Alterations in Cardiac Structure and Function Associated With Sub‐Severe Aortic Stenosis Progression: The ARIC Study

Abstract

Background

American Heart Association and American College of Cardiology guidelines articulate 4 aortic stenosis (AS) stages to highlight its progressive nature, but limited data exist on cardiac alterations in sub‐severe stages.

Methods

ARIC (Atherosclerosis Risk in Communities) study participants with protocol echocardiography at Visit 5 (V5; 2011–2013) and free of aortic valve (AV) replacement or cardiovascular disease were classified by ACC/AHA AS stages at V5 and Visit 7 (V7; 2018–2019). AS stage progression was defined as AV replacement or hospitalization, or a higher stage at V7. Associations of AS stage at V5 and AS stage progression from V5 to V7 with cardiac structure and function were assessed using multivariable linear regression. Associations of extra‐AV cardiac abnormality categories with AS stage progression were assessed by multivariable logistic regression.

Results

Of 5206 V5 participants (age 75±5 years, 40% men), AS stages A and B at V5 were associated with greater left ventricular wall thickness, mass, and filling pressure measures at both V5 and V7 compared with Stage 0. Among 1562 participants with assessable AS stage at V7, AS stage progression occurred in 370 and was associated with greater worsening of these measures (all P<0.02). The presence of both left ventricular and left atrial extra‐AV abnormalities was associated with greater likelihood of AS stage progression (odds ratio 1.7 [95% CI, 1.2–2.6], P=0.009).

Conclusions

Early AS stages are associated with greater left ventricular mass and diastolic dysfunction, and AS stage progression is associated with worsening of these measures. The presence of both left ventricular and left atrial extra‐AV abnormalities is associated with a greater likelihood of early AS stage progression.

Nonstandard Abbreviations and Acronyms

ACC/AHA

American College of Cardiology /American Heart Association

AS

aortic stenosis

AV

aortic valve

GLS

global longitudinal strain

IPAW

inverse probability of attrition weights

LAVi

indexed left atrial volume index

LVMi

left ventricular mass index

PASP

pulmonary artery systolic pressure

VTI

velocity‐time integral

Clinical Perspective.

What Is New?

• This study identified the prevalence and predictors of American Heart Association and American College of Cardiology aortic valve stenosis (AS) stages among older adults in the community‐based ARIC (Atherosclerosis Risk in Communities) study who underwent 2 serial protocol echocardiograms 6 years apart.

What Are the Clinical Implications?

• The graded associations of higher left ventricular mass and measures of left ventricular filling pressure (E/e′, left atrial volume index) with more advanced AS stages and greater likelihood of progression highlight the relevance of monitoring these parameters in patients at risk for or with sub‐severe AS.

• These extra‐aortic valve measures could serve as readily accessible imaging biomarkers to inform AS progression screening or help select individuals for clinical trials designed to prevent AS.

Calcific aortic valve disease and associated aortic valve stenosis (AS) increase with age, causing significant mortality and morbidity that increases with the severity of AS, even in its asymptomatic phases. ¹ , ² While current practice guidelines and clinical care focus on aortic valve (AV) interventions in severe symptomatic and asymptomatic AS, ³ , ⁴ American Heart Association and American College of Cardiology (AHA/ACC) guidelines have also developed the construct of AV stages to highlight the progressive nature of calcific aortic valve disease, defining 4 AS Stages based on AS hemodynamic severity and the AV morphology. ⁵ , ⁶ Hemodynamic progression of calcific aortic valve disease is known to culminate in left ventricular (LV) remodeling and dysfunction in the context of severe AS, ⁷ but myriad additional changes in cardiac structure and function may occur in the context of sub‐severe AS progression, including LV hypertrophy and concentric remodeling, diastolic dysfunction with elevated filling pressures, and concomitant myocardial ischemia, fibrosis, and scar formation. ⁸ , ⁹ , ¹⁰ An additional AS categorization scheme has recently been proposed using the extent of extra‐AV cardiac remodeling and dysfunction to predict prognosis in AS. ¹¹ , ¹² , ¹³ However, limited longitudinal data exist regarding changes in cardiac structure and function associated with sub‐severe AS and AS progression, and the potential utility of the newly proposed categorization system in predicting AS progression. This study leveraged longitudinal echocardiographic data in a large cohort of older adults to describe the alterations in cardiac structure and function associated with sub‐severe AS stages as defined by AHA/ACC and their progression, and to determine the prognostic role of extra‐AV abnormalities categorizations for AS stage progression.

METHODS

Dr AM Shah had full access to all the data in the study and took responsibility for the integrity of the data and the accuracy of the data analysis.

Study Population

We studied participants in the prospective community‐based ARIC (Atherosclerosis Risk in Communities) study, which originally enrolled middle‐aged adults in 4 US sites between 1987 and 1989: Forsyth County, NC; Jackson, MS; suburban Minneapolis, MN; and Washington County, MD. The design and procedures of ARIC have been previously described. ¹⁴ The study was approved by an institutional review committee of the 4 sites, and the subjects gave informed consent. The deidentified ARIC data set can be requested at https://biolincc.nhlbi.nih.gov/studies/aric/. This analysis included ARIC participants who underwent a protocol echocardiogram at study Visit 5 (2011–2013), had adequate images for assessing AS, and were free of prior AV replacement, heart failure, or history of myocardial infarction. Of 6538 participants attending Visit 5, 5206 were included, among whom 1991 (38%) had other valvular heart disease (VHD) and 264 (5%) had atrial fibrillation (AF) and were excluded from most analyses (see Statistical Methods section below). Protocol echocardiography was repeated in 2591 of these participants at study Visit 7 (2018–2019; Figure S1).

Echocardiography

The echocardiography procedures used in ARIC at Visit 5, including intra‐reader reproducibility metrics, have been previously published, and similar procedures were used at Visit 7. ¹⁵ Intra‐reader reproducibility assessments demonstrated a coefficient of variation <10% for all key echocardiographic measures. For example, coefficient of variation for global longitudinal strain (GLS) was 7.7% and for LV mass was 9%. Certified sonographers obtained the studies at Visits 5 and 7 using standard imaging devices (Philips iE33, Koninklijke Philips, The Netherlands), probes (Philips XMatrix), and acquisition protocols. Quantitative measurements for studies from both visits were similarly performed at the same central Echocardiography Reading Center by dedicated analysts blinded to clinical data and in accordance with standards from the American Society of Echocardiography. ¹⁶ , ¹⁷ , ¹⁸ At both Visits, staff cardiologists at the Echocardiography Reading Center with Core Cardiology Training Symposium (COCATS) level III advanced training in echocardiography and/or American Society of Echocardiography Board Certification in Comprehensive Adult Echocardiography overread all quantitative measures at both visits.

AV peak jet velocity (Vmax) and velocity‐time integral (VTI) and the LV outflow tract (LVOT) VTI by continuous‐wave and pulsed wave Doppler respectively were acquired from the apical 5 chamber view, and LVOT diameter was measured in the parasternal long axis view. Aortic valve area was calculated using the continuity equation as follows: aortiv valve area=[CSA_LVOT×VTI_LVOT]/VTI_AV, where CSA indicates LVOT cross‐sectional area. ⁷ , ¹⁰ The parasternal long‐axis view was used to measure the anterior and septal wall thickness of the LV. The modified Simpson’s rule was used to evaluate the LV volumes, LV ejection fraction (LVEF), and body surface area‐indexed left atrial volume index (LAVi). GLS was measured from the apical 4‐chamber and 2‐chamber views using the TomTec Cardiac Performance Analysis. Mitral inflow early peak velocity (E wave) was measured using pulsed‐wave Doppler, and the early septal mitral annular velocity (e′) was measured using tissue Doppler. Pulmonary artery systolic pressure was calculated from the tricuspid regurgitation jet velocity.

Definition of AS Stages

AS stages were defined uniformly at both Visits 5 and 7 based on the ACC/AHA guideline recommendations and operationalized in ARIC as previously described. ⁵ Stage A, which does not represent clinical AS but rather a state of risk, was defined as the presence of a bicuspid AV or aortic sclerosis based on a peak AV velocity of 1.5–2.0 m/sec. Stage B AS was defined based on a peak AV velocity of 2.0–3.9 m/sec. Stage C was based on a peak AV ≥4.0 m/sec, mean pressure gradient ≥40 mm Hg or aortic valve area <1 cm², and no symptoms of anginal pain, dyspnea, or exhaustion, while Stage D was defined by the same objective measures in the presence of symptoms. Dyspnea, exhaustion, and anginal pain were ascertained based on participant response to the annual follow‐up questionnaire closest to the visit date.

Other VHD Stages

The stages of other VHD, including aortic regurgitation, mitral regurgitation, and mitral stenosis, were operationalized based on the ACC/AHA guideline as previously described. ⁵ Participants who had another valvular lesion with a greater stage than the stage of AS by any level were classified as having other VHD.

AS Stage Progression

Among participants with echocardiographic data from both Visit 5 and Visit 7, participants with a higher AS Stage at Visit 7 than at Visit 5 were considered to have either incident AS stage (transitioning from Stage 0 to Stage A or greater) or AS stage progression. Among participants who attended Visit 5 but not Visit 7, those with a post‐Visit 5 hospitalization with an AS‐related International Classification of Diseases (ICD) code (424.1 in ICD9; I35.0 or I35.2 ICD10) ¹⁹ or with an ICD procedure code for AS intervention (35.21, 35.22, or 35.11in ICD9‐ PCS; 02NFxx, 027Fxx, 02QFxx, 02RFxx, X2RF3xx, 02RF0xx, 02RF4xx, or X2RF0xx in ICD10‐ PCS), ²⁰ , ²¹ including transcatheter aortic valve implantation post‐visit 5 (35.05 or 35.06 in ICD9‐ PCS; 02RF37H, 02RF38H, 02RF3J, 02RF3KH, 02RF37Z, 02RF38Z, 02RF3JZ, or 02RF3KZ in ICD10‐ PCS), ²² were classified as AS stage C/D at Visit 7. Performance characteristics of these ICD codes are provided in Table S1.

AS Categories Based on Extra‐AV Abnormalities in Cardiac Structure and Function

Participants were categorized into 5 categories according to the presence of extra‐AV abnormalities based on a modified version of the categorization previously proposed by Généreux P et al. ¹² Category 1 was defined by the absence of any associated extra‐AV abnormality; Category 2 by the presence only of LV abnormalities including LVEF <50%, LV mass index (LVMi) >115 g/m² in men and >95 g/m² in women, or E/e′>14; Category 3 by the presence of left atrial (LA) abnormalities including prevalent atrial fibrillation (AF), moderate to severe mitral regurgitation, or LAVi >34 mL/m²; Category 4 by the combined presence of any LV abnormality and any LA abnormality; and Category 5 by the presence of right ventricular fractional area change (RV‐FAC) <35% or pulmonary artery systolic pressure >60 mmHg. The modification in the categorization proposed by Généreux et al included combining the right ventricle (RV) and pulmonary artery (PA) abnormalities in 1 category (Category 5) because of the limited number of participants meeting these criteria, and assessing RV dysfunction using RV function with prognostic value only, as tricuspid annular plane systolic excursion and RV strain were not available in our data set. Therefore, isolated RV abnormalities were not analyzed as an independent group.

Clinical Covariates

Hypertension was defined based on participant report of antihypertensive medication use or blood pressure readings ≥140/90 mm Hg at any ARIC visit. Diabetes was defined based on self‐report of a physician’s diagnosis of diabetes, use of antidiabetic medications, fasting glucose ≥126 mg/dL, or non‐fasting glucose ≥200 mg/dL at any ARIC visit. Weight and height were recorded at Visit 5, and body mass index (BMI) was calculated. Heart rate was measured at the time of echocardiography, and blood pressure was measured 3 times at the time of ARIC visit 5 and 7, with the average of the last two used to determine averaged systolic blood pressure. The estimated glomerular filtration rate was calculated using the Chronic Kidney Disease Epidemiology Collaboration equation. ²³ Prevalent coronary heart disease (CHD) was defined by self‐report of CHD at Visit 1 and adjudicated CHD events occurring by Visit 5. The adjudication of CHD was previously described. ²⁴ AF prevalence was defined through hospitalization with ICD‐9427.31 or ICD‐10 I48.91 discharge codes by Visit 5. ²⁵ At Visit 5, the Fried criteria—which considers walking speed, grip strength, low energy expenditure, weight loss, and exhaustion—were used to measure frailty. ²⁶

Statistical Methods

For analyses of the correlates of AS Stages and of AS Stage progression, participants with other VHD or prevalent AF were excluded. Multivariable linear regression models were used to estimate the cross‐sectional associations of AS Stage at Visit 5 with measures of cardiac structure and function at Visit 5 and with longitudinal changes in measures of cardiac structure and function between Visits 5 and 7 (mean time interval 6.6±0.7 years). Models adjusted for age, sex, race, hypertension, diabetes, BMI, estimated glomerular filtration rate, CHD, systolic blood pressure, and study center in addition to heart rate at the time of echo (ie, at Visit 5 for the cross‐sectional analyses and at both visits for the longitudinal analysis). For analyses with the Visit 7 cardiac structure and function measure as outcome, models additionally adjusted for the value of that measure at Visit 5. The Visit 5 value was not included as a model covariate in analyses with change in the cardiac structure and function measure as the outcome given prior studies demonstrating that baseline adjustment in non‐randomized studies may lead to spurious associations. ²⁷ An additional sensitivity analysis was performed for the cross‐sectional Visit 5 analysis restricted to participants who also had echocardiographic data available at Visit 7.

We used multivariable linear regression to assess the association of progression in AS Stage between Visits 5 and 7 with concomitant changes in cardiac structure and function between Visits 5 and 7. Participants with Stage C/D AS at Visit 5 were excluded as they were not eligible to progress. As before, models adjusted for age, sex, race, hypertension, diabetes, BMI, estimated glomerular filtration rate, CHD, and study center in addition to s and heart rate at both visits. These analyses were performed in the study sample overall and stratified by Visit 5 AS Stage.

For analyses of the association between extra‐AV abnormalities categories and AS Stage progression, participants with other VHD or prevalent AF were not excluded. Multivariable logistic regression models were used adjusting for age, sex, race, hypertension, diabetes, BMI, estimated glomerular filtration rate, CHD, and study center in addition to systolic blood pressure and heart rate at Visit 5. Additional analyses were performed stratifying participants by Visit 5 AS stage (no AS stage [ie, Stage 0], Stage A, Stage B AS), and testing for effect modification by Visit 5 AS stage on the association between extra‐AV abnormality category and AS Stage progression.

For progression analyses, sensitivity analyses were performed incorporating inverse probability of attrition weights (IPAW) as to account for Visit 7 non‐attendance. For IPAW, Visit 7 non‐attendance was modeled among participants alive through the end of Visit 7 using the following covariates from Visit 5: age, sex, race, study center, frailty, cancer, diabetes, hypertension, and prevalent heart failure. The resulting calculated weights were incorporated to estimate AS stage progression among all Visit 5 participants who had AS Stage B or less and were alive at the time of Visit 7. Another sensitivity analysis was conducted using a fully adjusted multivariable linear regression model to assess the association between changes in peak AV velocity as a continuous variable and the changes in cardiac structure and function.

The estimates are reported as model‐based (ie, adjusted) along with the standard errors, and P values for trends across groups were compared using linear regression. All analyses were performed using STATA 16. Two‐sided P values of <0.05 were considered significant.

RESULTS

The mean age of the 5206 participants at Visit 5 was 75±5 years, 40% were men, and 22% reported Black race. Mean age of the 2591 study participants with available echocardiography at Visit 7 was 81±4 years. An additional 62 participants experienced an AS‐related hospitalization between visits 5 and 7 (Figure S1).

AS Stage and Changes in Cardiac Structure and Function

Further exclusion of participants with other VHD or prevalent AF at Visit 5 resulted in 3072 participants with mean age 75±5 years (42% men, 25% Black race) and the following prevalence of AS Stages: 20% Stage A, 6% Stage B, and 1% Stage C/D. Participants with Stage B or C/D AS had a higher prevalence of hypertension, CHD, and chronic kidney disease, in addition to higher pulse pressure and NT‐proBNP (N‐terminal pro b‐type natriuretic peptide) (Table 1). Higher AS stages at Visit 5 were associated with larger LVEDV, LVESV, wall thickness, and LVMi at Visit 5 in a graded manner as compared with Stage 0 in fully adjusted models (Table 2). Markers of LV filling pressure, including E/e′ and LAVi, were higher in Stage B AS in comparison to Stage 0 and Stage A. In contrast, LVEF was higher in Stage A and B AS when compared with Stage 0 AS (Table 2). These findings persisted in analyses restricted to participants who attended Visit 7 (Table 3).

Table 1.

Baseline Characteristics of Participants With AS Stages

	Stage 0 AS	Stage A AS	Stage B AS	Stage C/D AS	P value
	n=2249	n=617	n=172	n=34
Age	74±5	75±5	78±5	78±6	<0.001
Men	940 (42%)	260 (42%)	79 (46%)	16 (47%)	=0.30
Black	594 (26%)	133 (22%)	23 (13%)	7 (21%)	<0.001
Study center
Forsyth County	464 (21%)	119 (19%)	26 (15%)	11 (32%)	<0.001
Jackson	541 (24%)	123 (20%)	23 (13%)	6 (18%)
Minneapolis	726 (32%)	168 (27%)	63 (37%)	6 (18%)
Washington County	518 (23%)	207 (34%)	60 (35%)	11 (32%)
Smoking	1376 (61%)	363 (59%)	123 (72%)	19 (56%)	=0.38
Hypertention	1785 (79%)	541 (88%)	152 (88%)	30 (88%)	<0.001
Diabetes	839 (37%)	238 (39%)	66 (38%)	18 (53%)	=0.19
Chronic kidney disease	506 (23%)	168 (27%)	56 (33%)	11 (33%)	<0.001
Chronic heart disease	130 (6%)	56 (9%)	27 (16%)	5 (16%)	<0.001
Stroke	42 (2%)	21 (3%)	12 (7%)	1 (3%)	<0.001
Body mass index	29±6	30±6	30±6	28±6	<0.001
Heart rate at echo	67±11	67±12	68±12	66±9	=0.61
Pulse pressure	61±13	65±14	65±14	67±15	<0.001
Average SBP	129±17	131±18	128±17	131±24	=0.34
Average DBP	68±10	66±11	63±10	64±15	<0.001
CKD‐EPI eGFR	72±17	71±17	68±20	66±17	<0.001
Low‐density lipoprotein	108±35	103±32	102±38	103±34	=0.001
C‐reactive protein	2.0 [0.9, 4.1]	2.1 [1.0, 4.6]	2.2 [1.1, 4.3]	2.5 [0.8, 4.2]	=0.05
NT‐proBNP	93 [51, 169]	117 [64, 237]	191 [85, 391]	231 [132, 552]	<0.001

Participants with other valvular heart disease or prevalent atrial fibrillation at visit 5 were excluded. AS indicates aortic stenosis; CKD‐EPI, Chronic Kidney Disease Epidemiology Collaboration; DBP, diastolic plod pressure; eGFR, estimated glomerular filtration rate; NT‐proBNP, N‐terminal pro b‐type natriuretic peptide; and SBP, systolic blood pressure.

Table 2.

Cross‐Sectional Associations of Aortic Stenosis Stage With Cardiac Structure and Function at Visit 5

Adjusted value of	Stage 0 n=2249	Stage A n=617	Stage B n=172	Stage C/D n=34	P for trend
LVEDV, mL	79.3±0.4	83.57±0.7*	86.1±1.4*	87.3±3.2*	<0.001
LVESV, mL	27.4±0.2	27.8±0.4	29.2±0.7*	30.6±1.6*	=0.005
LV wall thickness, cm	0.98±0.00	0.99±0.00*	1.03±0.01* , †	1.13±0.02* , † , ‡	<0.001
LV mass index, g/m2	75.4±0.4	78.7±0.7*	81.5±1.3*	95.5±3.0* , † , ‡	<0.001
LVEF, %	65.8±0.1	67.3±0.2*	67.1±0.4*	65.1±1.0	<0.001
GLS, %	−18.1±0.1	−18.6±0.1*	−17.9±0.2†	−17.4±0.4†	=0.28
E/e′ septal ratio	11.5±0.1	12.6±0.1*	14.9±0.3* , †	14.4±0.7*	<0.001
LAVi, mL/m2	23.4±0.1	26.3±0.3*	29.2±0.5* , †	27.8±1.2*	<0.001
PASP, mm Hg	27.1±0.2	28.9±0.3*	28.4±0.5*	29.7±1.2*	<0.001

Participants with other valvular heart disease or prevalent atrial fibrillation at visit 5 were excluded. Models adjusted for age, sex, race, hypertension, diabetes, body mass index, estimated glomerular filtration rate, non‐myocardial infarction coronary heart disease, and study center in addition to systolic blood pressure and heart rate at the time of echo (ie, V5). GLS indicates global longitudinal strain; LAVi, indexed left atrial volume; LV, left ventricular; LVEDV, left ventricular end‐diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end‐systolic volume; and PASP, pulmonary artery systolic pressure. The reported estimates are model‐based (adjusted)±standard errors.

^* P<0.05 compared with Stage 0.

^† P<0.05 compared with Stage A.

^‡ P<0.05 compared with Stage B.

Table 3.

Association of Aortic Stenosis Stage at Visit 5 With Cardiac Structure Function at Visit 5, Visit 7, and the Change Between Visits 5 and 7 Among Participants With Echocardiographic Data at Both Visits

Adjusted value of		Stage 0 at Visit 5 n=1206	Stage A at Visit 5 n=287	Stage B at Visit 5 n=67	P for Trend
LVEDV, mL	V5 (who attended Visit 7)	81.0±0.5	84.6±1.1	85.5±2.2	<0.001
	V7 (adjusted for baseline)	85.7±0.4	86.9±0.9	86.9±1.9	=0.22
	Change from V5 to V7	4.1±0.5	4.0±1.0	3.6±2.0	=0.85
LVESV, mL	V5 (who attended Visit 7)	28.1±0.3	28.0±0.6	29.5±1.2	=0.44
	V7 (adjusted for baseline)	31.6±0.3	32.7±0.6	31.0±1.3	0.53
	Change from V5 to V7	3.5±0.3	4.6±0.6	2.6±1.3	=0.63
LV wall thickness, cm	V5 (who attended Visit 7)	0.97±0.00	0.99±0.01	1.00±0.01	<0.001
	V7 (adjusted for baseline)	1.02±0.00	1.03±0.01	1.07±0.01	=0.004
	Change from V5 to V7	0.05±0.00	0.04±0.01	0.09±0.01	=0.22
LV mass index, g/m2	V5 (who attended Visit 7)	74.0±0.5	76.6±0.9	77.9±1.9	0.003
	V7 (adjusted for baseline)	78.7±0.4	80.4±0.9	83.0±1.8	0.009
	Change from V5 to V7	4.3±0.4	4.9±0.9	7.2±1.9	=0.15
LVEF, %	V5 (who attended Visit 7)	65.7±0.2	67.4±0.3	66.4±0.7	< 0.001
	V7 (adjusted for baseline)	64.3±0.2	64.2±0.4	65.7±0.8	0.25
	Change from V5 to V7	−1.6±0.2	−2.6±0.4	−0.5±0.8	=0.58
GLS, %	V5 (who attended Visit 7)	−18.1±0.1	−18.7±0.1	−18.0±0.3	0.041
	V7 (adjusted for baseline)	−17.8±0.1	−17.7±0.1	−17.0±0.3	0.04
	Change from V5 to V7	−0.3±0.1	−0.8±0.2	−1.1±0.4	0.005
E/e′ septal ratio	V5 (who attended Visit 7)	11.4±0.1	12.2±0.2	14.1±0.4	<0.001
	V7 (adjusted for baseline)	14.4±0.1	15.7±0.3	16.6±0.5	<0.001
	Change from V5 to V7	2.9±0.1	3.9±0.3	4.1±0.5	<0.001
LAVi, mL/m2	V5 (who attended Visit 7)	23.2±0.2	25.8±0.4	27.2±0.8	<0.001
	V7 (adjusted for baseline)	26.0±0.2	27.3±0.4	28.8±0.8	<0.001
	Change from V5 to V7	2.4±0.2	3.0±0.4	4.1±0.8	=0.024
PASP, mm Hg	V5 (who attended Visit 7)	26.4±0.2	28.5±0.4	26.9±0.8	<0.001
	V7 (adjusted for baseline)	31.9±0.3	32.8±0.6	34.6±1.3	=0.027
	Change from V5 to V7	5.0±0.3	5.2±0.7	7.5±1.3	=0.17

Participants with other valvular heart disease or prevalent atrial fibrillation at visit 5 were excluded. Models adjusted for age, sex, race, hypertension, diabetes, body mass index, estimated glomerular filtration rate, non‐myocardial infarction coronary heart disease, and study center in addition to systolic blood pressure and heart rate at the time of echo (ie V5 and V7 for V7 and delta). Models for V7 echo measures were additionally adjusted for the baseline value of that measure at Visit 5. The reported estimates are model‐based (adjusted)±standard errors. GLS indicates global longitudinal strain; LAVi, indexed left atrial volume; LV, left ventricular; LVEDV, left ventricular end‐diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end‐systolic volume; and PASP, pulmonary artery systolic pressure.

In models adjusting for demographics, cardiovascular risk factors, and Visit 5 value of the echocardiographic measures of interest, AS Stages A and B at Visit 5 were associated with greater LV wall thickness, LVMi, E/e′ septal ratio, LAVi, and pulmonary artery systolic pressure at Visit 7 compared with Stage 0. Associations with greater increases in E/e′ septal and LAVi were also noted in analyses modeling change in echocardiographic measures from Visit 5 to Visit 7 as the outcome. In contrast to the cross‐sectional analysis, higher AS Stage at Visit 5 was associated with worse global longitudinal strain at Visit 7 and with greater decline in absolute global longitudinal strain between Visits 5 and 7 (Table 3).

Alterations in Cardiac Structure and Function Associated With Progression in AS Stage

Of the 1562 participants free of other VHD and AF who underwent echocardiography at Visit 7, 370 (24%) demonstrated progression in AS Stage from Visit 5 to 7: 273 progressed to incident AS Stage A or higher; 69 progressed from Stage A to B or C/D; and 28 progressed from Stage B to C/D. Participants with progression in AS Stage over 6 years demonstrated more prominent concomitant increases in LV wall thickness, LVMi, and diastolic dysfunction. Consistent results were observed in sensitivity analyses incorporating IPAW (Table 4; Table S2). In contrast, decreases in LVEF and GLS between Visits 5 and 7 were less prominent in participants who had progression in the AS stage compared with those who did not have progression (Table 4), mainly attibutable to participants who progressed to incident AS Stage A or higher (Table S3). The participants who had AS stage progression from Stage A to Stage B or C/D showed worsening in all diastolic function measures in addition to LVMi (Table S3). Consistent results were observed in analyses assessing the association of change in peak AV velocity as a continuous variable with concomitant changes in cardiac structural and functional measures (Table S4).

Table 4.

Changes in Cardiac Structure and Function Over 6 years (V5 to V7) Among Those With and Without Concomitant Progression in Aortic Stenosis Stage

Adjusted value of	Participants without progression n=1192	Participants with progression n=370	P value
Delta LVEDV, mL	4.0±0.5	4.2±0.9	=0.83
Delta LVESV, mL	3.9±0.3	3.1±0.6	=0.28
Delta LV wall thickness, cm	0.05±0.00	0.06±0.01	=0.020
Delta LV mass index, g/m2	3.8±0.4	6.7±0.8	=0.002
Delta LVEF, %	−2.1±0.2	−0.6±0.4	<0.001
Delta GLS, %	−0.6±0.1	0.1±0.2	<0.001
Delta E/e′ septal ratio	2.9±0.1	3.9±0.2	<0.001
Delta LAVi, mL/m2	2.3±0.2	3.3±0.4	=0.019
Delta PASP, mm Hg	4.6±0.3	7.4±0.7	<0.001

Participants with absent Visit 7 echocardiography and participants with Stage C/D aortic stenosis, other valvular heart disease, or atrial fibrillation at Visit 5 were excluded. Models adjusted for age, sex, race, hypertension, diabetes, BMI, estimated glomerular filtration rate, non‐myocardial infarction coronary heart disease, and study center in addition to systolic blood pressure and heart rate at the time of echo (ie, V5 and V7). The reported estimates are model‐based (adjusted)±standard errors. GLS indicates global longitudinal strain; LAVi, indexed left atrial volume; LV, left ventricular; LVEDV, left ventricular end‐diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end‐systolic volume; and PASP, pulmonary artery systolic pressure.

Extra‐AV Cardiac Dysfunction Categories and AS Stage Progression

The prevalence of extra‐AV abnormality categories among the 5170 participants liable to AS stage progression (ie, less than Stage C/D AS at Visit 5, not excluding those with other VHD or prevalent AF at Visit 5) was 67% for no associated abnormality, 17% for associated LV abnormality only, 10% for associated LA abnormality only, 6% for combined LV and LA abnormality, and 1% for RV or PA abnormality (Figure S2). The relative contribution of abnormal parameters in each category is demonstrated in Figure S3. Notably, among the 56 participants with RV or PA abnormality, 51 had RV abnormality only, 4 had PA abnormality only, and 1 had both.

The number of participants who progressed to a higher AS stage was 682 (Table S5). In multivariable logistic regression models among participants with available progression data, only the category defined by the presence of both LV and LA abnormalities demonstrated higher odds of AS stage progression (1.7; [95% CI, 1.2–2.6], P=0.009) relative to the No associated abnormality category (Figure [A]). Sensitivity analyses incorporating IPAW and stratifying by Visit 5 AS stage demonstrated generally consistent results (Figure [B]; Figure S4). There was no evidence of significant effect modification by baseline AS stage on the association between extra‐AV cardiac abnormality category and AS stage progression (p for interaction=0.40).

Figure . Odds of progression in aortic stenosis stage over 6 years among 5 categories of extra‐AV abnormality categories (A) among participants with echocardigraphic data at both timepoints, (B) in analyses accounting for non‐attendance at follow‐up using IPAW.

Isolated PA abnormality, Isolated RV abnormality, and their combination were combined in one category. Models adjusted for age, sex, race, hypertension, diabetes, body mass index, estimated glomerular filtration rate, coronary heart disease, and study center in addition to systolic blood pressure and heart rate at Visit 5. AV indicates aortic valve; IPAW, Inverse probability of attrition weights; LA, left atrial; LV, left ventricular; PA, pulmonary artery; and RV, right ventricular.

DISCUSSION

Using prospectively acquired serial protocol echocardiography in a community‐based cohort, we quantified the consequences of less than severe AS stages as defined in AHA/ACC guidelines on cardiac structure and function. Higher AS stages were associated with larger LV volumes, greater LV mass, and worse diastolic dysfunction, and with greater increases in LV filling pressures over 6‐year follow‐up. AS stage progression over 6 years was also associated with greater concomitant increases in LV volumes, mass, and measures of LV filling pressure. AS stage progression was not associated with systolic function among participants with Stage A or B AS, and was associated with a slight increase in contractility among participants who progress from Stage 0 to Stage A. Finally, the presence of extra‐AV cardiac abnormalities involving both the LV and LA was associated with greater progression to AS stages. These findings clarify the bidirectional relationships between sub‐severe AS stages, LV remodeling, and diastolic dysfunction, and identify extra‐valvular cardiac measures as risk markers for AS stage progression with potential implications for clinical monitoring.

Several studies have included serial echocardiographic assessment of AS, but the majority focused on either the effect of risk factors and potential medications on the hemodynamic assessment of the AV ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² or on the impact of AS progression on cardiovascular outcomes. ² , ³³ In our study, we used the AHA/ACC Stages to describe cardiac structure and function in older adults at risk (Stage A) of AS in comparison to those who had mild or moderate AS (Stage B). The association of AS with LV remodeling and diastolic dysfunction is well documented, though the majority of prior studies have focused on patients with hemodynamically significant AS. ³⁴ , ³⁵ , ³⁶ Previous studies establish the higher prevalence of LV remodeling and hypertrophy in patients with moderate AS, ³⁷ and impaired diastolic function evidenced by lower e′ and higher E/e′ among patients with even mild AS. ³⁸ The longitudinal association between AS severity and LV hypertrophy has also been described. Among patients with primarily moderate and severe AS, greater progression in LV wall thickness over 2 years of follow‐up was associated with greater concomitant increases AV mean gradient. ³⁹ Importantly, our study focused on early AS stages, including predominantly individuals with aortic sclerosis or hemodynamically mild stenosis, and protocolized follow‐up imaging was performed after a longer time interval. Changes in myocardial collagen architecture and microcirculatory dysfunction resulting from chronically elevated afterload in AS are potential explanations for the well‐recognized associations of AS stage with diastolic dysfunction and LV remodeling. ⁴⁰ , ⁴¹ However, myocardial adaptation to increased afterload is unlikely to be the sole explanation since we observed these associations even in earlier stages of AS, when hemodynamic stenosis is negligible. Shared risk factors, including hypertension, may also contribute to these associations as traditional cardiovascular risk factors also increase the risk of AS. ⁴²

We observed a gradual decline in systolic function as reflected in GLS across Visit 5 AS Stages A through C at both Visit 5 and Visit 7. These findings are concordant with prior studies demonstrating progressive impairment in GLS from hemodynamically mild to severe AS. ⁴³ In contrast, prior studies of individuals with aortic sclerosis (Stage A AS) and normal LV geometry have not observed alterations in GLS despite impairments in diastolic indices. ⁴⁴ Notably, in our study, Stage A was associated with modestly higher LVEF and GLS than Stage 0, and AS stage progression was associated with a subtle increase in LVEF and GLS driven primarily by participants with incident AS stage. We suspect this is related to LV geometric changes noted between Stage 0 and Stage A AS, which include modestly higher LV wall thickness and LVMi, as LV concentric remodeling and hypertrophy may help maintain normal endocardial shortening and preserve LVEF. ³⁹ However, the magnitude of mean observed changes in GLS is small. Caution is therefore required in interpreting GLS as a marker of early functional change in subclinical AS, and the clinical relevance of these findings is uncertain.

Beyond hemodynamic severity, a new categorization scheme of patients with moderate to severe AS based on cardiac consequences of pressure overload ¹² was developed that aimed to inform therapeutic decision in patients with borderline indications for AV intervention. ¹² , ¹³ , ⁴⁵ While existing studies have minor variations in classification criteria, these studies demonstrated incremental prognostic value in terms of all‐cause mortality and composite endpoint of all‐cause mortality, stroke, heart failure, or myocardial infarction. ¹² , ¹³ , ⁴⁵ In this study, we investigated the utility of a modified categorization scheme to predict AS stage progression among participants with non‐clinically significant AS. The presence of at least 1 LV abnormality and of at least 1 LA abnormality—which notably included AF and moderate or greater mitral regurgitation in addition to structural enlargement—was associated with AS stage progression. The association of these extravalvular findings with the progression of the AS stage likely does not reflect a causal relationship, but rather shared risk factors. While neither LV diastolic function nor LVMi predicted symptomatic deterioration of 183 patients with moderate or severe asymptomatic AS in a previous small study, ³⁵ our analysis suggests potential relevance to the likelihood of AS stage progression in earlier AS stages (ie, Stages A and B). Understanding the relationship between AS stages, LV structure, and diastolic functions in the preclinical setting may provide insight into predicting AS progression by extending the role of extra‐AV categorization in early AS stages. Furthermore, extra‐AV categorization could be a readily available imaging biomarker for AS stages with the potential to enhance AS screening or to select individuals for trials aimed at preventing AS progression at its early stages. ⁴⁶ , ⁴⁷ However, additional validation studies using the original and modified categorization scheme are needed to prove this concept in different populations.

This study has several limitations. The number of participants with Stage C and D AS was small, as was the number with RV abnormalities, as this was a community‐based study and participants with prevalent cardiovascular disease were excluded. As a result, we had to combine AS Stage C and D even though these stages represent distinct clinical states with different management strategies and prognostic implications. However, this study design allowed us to more precisely describe the alterations in cardiac structure and function associated with sub‐severe AS stages. The definition of AS stages followed the AHA/ACC guidelines, but AV thickening was not available, and sclerosis was instead defined based on a high normal V _max. Approximately 30% of the participants who were alive at Visit 7 did not attend Visit 7, and non‐attendance was nonrandom which may bias our results toward the null. However, we used IPAW and incorporated clinically detected progression to minimize this bias in analyses related to AV progression. Some extra‐AV abnormalities used in the original classification were not available in our data, including tricuspid annular plane systolic excursion and RV strain, so we used an alternative measure of RV function with prognostic value, ⁴⁸ , ⁴⁹ but it may be less sensitive than RV strain in identifying pseudo‐normalized RV function. ⁵⁰ We also combined RV and PA abnormalities in 1 category. Together, these changes may limit comparability with prior studies and may obscure the independent contribution of RV abnormalities. Power was limited for analyses stratified by baseline AS stage. ⁵¹ Finally, despite adjusting for several risk factors, residual confounding cannot be excluded because of the study’s observational design.

CONCLUSIONS

Sub‐severe AS stages and progression in these stages over 6 years in late‐life are associated with greater increases in LV wall thickness, LVMi, and diastolic dysfunction. The presence of both LV and LA abnormalities, driven mainly by abnormal LVMi, E/e′, and LAVi, is associated with greater AS stage progression. When combined with the ACC/AHA staging approach, extra‐aortic valve cardiac abnormalities hold promise in predicting progression even in early AS stages.

Sources of Funding

The Atherosclerosis Risk in Communities study has been funded in whole or in part with Federal funds from the National Heart, Lung, and Blood Institute, National Institutes of Health, Department of Health and Human Services, under Contract numbers 75N92022D00001, 75N92022D00002, 75N92022D00003, 75N92022D00004, 75N92022D00005. This work was also supported by NIH/NHLBI grants R01HL135008, R01HL143224, R01HL150342, R01HL148218, R01HL160025, and K24HL152008 to Dr Shah.

Disclosures

Dr Shah reports research funds from Novartis through Brigham and Women’s Hospital. Dr Skali reports consulting fees from Astellas Inc. and research support from ABT Associates. Dr Shelbaya reports medical monitoring for Nephrology and Hematology Clinical Development of Alexion Pharmaceutical.

Supporting information

Tables S1–S5

Figures S1–S4