4 and Figure 4.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4

Baseline and 9-Month Follow-Up of Mitral Ring Displacement Data According to the 4 Groups Studied

Patients with moderate AS showed higher values both at baseline and follow-up compared with patients with low-gradient AS, irrespective of ejection fraction. The potential for recovery of mitral ring displacement was best in patients with severe AS and a high gradient. p < 0.05: *versus moderate AS; †versus severe AS/high gradient; ‡versus baseline. Abbreviations as in Figure 1.

Clinical outcomes in relation to cardiac performance

The NYHA functional class for each group at baseline and at follow-up is shown in Tables 1 and 4. In regression analyses with fixed adjustment for age and sex, NT-proBNP (OR: 5.9 per quartile; 95% CI: 2.5 to 13.5; Wald: 17.0; AUROC: 0.89 ± 0.04; p < 0.001) and mitral ring displacement (OR per millimeter: 0.37; 95% CI: 0.23 to 0.59; Wald: 17.4; AUROC: 0.93 ± 0.03; p < 0.0001) were the only markers associated with fibrosis. Of note, the AUROC for age and sex alone to predict fibrosis was only 0.61 ± 0.07. Both markers, NT-proBNP and mitral ring displacement, also predicted improvement in NYHA functional class after AVR; for NT-proBNP, OR per quartile: 12.2; 95% CI: 3.8 to 38.9; Wald: 17.8; AUROC: 0.92 ± 0.04; p < 0.001 and for mitral ring displacement: OR per millimeter: 0.08; 95% CI: 0.01 to 0.49; Wald: 7.3; AUROC: 0.99 ± 0.01; p < 0.007). By contrast, the AUROC for age and sex alone to predict outcome was 0.73 ± 0.06. In multivariable analyses, only mitral ring displacement was retained in models for fibrosis (OR: 0.37, 95% CI: 0.23 to 0.58, Wald: 18.9; p < 0.0001) and functional improvement (OR: 0.09, 95% CI: 0.02 to 0.41, Wald: 9.7; p = 0.002).

Impact of myocardial fibrosis on the gradient

In an experimental setup, Derumeaux et al. (16) proved that an increased afterload induced by aortic banding causes hypertrophy and myocardial fibrosis, which both lead to reduced myocardial deformation. These changes in LV tissue structure were closely related to myocardial deformation, whereas the peak systolic strain rate correlated best with the transmural extent of myocardial fibrosis (16). Consistent with previous investigations, the current study showed that myocardial fibrosis is a feature of more advanced disease. In this disease stage, a high level of persistently elevated systolic wall stress and compromised myocardial perfusion is present (17). As soon as these intrinsic and extrinsic stressors to the myocardium cannot be compensated any longer by hypertrophy, they cause myocardial fibrosis. It is predominantly located at the subendocardial layers, leading first to interstitial and later to replacement fibrosis. The current study demonstrates by the use of endomyocardial biopsies (the reference standard for the assessment of myocardial fibrosis) that especially patients with a low transvalvular gradient showed a significantly higher degree of myocardial fibrosis. In addition, replacement fibrosis as a sign for more advanced fibrosis could be visualized by noninvasive LE imaging in most of these patients with low transvalvular gradients.

It can be assumed that both the decreased longitudinal function due to fibrotic noncontractile tissue at the subendocardium and a small LV cavity size induced by hypertrophy lead to a reduced stroke volume and, thus, a low transvalvular pressure gradient. Accordingly, the group of patients with a low gradient and a normal EF displayed the smallest LV cavity size and the highest relative wall thickness. In addition, global hemodynamics and valve resistance parameters were altered, as previously reported (18). Thus, stroke volume was reduced and valvular resistance (a parameter of valve stiffness) was significantly increased compared with the high-gradient group. Additionally, in patients with a low gradient, valvular arterial impedance was increased because these patients have both vascular and valvular damage, which indicates a more advanced stage in disease progression (18). Consistently, serum biomarkers for the degradation of collagen (PIIINP) and for myocardial stress (NT-proBNP) were highest in the group of patients with a low gradient. The latter finding is in good accordance with the TOPAS (Truly Or Pseudo-severe Aortic Stenosis) study and a recently published study with a new risk score in AS, in which patients with lower values of natriuretic peptides had a better survival and lower mortality risk after AVR (19,20).

With respect to the clinical outcome, our findings confirm published data demonstrating that patients with low-gradient AS are more likely to experience a worse outcome after AVR (21). Of note, 5 of 6 patients who died during follow-up were low-gradient patients, whereas 2 had preserved EF and 3 had reduced EF. One patient from the group with severe AS/high gradient died of cardiac arrest. She was a patient with a high fibrosis tissue score and high serum levels of PIIINP and NT-proBNP. The relatively high mortality might be due to the fact that most patients were severely symptomatic and patients were in an advanced stage of the disease.

LV function and stenosis severity

Because of the pronounced concentric remodeling, small LV cavities, impaired LV filling, and impaired longitudinal shortening, some AS patients may have decreased pump function and thus a low stroke volume (being also associated with a lower central venous oxygen saturation), low flow rate, and a low transvalvular gradient despite preserved LVEF (1,15). Hence, this “paradoxical” low-flow, low-gradient AS would be compatible with a severe AS and more advanced impairment of myocardial function. Other investigators, however, suggested that this disease pattern is most often due to inconsistency in the criteria of AVA and gradients proposed in the guidelines for AS grading and that it generally reflects moderate or moderate to severe AS (15). The results of the present study provide important new insight in this regard and confirm that a substantial proportion of patients with severe AS may present with low-gradient AS despite the presence of preserved LVEF.

Current guidelines for the assessment of patients with AS mention EF as the only established functional parameter both for the evaluation of the stenosis severity and estimation of prognosis (9,10). However, EF seems to have very little clinical relevance for patients with a low transvalvular gradient and preserved LV function (i.e., EF >50%) (3,4,22). This group of patients had severe myocardial fibrosis and a poor clinical outcome after AVR, and 2 of them even died during follow-up. EF is mainly determined by radial function, which can be compensated for a long time, even in the presence of subendocardial fibrosis (23). It has been shown that EF will decrease at the very advanced disease stages, when radial and longitudinal function is decreased, as present in our last group of patients with a low-gradient AS and decreased EF (8,19).

In contrast, longitudinal myocardial kinetics as measured by, for example, longitudinal mitral ring displacement or tissue Doppler strain or strain rate, is more sensitive for the detection of subtle LV dysfunction. It is related to subendocardial fibers where wall stress is highest and fibrosis starts in AS. Thus, patients with moderate AS had the best radial and longitudinal function. Patients with a high-gradient AS had normal radial and only slightly decreased longitudinal values. In these patients, the stroke volume was normal and, thus, the left ventricle could generate a high transvalvular gradient. In patients with a low gradient but preserved EF, radial function was almost normal (responsible for the normal EF), but longitudinal function was markedly decreased (responsible for the decreased stroke volume). Only in patients with severe AS and EF <50% were all functional parameters reduced (EF + radial function + longitudinal function). Obviously, AVR should be performed before the development of this stage of disease progression, because the potential for recovery of mitral ring displacement was best in patients with severe AS and a high gradient (Fig. 4). However, although these patients with low-gradient AS showed no significant improvement in NYHA functional class, it is known that they have a survival benefit after AVR and should not be treated conservatively (2).

Clinical impact

Clinical management of patients with AS is mainly based on the assessment of valvular parameters, EF, and symptoms (24,25). The current study suggests that it is also important to carefully assess longitudinal myocardial kinetics and LV hemodynamics to evaluate the stage of the disease and to estimate short-term prognosis. Thus, as shown also by others (4,7,8,12,26,27), the presence of low peak velocity or low peak/mean gradients combined with preserved EF does not necessarily indicate the absence of severe stenosis. A more comprehensive evaluation is required in patients with a small AVA and low gradient.

In the clinical setting in patients with poor echocardiographic image quality, it is sometimes difficult to distinguish between a moderate AS (i.e., low gradient because of no severe valve abnormality) and a severe AS with low gradient and preserved EF (i.e., low gradient because of decreased stroke volume). The current data argue for routine assessment of mitral ring displacement in these patients, which can be conveniently and reliably acquired in any patient with AS using standard echocardiographic methods. Thus, in patients with isolated low-gradient AS, a cutoff value of <9 mm had an excellent sensitivity (100%) and specificity (100%) to distinguish between moderate and severe AS.

Longitudinal myocardial kinetics (e.g., mitral ring displacement, a well-known but infrequently used marker for myocardial structural abnormalities) captures the functional consequences of myocardial fibrosis and predicts functional improvement after AVR. As shown in our previous study, the mitral ring displacement demonstrates good diagnostic utility to predict improvement in NYHA functional class in patients with severe aortic valve stenosis (8). In contrast, an EF >50% was less predictive. However, the clinical utility of this promising marker (e.g., in an improved pre-operative diagnostic algorithm) needs to be investigated in larger prospective cohorts.

Study limitations

The frequency of patients with severe AS and a low gradient was low, resulting in limited power of the analyses. This order of magnitude, however, is well in line with that in the literature in which a prevalence of low-flow gradient/severe AS with preserved LV EF of 16% by Doppler echocardiography and 14% by catheterization is reported (1,14). Because of the relatively small sample size, our findings should be regarded as preliminary. In particular, the clinical utility of mitral ring displacement as a clinically promising surrogate of risk needs to be replicated in larger studies. The reference standard for the assessment of LV function (i.e., end-systolic elastance) was not available, and in a sizable number of these older patients (n = 18), cMRI was not feasible because of claustrophobia.