Prevalence, Characteristics, and Impact on Prognosis of Aortic Stenosis in Patients With Cardiac Amyloidosis

Abstract

Background

Cardiac amyloidosis (CA) is frequently found in older patients with aortic stenosis (AS). However, the prevalence of AS among patients with CA is unknown. The objective was to study the prevalence and prognostic impact of AS among patients with CA.

Methods and Results

We conducted a retrospective analysis of a prospective registry comprising 976 patients with native aortic valves who were confirmed with wild type transthyretin amyloid (ATTRwt), hereditary variant transthyretin amyloid (ATTRv), or immunoglobulin light‐chain (AL) CA. CA patients' echocardiograms were re‐analyzed focusing on the aortic valve. Multivariable Cox regression analysis was performed to assess the mortality risk associated with moderate or greater AS in ATTRwt CA. The crude prevalence of AS among patients with CA was 26% in ATTRwt, 8% in ATTRv, and 5% in AL. Compared with population‐based controls, all types of CA had higher age‐ and sex‐standardized rate ratios (SRRs) of having any degree of AS (AL: SRR, 2.62; 95% Confidence Interval (CI) [1.09–3.64]; ATTRv: SRR, 3.41; 95%CI [1.64–4.60]; ATTRwt: SRR, 10.8; 95%CI [5.25–14.53]). Compared with hospital controls, only ATTRwt had a higher SRR of having any degree of AS (AL: SRR, 0.97, 95%CI [0.56–1.14]; ATTRv: SRR, 1.27; 95%CI [0.85–1.44]; ATTRwt: SRR, 4.01; 95%CI [2.71–4.54]). Among patients with ATTRwt, moderate or greater AS was not associated with increased all‐cause death after multivariable adjustment (hazard ratio, 0.71; 95%CI [0.42–1.19]; P=0.19).

Conclusions

Among patients with CA, ATTRwt but not ATTRv or AL is associated with a higher prevalence of patients with AS compared with hospital controls without CA, even after adjusting for age and sex. In our population, having moderate or greater AS was not associated with a worse outcome in patients with ATTRwt.

Nonstandard Abbreviations and Acronyms

AL

immunoglobulin light‐chain

AS

aortic stenosis

ATTR

transthyretin amyloidosis

ATTRv

hereditary variant transthyretin amyloid

ATTRwt

wild‐type transthyretin amyloid

AVA

aortic valve area

AVR

aortic valve replacement

CA

cardiac amyloidosis

Clinical Perspective.

What Is New?

• Aortic stenosis (AS) is common among patients with cardiac amyloidosis (CA), especially patients with wild‐type CA where one quarter of patients are affected by AS.

• In our study, AS in CA was not associated with a poorer prognosis in the wild‐type CA.

What Are the Clinical Implications?

• Patients with a diagnosis of wild‐type CA should be considered screened for AS.

• Further studies are needed to explore the possible common pathophysiological pathway between AS and CA.

Amyloidosis is a systemic disease caused by misfolded proteins forming amyloid fibrils that deposit in organs. ¹ , ² There are 3 main types of cardiac amyloidosis (CA): immunoglobulin light‐chain (AL) amyloidosis, due to amyloidogenic monoclonal light‐chain production by a plasma cell clone; and transthyretin amyloidosis caused by the deposition of misfolded transthyretin, which is divided into a hereditary variant (ATTRv) and a wild‐type/nonhereditary variant (ATTRwt). ¹ , ³

The association between aortic stenosis (AS) and CA has received particular attention since 8% to 15% of patients with AS undergoing aortic valve replacement (AVR) present evidence of CA. ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ Kristen et al reported that 74% of patients with AS, at the time of AVR, had amyloid deposition on their valves, ¹² while other studies reported evidence of aortic valve leaflet thickening on echocardiography in 13% to 31% of patients with CA. ¹³ , ¹⁴ , ¹⁵ However, little is known on the prevalence of AS among patients with an established CA diagnosis and whether a particular CA type is more closely associated with AS. CA was previously associated with a higher prevalence of low‐flow, low‐gradient (stages D2 and D3) pattern of AS, ⁷ , ¹⁶ but other AS series reported a majority of high‐gradient severe AS (stage D1) among patients with CA. ⁵ Thus, the distribution of flow‐gradient phenotypes in patients with both CA and AS remains unclear. Finally, the prognosis of AS in patients with established CA is unknown.

The objectives of this study were thus to (1) describe the point prevalence of aortic valve sclerosis and AS in a large CA cohort according to CA subtypes, (2) study flow/gradient profiles among patients with both CA and AS, and (3) study the impact of AS on patient prognosis.

METHODS

Population

We retrospectively analyzed the data of patients recruited prospectively in the CA registry of the referral center for CA at Hôpital Henri Mondor, Assistance Publique–Hôpitaux de Paris, Créteil, France. The registry included patients aged >18 years who were referred for a diagnostic workup of amyloidosis. For the present analysis, we included patients with confirmed ATTR or AL CA and excluded those with other types of CA. Patients for whom CA was ruled out were used as hospital‐based controls. Patients with CA with missing baseline echocardiogram and patients with prosthetic aortic valve at baseline were also excluded (see study flowchart in Figure 1). As per the prospective CA registry protocol, each patient underwent basic clinical evaluation, for which baseline demographic characteristics, medical history, cardiovascular risk factors, and medication taken were collected. An ECG and a transthoracic echocardiography were performed at inclusion. Standard assessment was completed by biological tests including cardiac biomarkers and genotyping. To compare with the prevalence of AS in the community, we further extracted data from the report of Nkomo et al where they pooled Western population‐based studies ¹⁷ (population‐based controls; see Data S1 for details). The study was approved by the French Patient Protection Committee (ie, ethics committee). All patients provided written informed consent. As the primary objective was to assess the point prevalence of AS in different CA groups, we did not perform a power sample calculation. The data underlying this article will be shared on reasonable request to the corresponding author.

Figure 1. Study flowchart.

Shows inclusion and exclusion of patients into the study. The study population comprised 976 patients with confirmed cardiac involvement of either ATTR or AL amyloidosis. Patients who were not confirmed with CA were used as hospital‐based controls for the comparison of AS prevalence. AA indicates serum amyloid amyloidosis; AL, immunoglobulin light‐chain; AS, aortic stenosis, ATTR wt/v, transthyretin amyloidosis wild‐type/hereditary variant type; and CA, cardiac amyloidosis.

Echocardiography

Experienced operators performed echocardiograms following a standard operating procedure, using Vivid S6, 7, or E9 (GE Vingmed, Horten, Norway) ultrasound systems. The echocardiograms were stored in an image server and were retrieved for image analysis using ECHOPac software (GE Healthcare, Chicago, IL) following a previously reported method. ¹⁸ Global longitudinal strain was measured offline using the 3 standard left ventricular (LV) apical views using an automated function, as the average of all 17 segments.

For the purpose of the present study, the echocardiograms of patients with CA were analyzed, focusing on the aortic valve by a single operator (M.‐S.A.) after having received training in the Echocardiography CoreLab of Québec Heart and Lung Institute and following a protocol (Table S1) adapted from American Society of Echocardiography and European Association of Cardiovascular Imaging guidelines. ¹⁹ AS was graded using a multiparameter integrated approach based on aortic valve area (AVA) and transvalvular gradients, but also on the severity fibrocalcific remodeling and the degree of impairment of the cusps' opening. The primary criterion for severe AS was an AVA ≤1.0 cm² or indexed AVA ≤0.6 cm²/m².

Patients with severe AS were classified into flow gradient–ejection fraction phenotypes using the staging classification of the American Heart Association. ²⁰ Low flow was defined as stroke volume index <35 mL/m², high gradient (type D1) was defined as mean gradient ≥40 mm Hg; patients with low flow, low gradient (type D2) were divided into “classical” low flow, low gradient (LV ejection fraction <50%) and paradoxical low flow, low gradient (LV ejection fraction ≥50%).

CA Workup

Diagnostic workup for CA confirmation was previously described. ²¹ , ²² It was based on a clinical workup comprising medical and family history, biomarkers (serum and urine immunoglobulin measurements), cardiac uptake at ⁹⁹Tc‐bisphosphonate bone scintigraphy, cardiac magnetic resonance (see protocol description in Data S1), evidence of amyloid deposits using Congo red staining and immunohistochemistry in endomyocardial or extracardiac biopsies. ATTR gene sequencing was performed to differentiate ATTRv and ATTRwt. The diagnosis of CA was made when amyloidosis was associated with a hypertrophic cardiomyopathy defined by an interventricular septal thickness >12 mm and a cardiac uptake at the bone scintigraphy without gammopathy for ATTR and in case of a gammopathy, a biopsy was systematically performed (extracardiac or cardiac) showing an immunostaining κ or λ for AL or transthyretin for ATTR.

Outcomes and Follow‐Up

The follow‐up began at the time of CA confirmation. The primary end point was all‐cause death. The secondary end point was the composite of all‐cause death or heart failure–related hospitalization. Observations were censored at the date of last contact alive or heart transplantation/LV assistance. Status and dates of events were obtained from the patients' usual follow‐up visits, by patient phone call, or using medical records. The first patient was recruited in June 2007 and the last one in September 2019. The database was frozen in April 2020 for statistical analysis. Patients with missing follow‐up were excluded from the analyses.

Statistical Analysis

Continuous variables are expressed as mean±SD or median (25th–75th percentile) for normally and nonnormally distributed variables, respectively (as tested by the Shapiro–Wilk test) and were compared using Student's t test or Wilcoxon‐Mann–Whitney U test as appropriate. Proportions are expressed as percentages and were compared using the χ² test unless the number of events was <5, in which case Fisher's exact test was used. Comparison between >2 groups were performed with the 1‐way ANOVA test for normally distributed data and the Kruskal–Wallis test for nonparametric data. We also calculated standardized (ie, age‐ and sex‐adjusted) prevalence of AS using direct standardization (see details in Data S1). All missing variables were excluded from the analyses.

The association between CA types and AS was assessed by (1) using multiple logistic regression following backward stepwise selection where the CA type was added to all clinically relevant variables with a P value ≤0.1 in univariable analysis; (2) comparing the standardized (ie, age‐ and sex‐adjusted) prevalence of moderate to severe AS (ie, grouping moderate and severe AS in a single category) in each CA subtype versus hospital‐based controls and population‐based controls.

To study the prognostic impact of moderate or greater AS in ATTRwt, survival analyses using multivariable Cox proportional hazards regression was performed. Survival analyses were not performed for ATTRv or AL due to a low proportion of significant AS in these subtypes. Multivariable adjustment was done with carefully selected clinically meaningful covariates (age, sex, New York Heart Association class 3 to 4, coronary artery disease, diabetes, stroke volume index, NT‐proBNP [N‐terminal pro‐B‐type natriuretic peptide]–estimated glomerular filtration rate staging, troponin T, and AVR). Proportional hazards assumption was verified using Shoenfield's residuals in all the reported models. Multicollinearity was ruled out by verifying that the variance inflation factor was ≤2.5 in a multiple linear regression model, which included the variables selected in the Cox model, and for this reason global longitudinal strain was deselected.

A 2‐sided P value <0.05 was considered statistically significant. Statistical analyses were performed with STATA BE version 18.0 (StataCorp, College Station, TX).

RESULTS

Overall, 2287 patients were referred for CA workup (see the flowchart in Figure 1), and CA was ruled out in 1197. Among 1090 remaining patients with confirmed CA, 114 were excluded (see Figure 1 for details). The final CA population was composed of 976 patients. The baseline clinical and imaging characteristics of the entire study cohort are provided in Tables 1 and 2, respectively. The mean age of the study population was 71±12 year and 72% were men. ATTRwt represented 41% (n=398) of the total population followed by AL (32% [n=317]) and ATTRv (27% [n=261]).

Table 1.

Baseline Clinical Characteristics and Comparison Between CA Subtypes

	ATTRwt	ATTRv	AL
	n=398	n=261	n=317
Age, y	80±7	68±13	65±11
Male sex	344 (86)	176 (67)	195 (62)
BMI, kg/m2 (1)	25±4	24±4	24±4
NYHA functional class III–IV (80)	147 (40)	88 (37)	151 (53)
Heart rate, bpm (87)	76±18	77±14	77±17
Systolic blood pressure, mm Hg (59)	129±21	124±21	115±20
Diastolic blood pressure, mmHg (60)	75±13	75±12	71±12
Atrial fibrillation/flutter on ECG (93)	148 (40)	43 (19)	36 (12)
Pacemaker	190 (48)	81 (31)	39 (12)
Defibrillator	65 (16)	84 (32)	128 (40)
Coronary artery disease	109 (27)	29 (11)	37 (12)
Stroke	50 (13)	23 (9)	19 (6)
Peripheral artery disease	15 (4)	6 (2)	3 (1)
Diabetes	65 (16)	43 (16)	46 (15)
Hypertension	235 (59)	125 (48)	117 (37)
Diuretics	293 (74)	133 (51)	219 (66)
Smoking	111 (28)	70 (27)	94 (30)
NT‐proBNP, ng/L (35)	3295 (1762–5926)	1930 (472–4183)	4786 (2157–10 328)
Creatinine, μmol/L (7)	110 (91–134)	91 (75–127)	102 (80–138)
eGFR, mL/min per 1.73 cm/m2 (8)	48 (37–60)	61 (41–84)	56 (37–76)
NT‐proBNP–eGFR staging (36)
Stage 1	137 (35)	118 (30)	133 (34)
Stage 2	142 (57)	59 (24)	46 (19)
Stage 3	91 (30)	128 (42)	86 (28)
Ultra‐sensitive troponin T, ng/L (99)	66 (44–93)	55 (25–86)	79 (47–133)
Hemoglobin, g/dL (40)	12.9±2.0	12.8±1.8	12.9±1.9
Surgical management
AVR	16 (4)	2 (1)	3 (1)
Conservative	376 (96)	247 (96)	290 (92)
Heart transplant	0 (0)	8 (3)	22 (7)
Tafamidis	218 (55)	163 (62)	…
Chemotherapy (59)	…	…	242 (94)

Data are presented as mean±SD, median (interquartile range), or number (%) Numbers in parentheses after variable name are number of missing variables if any. AL indicates immunoglobulin light‐chain amyloidosis; AS, aortic stenosis; ATTRwt/v, transthyretin amyloid wild‐type/hereditary transthyretin amyloidosis; AVR, aortic valve replacement; BMI, body mass index; eGFR, estimated glomerular filtration rate; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; and NYHA, New York Heart Association.

Table 2.

Baseline Imaging Characteristics and Comparison Between CA Subtypes

	ATTRwt	ATTRv	AL	P value
	n=398	n=261	n=317
Peak velocity, m/s (11)	1.7±0.8	1.4±0.5	1.4±0.6	<0.001
Mean gradient, mm Hg (23)	9.3±11.5	5.4±3.9	5.3±5.2	<0.001
Aortic valve area, cm2 (5)	2.3±0.9	2.5±1.4	2.7±0.7	<0.001
Aortic valve				<0.001
Morphologically normal valve, no AS	134 (34)	146 (56)	203 (64)
Abnormal valve, no AS	162 (41)	94 (36)	97 (31)
Any degree of AS	102 (26)	21 (8)	17 (5)
Greater than mild AS	75 (18)	13 (5)	9 (3)	<0.001
Indexed aortic valve area, cm2/m2 (9)	1.2±0.5	1.4±0.4	1.5±0.4	<0.001
Stroke volume index, mL/m2 (5)	30±9	31±10	31±11	0.988
LV ejection fraction, % (19)	49±12	50±14	54±12	<0.001
LV global longitudinal strain, % (41)	−10.3±3.6	−12.0±5.0	−11.6±4.5	<0.001
Relative wall thickness (41)	0.81±0.21	0.74±0.26	0.75±0.22	<0.001
Indexed LV mass, g/m2 (99)	186±52	170±62	145±44	<0.001
E/E' ratio (134)	18.8±8.4	16.7±8.0	19.3±10.1	0.003
LV filling pressure				<0.001
Normal	36 (9)	65 (25)	66 (21)
Intermediate/indeterminate	108 (27)	68 (26)	68 (21)
Elevated	254 (64)	128 (49)	183 (58)
Greater than moderate aortic regurgitation	0 (0)	0 (0)	1 (0)	0.592
Greater than moderate mitral regurgitation	22 (5)	13 (5)	12 (4)	0.552
MRI late gadolinium enhancement (464)	196 (49)	112 (43)	138 (44)	0.180

Data are presented as mean±SD, median (interquartile range), or number (%). Numbers in parentheses after variable name are number of missing variables if any. AL indicates immunoglobulin light‐chain amyloidosis; AS, aortic stenosis; ATTRwt/v, transthyretin amyloid wild‐type/hereditary transthyretin amyloidosis; CA, cardiac amyloidosis; and LV, left ventricular.

Overall, patients with ATTRwt were older, were more often of male sex, and had a higher proportion of atrial fibrillation, pacemakers, and coronary artery disease compared with the other CA types, while AL had the highest NT‐proBNP and troponin T levels (Table 1).

Crude Prevalence of AS and Characteristics of Patients With CA With Versus Without AS

The overall crude prevalence of mild or greater AS in this CA population was 15% (n=140). It is noteworthy that 36% of the overall cohort had aortic sclerosis. A comparison of patients with versus without AS for each of the CA subtypes is shown in Table S2 for clinical characteristics and Table S3 for imaging characteristics. Overall, compared with patients with CA with no AS, patients with both CA and AS were older with a higher prevalence of coronary artery disease. They had higher NT‐proBNP and troponin T levels (Table S2) as well as a higher filling pressure on echocardiography, but LV ejection fraction, LV wall mass, and late gadolinium enhancement proportion between patients with and without AS were similar (Table S3).

Association Between CA Type and AS

AS was more common among patients with ATTRwt with 26% versus 8% in ATTRv and versus 5% in AL (Figure 2, both P<0.001). The standardized (ie, age‐ and sex‐adjusted) prevalence of any grade of AS was also significantly higher among patients with ATTRwt (18%; 95% confidence interval (CI) [14%–21%]) compared with ATTRv (11%; 95%CI [6%–16], corrected P<0.05) and versus AL (10%; 95%CI [4%–15%], corrected P<0.05). Across all 3 CA subtypes, the presence of any grade of AS was associated with higher age and was further associated with coronary artery disease in patients with ATTRwt in multivariate logistic regression analysis (Table 3).

Figure 2. Crude prevalence of aortic stenosis and of aortic sclerosis in the overall population and according to CA subtype.

Histogram showing the crude prevalence of aortic stenosis in the whole cohort and according to cardiac amyloidosis type. AL indicates immunoglobulin light‐chain amyloidosis; AS, aortic stenosis; ATTRwt/v, transthyretin wild‐type/hereditary variant amyloidosis. *Corrected P<0.05.

Table 3.

Multivariate Predictors of Any Grade of AS Across CA Subtypes

	ATTRwt		ATTRv		AL
	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
Age, y	1.08 (1.04–1.13)	<0.001	1.13 (1.06–1.21)	<0.001	1.11 (1.04–1.19)	0.002
Male sex	0.93 (0.46–1.90)	0.853	1.50 (0.57–3.91)	0.410	0.68 (0.19–2.46)	0.560
Coronary artery disease	1.77 (1.06–2.95)	0.028
NT‐proBNP/1000 ng/L	1.04 (1.00–1.07)	0.065
Troponins/10 ng/L					1.04 (1.00–1.09)	0.063

AL indicates immunoglobulin light‐chain amyloidosis; AS, aortic; ATTR, transthyretin amyloidosis; ATTRwt/v, transthyretin amyloidosis wild‐type/hereditary; CA, cardiac amyloidosis; and NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide.

The standardized prevalence of moderate or greater AS was higher among hospital‐based controls versus population‐based controls (standardized rate ratio [SRR], 2.70; 95%CI [1.94–3.2]; Figure 3, Table S4), reflecting a certain degree of selection bias in our cohort. Hence, compared with population‐based controls, AL (SRR, 2.62; 95%CI [1.09–3.64]) and ATTRv (SRR, 3.41; 95%CI [1.64–4.60]) patients also had higher standardized prevalence of significant AS versus population‐based controls, yet they had similar rates as hospital‐based controls (SRRs versus hospital‐based controls, 0.97; 95%CI [0.56–1.14] and 1.27; 95%CI [0.85–1.44] for AL and ATTRv, respectively). Patients with ATTRwt, however, had higher standardized rates of significant AS both versus population‐based (SRR, 10.8; 95%CI [5.25–14.53]) and versus hospital‐based (SRR, 4.01; 95%CI [2.71–4.54]) controls. A comparison of baseline characteristics between patients with CA and hospital controls can be seen in Table S5.

Figure 3. Age‐ and sex‐adjusted prevalence of AS according to CA subtype.

Forest plot of the standardized (age‐ and sex‐adjusted) rate ratios and 95% CIs, to compare the prevalence of significant (ie, moderate to severe) AS among the different CA subtypes with our hospital‐based controls and population‐based reference as previously reported by Nkomo et al. ¹⁷ AL indicates immunoglobulin light‐chain amyloidosis; AS, aortic stenosis; ATTRwt/v, transthyretin wild‐type/hereditary variant amyloidosis; and CA, cardiac amyloidosis. *Corrected P<0.05.

Distribution of Flow Gradient–Ejection Fraction Patterns of AS in CA

Among patients with CA and AS, 43 (32%) had mild AS, 36 (25%) had moderate AS, and 61 (43%) had severe AS, with a higher proportion of moderate to severe AS in the ATTRwt group (compared with ATTRv and AL, both P<0.05) (Figure 4A).

Figure 4. AS severity and flow gradient patterns of severe AS in patients with cardiac amyloidosis.

A, Shows that ATTRwt patients had the highest proportion of severe AS compared with the other CA subtypes. B, Shows that only a small minority of patients with severe AS had a high gradient (ie, a mean gradient ≥40 mm Hg or peak velocity≥4 m/s) and a relative majority had either classical or paradoxical LFLG AS. AL indicates immunoglobulin light‐chain amyloidosis; AS, aortic stenosis; ATTRwt/v, transthyretin wild‐type/hereditary variant amyloidosis; LFLG, low flow/low gradient; and NFLG, normal flow/low gradient. *Corrected P<0.05.

The majority of patients (69%) had a low‐flow state. Among patients with both CA and severe AS (n=61; Figure 4B), there was a minority of high‐gradient severe AS (n=7 [11%]). The majority had “classical” low‐flow, low‐gradient severe AS (48%), followed by “paradoxical” low‐flow, low‐gradient severe AS (25%), and normal‐flow, low‐gradient severe AS (16%).

Prognostic Impact of AS in Patients With ATTRwt

Only 21 patients with severe AS underwent an aortic valve procedure (16 transcatheter AVRs, 3 surgical AVRs, and 2 balloon aortic valvuloplasties) at follow‐up. Of these, 18 (86%) had severe AS and 3 (14%) had moderate AS at baseline.

Due to a low proportion of moderate or greater AS in AL and ATTRv groups, survival analyses were performed only in the ATTRwt group. In the ATTRwt group, 1 patient was lost to follow‐up, and 140 patients reached the primary end point of all‐cause death during a median follow‐up of 18 (9–30) months. Among patients with ATTRwt, moderate or greater AS (n=75) was neither associated with all‐cause death in univariable or multivariable analysis (adjusted hazard ratio, 0.71 [95% CI, 0.42–1.19]; P=0.19; Figure 5, Table 4).

Figure 5. Prognostic impact of AS in patients with ATTRwt cardiac amyloidosis.

Kaplan–Meier survival curve for all‐cause death of patients with ATTRwt with moderate or greater AS (red line) compared with patients with ATTRwt with no or mild AS (blue line). AS indicates aortic stenosis; ATTRwt, wild‐type transthyretin amyloidosis; and HR, hazard ratio.

Table 4.

Multivariable Predictors of All‐Cause Death in Patients With ATTRwt

Number of events=140/397	Univariable			Multivariable
	HR	95% CI	P value	HR	95% CI	P value
≥Moderate aortic stenosis	1.43	(0.95–2.15)	0.08	0.71	(0.42–1.19)	0.19
Age, /y	1.07	(1.03–1.10)	<0.001	0.99	(0.95–1.03)	0.55
Male sex	1.89	(1.02–3.50)	0.04	1.41	(0.65–3.05)	0.39
NYHA functional class III–IV	2.62	(1.85–3.71)	<0.001	1.07	(0.70–1.64)	0.74
Coronary artery disease	1.73	(1.22–2.45)	0.002	1.39	(0.92–2.09)	0.12
Diabetes	1.40	(0.92–2.14)	0.12	1.60	(1.00–2.58)	0.052
Stroke volume index, mL/m2	0.95	(0.93–0.97)	<0.001	0.97	(0.94–0.99)	0.01
NT‐proBNP–eGFR staging, /stage	2.32	(1.92–2.80)	<0.001	1.53	(1.14–2.07)	0.005
Natural log troponin	3.57	(2.63–4.85)	<0.001	2.13	(1.42–3.21)	<0.001
Tafamidis treatment	0.16	(0.10–0.24)	<0.001	0.17	(0.10–0.28)	<0.001
AVR treatment	0.76	(0.44–1.33)	0.34	0.77	(0.28–2.13)	0.62

ATTRwt indicates ATTR wild‐type/hereditary transthyretin amyloidosis; AVR, aortic valve replacement; eGFR, estimated glomerular filtration rate; HR, hazard ratio; NT‐proBNP, N‐terminal pro‐B‐type natriuretic peptide; and NYHA, New York Heart Association.

DISCUSSION

The major findings of this large CA registry were (1) AS is highly prevalent in patients with ATTRwt, with 26% affected by mild or greater AS versus only 8% and 5% in patients with ATTRv and patients with AL, respectively; (2) the predominant pattern in patients with CA and severe AS is low‐flow, low‐gradient AS; and (3) AS was not associated with an increased mortality rate in the ATTRwt group.

AS, a Common Comorbidity of ATTRwt

To our knowledge, this is the first study in which aortic sclerosis and AS were screened and graded in a large sample of patients with CA. Our results suggest that 15% of the patients with CA and more than one quarter of those with ATTRwt harbor some degree of AS. When considering the proportion of patients with aortic sclerosis, only one third of patients with ATTRwt had a normal aortic valve. Our findings confirm the frequent coexistence of AS and ATTRwt reported in previous AS series. ⁵ , ⁷ It is also the first to demonstrate that among CA patients, ATTRwt is associated with a higher prevalence of AS compared with ATTRv and AL even after controlling for age and sex. Chacko et al reported the prevalence of AS in patients with CA and found that 7% of the patients with ATTR were found to have at least mild AS, with a higher prevalence of severe AS in ATTRwt compared with ATTRv. ²³ The reason for the discrepancy in AS prevalence between our study and the study by Chacko et al is unclear but could be related to the differences in inclusion/exclusion criteria and imaging analysis approach. While the presence of AS in our CA population was largely associated with older age, our findings together with recent studies raise the hypothesis that the coexistence of CA and AS could in part be explained by transthyretin amyloid deposits within aortic valve leaflets. Heuschkel et al recently reported multiomics evidence of ATTR deposits in surgically explanted stenotic aortic valves. ²⁴ Furthermore, Hussain et al recently found that aortic valve remodeling in patients with ATTR is not related to aortic valve calcium deposits as opposed to matched controls, further suggesting the potential role of ATTR amyloid deposits within the valve as a culprit lesion. ²⁵ In the present report, the sex‐ and age‐adjusted rates of AS in patients with ATTRwt were 10‐fold higher than population‐based controls and 4‐fold higher than hospital‐based controls (Figure 3). Further studies should be considered to investigate the role of transthyretin tetramer instability in the initiation and progression of AS in patients with ATTR.

Grading AS Severity in Patients With CA

Since CA is usually associated with severe LV concentric remodeling and LV function impairment, it is anticipated that AS would frequently present with low‐flow status, a condition that leads to underestimation of AS severity by the mean gradient. ²⁶ In this study, only 11% of patients with CA with severe AS had a mean gradient ≥40 mm Hg. This is much lower than the proportion (≈60%) of high‐gradient patterns generally reported in the population with severe AS, ²⁷ but consistent with those of Chacko et al, who reported a high‐gradient pattern in only 9% with severe AS. ²³ According to the Gorlin formula $\left(\mathrm{AVA}=\frac{\text{Mean flow rate}}{44.3\times \sqrt{\mathrm{MG}}}\right)$, a mean flow rate >280 mL/s is required to generate a mean gradient >40 mm Hg across a fixed AVA of 1.0 cm². In our patients with CA, only 5% had a flow rate ≥280 mL/s, and for that reason mean gradient cannot stand alone, while the calculation of AVA should be included when grading AS severity in patients with CA, as guidelines suggest. ¹⁹ In patients without CA, a low‐flow severe AS condition is often confirmed with a noncontrast computed tomography scan, but this may introduce a diagnostic trap as computed tomography contrast scores are often significantly lower among CA‐AS patients due to amyloid deposits that do not calcify. ²⁵ Special care should therefore be taken among patients with CA with suspected severe low‐flow AS, and a multimodality approach is needed to confirm severity as previously suggested. ²⁸

Prognostic Implications of AS in Patients With CA

To our knowledge, only Chacko et al reported severe AS to be associated with lower survival in ATTR. ²³ In the present report, moderate/severe AS was associated with a slight trend toward lower survival in univariate analysis among patients with ATTRwt, but not in multivariate analysis. This lack of predictive ability of AS could be explained by several factors. First, one third of patients with severe AS underwent an aortic valve procedure, mainly transcatheter AVR, which has changed the natural history of AS and may have neutralized its effect on outcomes, ⁷ whereas in the study by Chacko et al, a large part of the inclusion was done in the pretranscatheter AVR era. ²³ Second, the overall degree of AS severity might have been too mild to influence the outcome with sufficient magnitude to be statistically detectable. However, other studies suggest that moderate AS has a significant impact on prognosis in patients with severe LV dysfunction. ²⁹ , ³⁰ , ³¹ Third, in contrast to the study by Chacko et al, ²³ we perform multivariate adjustment, adjusting for age among others. Since patients with AS in our study were older than patients without AS, this was absolutely essential and may have neutralized the effect of AS on outcome in multivariate analysis.

Limitations

First, in this hospital‐based study, selection bias artificially increased the prevalence of AS, as evidenced by the higher standardized rates of significant AS in hospital‐based controls versus population‐based controls. Conversely, the exclusion of patients with prosthetic aortic valves (n=20) may have led to the underestimation of the prevalence of AS. Nevertheless, large CA registries from tertiary referral centers offer the best available setting to study the prevalence of AS in patients with CA. Second, this study lacked flow‐independent grading of AS severity (eg, projected AVA using dobutamine stress echocardiography, aortic valve calcium scoring), and AVA was the primary criterion to classify severe AS. Thus, the proportion of severe AS may have been overestimated. Third, despite multivariable adjustment, a residual confounding effect cannot be excluded in this observational study. We could not estimate to what extent the observed cardiac damage was related to CA rather than AS, which could have helped discern the prognostic impact of the heart valve disease from that of CA. In some subgroup analysis such as ATTRv and AL, the number of patients with AS was low, and the results should be interpreted with caution. The subcohort with moderate or greater AS was small, with heterogeneous delays from inclusion to AVR, which precluded studying the benefit of AS therapeutic management. Finally, we did not have complete data on the ethnicity of the patients, and were therefore not able to study the prognostic impact, for example, of being of African descent.

CONCLUSIONS

According to the results of this large hospital‐based registry, 15% of patients with confirmed CA had at least mild native AS, particularly patients with ATTRwt, of whom 26% had AS and an additional 40% had aortic sclerosis, leaving a minority with normal aortic valves. This excess of prevalence appears to be mainly driven by age, yet even after adjusting for age and other confounding factors, ATTRwt remains associated with a higher prevalence of AS. These results would suggest to screen for AS in patients with ATTRwt. However, we did not find significant AS to be associated with a negative prognosis, possibly because a proportion of patients with AS underwent AVR.

Further studies are needed to (1) elucidate whether transthyretin amyloid deposits participate in the initiation/progression of AS; (2) clarify the role of aortic valve calcium scoring of patients with CA to assess AS severity; and (3) clarify the optimal therapeutic management of AS in patients with CA, especially whether the less invasive transcatheter AVR would be a better option in these high‐risk patients.

Sources of Funding

None.

Disclosures

Dr Damy reports grants and personal fees from ALNYLAM, grants and personal fees from AKCEA, grants and personal fees from IONIS, and grants and personal fees from Pfizer, outside the submitted work. Dr Oghina reports personal fees from Pfizer outside the submitted work. Dr Pibarot reports grants from Edwards Lifesciences, Pi‐Cardia, and Medtronic, outside the submitted work. Dr Clavel reports grants from Edwards Lifesciences, Medtronic, and Pi‐Cardia, outside the submitted work. Dr. Carter‐Storch reports travel grants from Astra Zeneca and Abbott outside the submitted work; further, he is the chairman of the Danish Echocardiographic Society for which he receives no compensation. The remaining authors have no disclosures to report.

Supporting information

Data S1

Tables S1–S5

Reference 32