Prevalence of Aortic Valve Calcium and the Long-Term Risk of Incident Severe Aortic Stenosis

Abstract

BACKGROUND

Aortic valve calcification (AVC) is a principal mechanism underlying aortic stenosis (AS).

OBJECTIVES

This study sought to determine the prevalence of AVC and its association with the long-term risk for severe AS.

METHODS

Noncontrast cardiac computed tomography was performed among 6,814 participants free of known cardiovascular disease at MESA (Multi-Ethnic Study of Atherosclerosis) visit 1. AVC was quantified using the Agatston method, and normative age-, sex-, and race/ethnicity-specific AVC percentiles were derived. The adjudication of severe AS was performed via chart review of all hospital visits and supplemented with visit 6 echocardiographic data. The association between AVC and long-term incident severe AS was evaluated using multivariable Cox HRs.

RESULTS

AVC was present in 913 participants (13.4%). The probability of AVC >0 and AVC scores increased with age and were generally highest among men and White participants. In general, the probability of AVC >0 among women was equivalent to men of the same race/ethnicity who were approximately 10 years younger. Incident adjudicated severe AS occurred in 84 participants over a median follow-up of 16.7 years. Higher AVC scores were exponentially associated with the absolute risk and relative risk of severe AS with adjusted HRs of 12.9 (95% CI: 5.6-29.7), 76.4 (95% CI: 34.3-170.2), and 380.9 (95% CI: 169.7-855.0) for AVC groups 1 to 99, 100 to 299, and ≥300 compared with AVC = 0.

CONCLUSIONS

The probability of AVC >0 varied significantly by age, sex, and race/ethnicity. The risk of severe AS was exponentially higher with higher AVC scores, whereas AVC = 0 was associated with an extremely low long-term risk of severe AS. The measurement of AVC provides clinically relevant information to assess an individual’s long-term risk for severe AS. (J Am Coll Cardiol Img 2024;17:31–42) © 2024 by the American College of Cardiology Foundation.

Calcific aortic valve disease (CAVD) is the most common valvular disorder in the United States, occurring in 2.7 million individuals ≥75 years of age, nearly 700,000 of whom have severe aortic stenosis (AS).¹ As the percentage of older persons in the United States increases, it is estimated that nearly 1 million patients will have severe AS in 2025, and without aortic valve replacement (AVR) the 5-year survival is just 15%.²

CAVD is principally characterized by the underlying pathophysiologic process of aortic valve calcification (AVC), which begins in middle to older age and has a decades-long slowly indolent course characterized by progressive AVC.³ Trials targeting traditional cardiovascular risk factors to slow CAVD have had largely negative results to date, although treatment too late in the CAVD process and the inability to identify those at highest risk for progression have been serious limitations.

Little is known about personalized risk prediction for the long-term risk of severe AS. There is no accepted risk score or algorithm for the early identification of those at elevated long-term risk for severe AS. The current treatment approach for AS focuses nearly exclusively on determining the optimal timing for surgical or transcatheter AVR. Therefore, it is imperative to develop strategies that can identify patients with early CAVD at the highest risk for severe AS because these persons may be most likely to benefit from early and/or novel tailored treatments.⁴

Noncontrast cardiac computed tomography (CT), most commonly performed for coronary artery calcium (CAC) scoring, is uniquely positioned for quantifying early AVC because it provides a highly sensitive and reproducible measurement of AVC. However, while it is recommended that the presence of AVC should be quantified and reported, this often does not occur because of a paucity of information with regard to normative values and the proper clinical interpretation, especially for persons with a mild or moderate burden AVC. Additionally, although AVC has been correlated with short-term (2-3 years) hemodynamic progression in a small study of patients with mild to moderate AS, the association of AVC with long-term risk for severe AS is unknown.⁵

Therefore, in this community-based U.S. cohort free of known cardiovascular disease (CVD), we calculated reference age-, sex-, and race/ethnicity-based population percentiles—a clinical cornerstone for the interpretation of CAC scoring that has not been established for AVC.^6,7 We also examine the association of AVC with long-term risk for severe AS using a newly derived outcome of physician-adjudicated severe AS in MESA (Multi-Ethnic Study of Atherosclerosis).

METHODS

STUDY PARTICIPANTS.

A total of 6,812 participants from MESA had noncontrast cardiac CT performed at visit 1 (2000-2002) with AVC scoring. Participants 45 to 84 years of age who were free of clinically apparent CVD were recruited from 6 communities within the United States who self-identified their race/ethnicity as White, Black, Hispanic, or Chinese.⁸ The MESA protocol was approved by the Institutional Review Boards at the participating institutions, and written informed consent was given by all participants.

QUANTIFICATION OF AVC.

Calcified lesions of the aortic valve leaflets and those extending from the aortic valve into the aortic root were classified as AVC.⁹ Calcification occurring at the level of the aortic valve ring was classified as aortic ring calcification (ARC).¹⁰ The Agatston method was used to quantify AVC. AVC scores were not reported to participants or their care providers.

AVC was defined as absent (AVC = 0) or prevalent (AVC >0) and categorized by severity as 0, 1 to 99, 100 to 299, and ≥300 Agatston units (AU). ARC was scored separately from AVC, although distinguishing AVC from ARC can be challenging given their proximity to each other. Scans with AVC = 0 and ARC >0 (n = 46) were rescored for this analysis with minimal differences between the original and rescored values. Therefore, the original AVC scores were used for the present analyses to maintain consistency with prior MESA AVC publications.^9,11 Scans were processed and interpreted at a centralized reading center (Harbor University of California, Los Angeles, California, USA), and there was low intrareader and interscan variability of 4.4% and 9.7%, respectively, for AVC scoring.⁹

CALCULATION OF AORTIC VALVE PERCENTILES.

The AVC score distribution is 0 inflated and right skewed, but the distribution of positive values (AVC >0) normalizes with logarithmic transformation, which is the same general distribution as CAC.¹² Therefore, we applied the same statistical methods previously used to calculate population percentiles for CAC to calculate population percentiles for AVC.⁶ Using a locally weighted scatterplot smoothing regression, we performed nonparametric modeling of the mean logarithmically transformed AVC score as a function of age separately for each sex and race. The pooled residuals were calculated by subtracting the estimated value from each observed value on the log-transformed AVC distribution. These pooled residuals were then ranked, and percentiles from 1 to 100 were calculated. These values were then fitted to the age, sex, and race value, which provides a log-transformed estimated AVC percentile.¹³ Exponentially transforming the log-transformed AVC percentile gives an AVC percentile for positive AVC scores, which is then used to model the distribution of AVC as a function of positive AVC scores, taking into account the proportion with AVC = 0.

This method of formulating population percentiles is superior to calculating age-categorized sex and race estimates because it allows for the use of age as a continuous variable.⁶ In addition, higher precision is achieved by using the pooled residual distribution.⁶ This modeling technique is also superior to an age-categorized approach because it is not reliant on the assumption of a normalized distribution, which provides a more reliable estimate for the middle of the distribution rather than the tails/extremes. This is an important distinction because the higher percentiles (particularly ≥75th percentile) are of particular interest for AVC given the large proportion of participants with AVC = 0.

ADJUDICATION OF SEVERE AS.

MESA participants and/or their families were contacted every 9 to 12 months over follow-up regarding new hospital admissions, outpatient CVD diagnoses, or death. Copies of medical records for hospitalizations, outpatient CVD diagnoses, and death certificates were obtained to verify self-reported diagnoses. Previous reports from MESA indicate that medical records have been obtained for ≥95% of hospital events and outpatient CVD diagnostic encounters.¹⁴ The MESA coordinating center searched all hospital encounters for International Classification of Diseases (ICD) codes broadly related to aortic valve disease, rheumatic heart failure, aortic valvuloplasty, and AVR.

Participants’ follow-up data and associated medical records were available through December 31, 2017. A total of 187 participants had an encounter with an ICD code identifying it as a candidate event, and complete medical records were obtained for all these participants. The medical records were then independently reviewed by 2 cardiologists from the MESA CVD events adjudication committee, and any disagreements on event classification were discussed to reach a consensus on severe AS event classification. Severe AS was defined according to standard clinical criteria from: 1) echocardiography (eg, peak velocity ≥4.0 m/s, mean gradient ≥40 mm Hg, aortic valve area <1.0 cm², and so on); 2) AVR (surgical or transcatheter) for documented severe AS; 3) AVR for moderate AS when part of coronary artery bypass graft surgery; or 4) clinical documentation of severe AS diagnosis.

Medical chart review revealed 2 participants who had severe AS at baseline and 2 participants in whom AVC values were missing at baseline who were excluded from this analysis for a total of 77 participants with severe AS identified by medical chart review. We supplemented medical chart review records with data from echocardiography, which was performed among 92% of 3,303 participants who attended MESA visit 6 (September 2016 to June 2018). A review of this echocardiography data resulted in the identification of an additional 7 participants with severe AS who were not identified by the ICD code and medical chart review algorithm. This resulted in a total of 84 MESA participants identified with incident severe AS who were included in this analysis (Supplemental Figure 1). A bicuspid aortic valve was reported in 11 participants who developed severe AS.

OUTCOME ANALYSES.

Unadjusted incident rates were calculated per 1,000 person-years follow-up stratified by AVC group. We also performed Kaplan-Meier survival analyses stratified by AVC score and multivariable-adjusted Cox proportional hazards models to estimate HRs of long-term severe AS according to AVC. The Cox models included the covariates of age, sex, race/ethnicity, systolic blood pressure, diastolic blood pressure, hypertension medication use, total cholesterol, high-density lipoprotein cholesterol, lipid-lowering medication use, fasting glucose, diabetes, body mass index, smoking, and lipoprotein(a) (Lp[a]). We performed sensitivity analyses additionally adjusting for CAC, estimated glomerular filtration rate, income, and education. We also performed additional sensitivity analyses excluding participants with bicuspid aortic valve and/or potentially undiagnosed severe AS at baseline, which was defined as a baseline AVC score ≥1,200 for women (n = 2) or ≥2,000 for men (n = 7) or time to diagnosis <2 years (n = 2).¹⁵

In order to examine whether ARC or the combination of ARC and AVC were more strongly associated with long-term severe AS than AVC alone, we repeated the analysis using AVC + ARC and ARC by itself as the outcome instead of AVC alone. We also performed a sensitivity analysis for ARC by itself after excluding participants with AVC >0 at baseline. Additionally, we examined the association of CAC with the long-term hazard of severe AS.

RESULTS

The mean age of participants was 62.1 ± 10.2 years of age, and 53% were women. Overall, there were 913 participants (13.4%) with AVC >0. Participants with AVC >0 were approximately 10 years older on average than participants with AVC = 0 and had a higher burden of traditional CVD risk factors (Table 1). Among participants with AVC >0, 211 (23%) had AVC between 100 and 299 AU, and 114 (13%) had AVC ≥300 AU (Figure 1).

TABLE 1.

Participant Demographics Stratified by the Presence of AVC at Baseline

	AVC = 0 (n = 5,899)	AVC >0 (n = 911)	P Value
Age, y	60.9 ± 9.9	70.5 ± 8.1	<0.01
Women	54.9	40.0	<0.01
Race
White	37.4	45.3	<0.01
Black	28.1	25.5	0.09
Chinese	12.5	7.4	<0.01
Hispanic	22.0	21.8	0.93
Heart rate, beats/min	63.1 ± 9.6	63.6 ± 10.1	0.10
Body mass index, kg/m2	28.3 ± 5.6	28.5 ± 5.0	0.05
Systolic BP, mm Hg	125.3 ± 21.1	135.0 ± 22.1	<0.01
Diastolic BP, mm Hg	71.9 ± 10.3	72.3 ± 10.0	0.27
Hypertension	56.7	78.4	<0.01
Antihypertensive medication	34.4	55.3	<0.01
LDL-C, mg/dL	117.0 ± 31.0	118.7 ± 34.4	0.41
HDL-C, mg/dL	51.3 ± 14.9	48.9 ± 14.2	<0.01
Triglycerides, mg/dL	110 (77-159)	121 (84-172)	<0.01
Lipoprotein(a), mg/dL	17.4 (8-38.3)	19.7 (8.1-54.6)	<0.01
Hyperlipidemia	45.1	55.5	<0.01
Lipid-lowering medication	14.7	26.0	<0.01
Glucose, mg/dL	96.5 ± 29.1	102.7 ± 36.7	<0.01
Diabetes	11.5	19.9	<0.01
Current smoking	13.3	10.5	0.02
10-year ASCVD risk score, %	11.8 ± 11.9	24.7 ± 15.5	<0.01
Coronary artery calcium score	0 (0-50.5)	117 (10-455)	<0.01

Values are mean ± SD, %, or median (IQR).

ASCVD = atherosclerotic cardiovascular disease; AVC = aortic valve calcification; BP = blood pressure; HDL-C = high-density lipoprotein cholesterol; LDL-C = low-density lipoprotein cholesterol.

FIGURE 1. Severity and Prevalence of AVC at Baseline.

AVC = aortic valve calcification.

The probability of AVC >0 was higher with older age for both men and women in a curvilinear manner (Central Illustration). There was generally a low probability of AVC >0 for men and women in early middle age. Overall, the probability of AVC >0 was higher for men than women with larger differences between the sexes apparent at older ages. Women generally developed AVC about 10 years later in life than men.

CENTRAL ILLUSTRATION. Prevalence of AVC and Incident Severe AS: Probability of AVC by Age, Sex, and Race/Ethnicity and Survival Free From Severe AS by AVC Burden.

AS = aortic stenosis; AVC = aortic valve calcification.

There were significant differences between races/ethnicities with regard to the probability of AVC >0. White men and White women had a higher probability of AVC at any given age compared with participants of other races/ethnicities of the same sex. In addition, White men and Hispanic men had a substantially higher probability of AVC than White women and Hispanic women at any given age. The probability of AVC was slightly higher for Black men than Black women at any given age. Chinese participants had the lowest probability of AVC >0, which was similar for men and women.

The age-, sex-, and race/ethnicity-based percentile (eg, 75th, 90th, and 95th percentile) associated with a participant’s AVC score was higher for men compared with women within each age group with similar trends and associations between sexes and races/ethnicities as seen with the probability of AVC >0 (Figure 2). For example, among persons 70 to 74 years of age, the AVC score associated with the 95th percentile was 223 for White men compared with 71 for White women and 168 for Black men compared with 64 Black women (Tables 2 and 3).

FIGURE 2. Estimated Percentiles of AVC for Women and Men.

Estimated percentiles of AVC for women (A) and men (B) as a function of age and stratified by race/ethnicity. Abbreviation as in Figure 1.

TABLE 2.

Estimated AVC Percentiles for Women Stratified by Age and Race/Ethnicity

	Age, y
	44-49	50-54	55-59	60-64	65-69	70-74	75-79	80-84
White
75th	0	0	0	0	0	0	17	73
90th	0	0	0	0	17	48	99	226
95th	0	0	0	24	71	126	158	229
Black
75th	0	0	0	0	0	0	14	60
90th	0	0	0	0	17	49	96	186
95th	0	0	0	29	60	95	168	353
Chinese
75th	0	0	0	0	0	0	0	0
90th	0	0	0	0	0	19	27	63
95th	0	0	0	0	20	93	107	110
Hispanic
75th	0	0	0	0	0	0	0	29
90th	0	0	0	0	9	27	77	229
95th	0	0	0	35	65	92	217	425

Abbreviation as in Table 1.

TABLE 3.

Estimated AVC Percentiles for Men Stratified by Age and Race/Ethnicity

	Age, y
	44-49	50-54	55-59	60-64	65-69	70-74	75-79	80-84
White
75th	0	0	0	0	0	28	94	165
90th	0	0	2	28	84	173	259	371
95th	0	0	52	103	233	449	564	682
Black
75th	0	0	0	0	0	0	25	68
90th	0	0	0	9	51	104	154	184
95th	0	0	12	58	168	242	248	311
Chinese
75th	0	0	0	0	0	0	7	32
90th	0	0	0	0	5	20	51	91
95th	0	0	0	10	25	55	108	191
Hispanic
75th	0	0	0	0	11	39	57	93
90th	0	0	16	56	99	175	230	371
95th	0	10	72	167	354	503	673	924

Abbreviation as in Table 1.

The median overall follow-up time for the ascertainment of incident severe AS was 16.7 years. Participants diagnosed with severe AS were approximately 8 years older on average at baseline than participants who were not diagnosed with severe AS, but there was no statistically significant difference in baseline low-density lipoprotein cholesterol, lipid-lowering medication use, or Lp(a) (Table 4). The mean age at diagnosis of severe AS was 80.6 years of age, and the mean follow-up time from the first visit to a diagnosis of severe AS was 9.5 years. Among participants diagnosed with severe AS, 86% had prevalent AVC at baseline with a median baseline AVC score of 211 (Figure 3A). Among the 5,899 participants with AVC = 0 at baseline, 12 (0.2%) developed severe AS (Figure 3B) with a mean time to diagnosis of 14.5 years, and of those, 8 of 12 (67%) had ARC.

TABLE 4.

Baseline Demographics of Participants Who Developed and Who Do Not Develop Incident Severe AS

	No Severe AS (n = 6,726)	Severe AS (n = 84)	P Value
Age baseline, y	62.0 ± 10.2	70.1 ± 7.9	<0.01
Age at diagnosis of AS, y	NA	80.6 ± 8.4	NA
Follow-up time to diagnosis, y	NA	10.5 ± 4.6	NA
Women	53.0	44.1	0.10
Race
White	38.3	54.8	<0.01
Black	28.0	13.1	<0.01
Chinese	11.9	3.6	0.02
Hispanic	21.9	28.6	0.14
Systolic BP, mm Hg	126.5 ± 21.4	136.3 ± 23.9	<0.01
Antihypertensive medication	37.0	54.8	<0.01
LDL-C, mg/dL	117.2 ± 31.5	117.9 ± 31.2	0.99
Lipid-lowering medication	16.2	18.1	0.65
Lipoprotein(a), mg/dL	17.3 (7.5-40.5)	19.0 (8.0-50.0)	0.44
Diabetes	12.6	15.5	0.42
Baseline CAC score	0 (0-83.0)	121.2 (5.6-376.7)	<0.01
Baseline AVC score	0 (0-0)	210.6 (39.9-846.8)	<0.01

Values are mean ± SD, %, or median (IQR).

AS = aortic stenosis; CAC = coronary artery calcium; other abbreviations as in Table 1.

FIGURE 3. Prevalence of AVC and Incident Severe AS.

(A) The prevalence of baseline AVC among participants diagnosed with severe aortic stenosis (AS) and (B) the proportion of participants with baseline AVC = 0 diagnosed with severe AS. Abbreviation as in Figure 1.

There was an exponential increase in the rate of incident severe AS with higher baseline AVC (Figure 4). Approximately 10% of participants with AVC of 100 to 299 were diagnosed with severe AS with an event rate of 8.7 per 1,000 person-years, whereas 34% of participants with AVC ≥300 were diagnosed with severe AS with an event rate of 37.7 per 1,000 person-years. Participants with AVC = 0 had an extremely low incidence of severe AS at 0.1 per 1,000 person-years. At 15 years of follow-up, the estimated cumulative incidence of severe AS was approximately 25% among participants with AVC of 100 to 299 and approximately 50% among participants with AVC ≥300, whereas participants with AVC = 0 had an extremely low cumulative incidence of severe AS (Central Illustration). The adjusted HR of severe AS also increased exponentially with higher AVC, and compared with those with AVC = 0, the HR for severe AS was 12.9 (95% CI: 5.6-29.7), 76.4 (95% CI: 34.33-170.2), and 380.9 (95% CI: 169.7-855.0), respectively, for AVC groups 1 to 99, 100 to 299, and ≥300 (Table 5).

FIGURE 4. Incident Rate of Severe AS by AVC Groups.

Vertical lines denote 95% CIs for incident rate. Abbreviations as in Figures 1 and 3.

TABLE 5.

Association Between AVC and Long-Term Incidence of Severe AS

	AVC = 0 (n = 5,899)	AVC 1-99 (n = 586)	AVC 100-299 (n = 211)	AVC ≥300 (n = 114)
Model 1	Ref.	12.2 (5.4-27.8)	69.5 (32.3-149.7)	404.5 (189.1-865.4)
Model 2	Ref.	12.9 (5.6-29.7)	76.4 (34.33-170.2)	380.9 (169.7-855.0)

Values are 95% CI. Model 1: age, sex, and race/ethnicity. Model 2: model 1 + systolic BP, diastolic BP, hypertension medication, total cholesterol, HDL-C, lipid-lowering medication, fasting glucose, diabetes, body mass index, current smoking, and lipoprotein(a).

Ref. = reference; other abbreviations as in Tables 1 and 4.

Sensitivity analyses that were additionally adjusted for CAC, estimated glomerular filtration rate, income, and education had similar HR estimates to the primary results (Supplemental Table 1). Analyses excluding participants with potentially undiagnosed severe AS and/or bicuspid aortic valve had similar HR estimates for AVC values <300 AU, whereas among participants with AVC ≥300, the HR estimates were slightly attenuated but still highly statistically significant (HR: 222.9 [95% CI: 89.0-558.2]). CAC was significantly associated with the long-term risk for severe AS but with HR values that were substantially lower than the observed relationship between AVC and severe AS (Supplemental Table 2). The combination of AVC + ARC as a combined variable produced lower HR estimates for the risk of long-term severe AS than those observed for AVC alone (Supplemental Table 3). ARC by itself was significantly associated with the long-term hazard of severe AS, but the association was substantially lower compared with AVC and AVC + ARC even when examined among participants with AVC = 0 at baseline. The cumulative incidence of severe AS was also lower for AVC + ARC than that for AVC alone (Supplemental Figure 2). Among the 12 participants with AVC = 0 at baseline who developed incident severe AS, 2 had a bicuspid aortic valve, 6 had ARC at baseline, 1 had AVR as part of coronary artery bypass graft surgery, and the median follow-up time for diagnosis of severe AS was 15.3 years.

DISCUSSION

To the best of our knowledge, this is the first report to: 1) provide age, sex, and race/ethnicity reference AVC percentiles from a community-based sample of people free of CVD from the United States; and 2) examine the association of CT-measured AVC with the long-term risk for adjudicated severe AS. Our results provide the foundation for a reference tool highlighting the importance of age, sex, and race/ethnicity in the clinical interpretation of AVC scoring. They also provide a robust framework for the long-term risk stratification of severe AS and may facilitate the implementation of a new paradigm to the approach of preventing AS that emphasizes early detection using CT-based imaging. Furthermore, these results may inform both the clinical interpretation of AVC detection and the potential for enriched selection of participants for future clinical trials of CAVD slowing or prevention.

Reference age-, sex-, and race/ethnicity-based population percentiles, which are a clinical cornerstone for the interpretation of CAC scoring, have not previously been derived for AVC. In this context, our results provide a critical clinical reference to define the relative burden/severity of AVC. We found that the presence of any AVC is very uncommon in men before 60 years of age and in women before 65 years of age. Accordingly, the presence of any AVC in a middle-aged person should be regarded as abnormal and premature. Conversely, there is a high prevalence of AVC among persons in their late 70s and particularly those in their 80s. Among men in their 80s, the prevalence of AVC was approximately 60% for White men and about 40% for men of other races/ethnicities. The increased prevalence and severity of AVC in this age group correspond with a significantly increased prevalence of AS among persons >75 years of age. Therefore, it is essential to interpret AVC not only as an absolute number but also as a percentile because a low AVC score in a young or middle-aged individual has significantly different long-term implications for the risk of future AS compared with a low score in an older individual.⁶

Our results demonstrating a higher prevalence and burden of AVC among men compared with women and among White persons compared with other races/ethnicities are consistent with what has been observed for CAC. These similarities between the epidemiology of CAC and AVC are likely because the pathophysiology of early AVC has many shared risk factors with CVD.³ Indeed, there is a strong correlation between CAC and AVC; approximately 80% of persons with AVC also have CAC.¹¹

Prior studies examining AS have reported a significantly lower incidence of AS and AVR among Black compared with White persons.^16-19 The underlying reason for these differences is unclear because Black persons generally have higher Lp(a) levels compared with other races/ethnicities, and in a large study of >2.1 million patient records, there was no evidence of referral bias for severe AS for Black persons vs White persons.^16,20 Our findings of a lower prevalence and severity of AVC for Black persons compared with White persons may help to explain these observed race/ethnicity-related differences in the clinical profile of AS and are also generally in keeping with prior studies demonstrating that Black individuals have a lower prevalence and severity of CAC compared with White individuals.⁶ It is unclear if these race/ethnicity-related differences reflect true biological differences, and further research is needed to understand the underlying pathophysiology with proposed mechanisms of AVC including differences in bone-vascular axis metabolism, renal function, and/or competing risks of mortality.²¹

Beyond previous cross-sectional investigations, a limited number of studies have examined AVC and its short-term progression over a follow-up of between 3 and 5 years. For example, in a study of 149 patients with at least mild AS by echocardiography, the baseline AVC burden was a very strong predictor for the rate of AVC progression over the 2.9-year follow-up.⁵ In another study of 100 consecutive patients with AVC scoring and echocardiography, an AVC score >500 was associated with a high risk of incident severe AS at the 5-year follow-up.²² However, these relatively small studies of patients with clinically apparent AS lacked sufficient follow-up time and power to accurately characterize the long-term association of AVC with severe AS in detail.

We observed an exponentially higher relative hazard between AVC and the long-term risk of severe AS, which is an order of magnitude stronger than the observed association for traditional risk factors or CAC with atherosclerotic CVD or the association between Lp(a) with AS.^23-25 The stronger relationship between AVC and AS compared with the association between CAC and atherosclerotic CVD is likely because AVC is the primary underlying pathophysiology mechanism leading to aortic valvular dysfunction, whereas acute atherosclerotic CVD has a stochastic element (ie, plaque rupture and thrombosis) that is likely not fully encapsulated by CAC. This exceedingly strong association between AVC and severe AS establishes that AVC is unequivocally the most important clinically accessible risk marker for the long-term prediction of severe AS.

The extremely protective association of AVC = 0 with the long-term risk of severe AS is an equally important prognostic finding in our understanding of the long-term risk for severe AS. It may have significant implications for future trials on the prevention of severe AS because these participants can be expected to have an extremely low event rate and a very low likelihood of benefit from any treatment, especially over the typical 3- to 5-year time period of a randomized clinical trial. Conversely, persons with an AVC score >100 had a much higher relative and absolute rate of long-term severe AS and would likely be ideal candidates for future trials examining novel treatment strategies or therapeutic targets. Indeed, although clinical trials to slow or prevent severe AS have had negative results to date, this has in large part been attributed to the inclusion of participants whose degree of AS is too advanced to benefit from lipid-lowering treatment. However, a subgroup analysis from the SEAS (Intensive Lipid Lowering with Simvastatin and Ezetimibe in AS) trial demonstrated a reduction in severe AS for participants with echocardiographically graded mild AS and low-density lipoprotein cholesterol ≥155 mg/dL, suggesting that persons in the earlier stages of AS who have a high atherosclerotic lipid burden may benefit from traditional lipid-lowering strategies.²⁶

Another important observation from this study is that the measurement of ARC did not improve risk stratification for severe AS. Over a median follow-up of 16.7 years, only 0.2% of participants with AVC = 0 developed severe AS. Although 8 of the 12 participants with AVC = 0 who developed severe AS had ARC, we found that the combination of AVC + ARC had a weaker association with the long-term risk for severe AS, even among participants with AVC = 0 at baseline. The weaker observed association with the addition of ARC to AVC is likely because noncontrast cardiac CT scans are reconstructed with a 3-mm slice thickness, and it can be difficult to distinguish calcification of the aortic ring vs the aortic root, which may lead to misclassification of whether or not the calcification is valvular/annular. In addition, ARC likely causes less hemodynamic obstruction compared with AVC. Accordingly, these findings support the measurement and reporting of only AVC for the long-term risk stratification of severe AS along with underscoring the importance for CT readers to carefully differentiate AVC and ARC, with potential for repeating the study in short-term follow-up if AVC = 0 and ARC >0.

STUDY LIMITATIONS.

Limitations of the present study include that all MESA participants were without known CVD at baseline. Therefore, participants in this study are somewhat healthier than the general U.S. population, some of whom have established CVD and may be at risk for AS. However, CAC percentiles calculated from MESA are recommended as a core criteria by the American Heart Association and the American College of Cardiology to interpret CAC scores in primary prevention, and our approach to AVC percentile calculation is identical to that for CAC percentiles and consistent with those from other large cohort studies.^6,7,27 Additionally, the total number of severe AS outcomes is relatively small, which results in reduced precision of the HR estimate and wider 95% CIs. However, the lower limit of the 95% CIs still shows an extremely strong association between AVC and severe AS. This relatively small number of severe AS outcomes also limits the ability to perform subgroup analyses. In addition, the MESA adjudication process relied on ICD codes from hospital encounters, which could miss some cases of AS. However, our emphasis on the outcome of severe AS protects against ascertainment bias because the majority of individuals with severe AS are clinically symptomatic and much less likely to not have an associated ICD code. Although we do not have baseline data on the prevalence of bicuspid aortic valve, at MESA visit 6 only 3 of 3,032 participants with echocardiography data had a reported bicuspid aortic valve. The strengths of this investigation include the use of noncontrast cardiac CT measurement of AVC, which is a highly sensitive and reproducible method of quantifying AVC, and the long follow-up period (median = 16.7 years). In addition, formal adjudication of severe AS was performed using medical records reviewed by 2 physicians with expertise in CVD event adjudication. Importantly, AVC measurement was not reported to participants or their clinicians and therefore did not alter the natural history of disease progression to severe AS.

CONCLUSIONS

These results provide the first age-, sex-, and race/ethnicity-specific reference values for AVC, which will help inform cutoffs for defining an elevated AVC score and provide an important framework for the clinical interpretation of AVC scores. These results also demonstrate that AVC is very strongly associated with the long-term risk for adjudicated severe AS and that the risk increases exponentially with higher AVC scores. Of similar importance, AVC = 0 is associated with an exceptionally low long-term risk for severe AS. Accordingly, these findings provide added support for the routine reporting of AVC in clinical practice, which delivers clinically relevant information for an individual’s long-term risk for severe AS. AVC appears to be a critically important phenotype of early valve disease and as such may also be an important tool to identify high-risk patients and enrich clinical trials with individuals who may benefit from early or novel treatment strategies to slow or prevent the development of severe AS.

PERSPECTIVES.

COMPETENCY IN MEDICAL KNOWLEDGE:

These results provide normative age-, sex-, and race/ethnicity-specific aortic valve calcium percentiles from a community-based sample of people free of CVD from the United States. This is the first study to show the association between aortic valve calcium measured from noncontrast cardiac CT scans with the long-term risk for adjudicated severe AS. Higher aortic valve calcium scores are associated with an exponentially increased long-term risk for severe AS, whereas AVC = 0 is associated with an extremely low long-term risk of severe AS. This study will help inform cutoffs for defining an elevated aortic valve calcium score. These results will help to identify persons at increased long-term risk for AS. Aortic valve calcium is an important early phenotype that should be considered for clinical trials aimed at testing novel strategies for preventing or slowing the development of severe AS.

TRANSLATIONAL OUTLOOK:

This is the first to study provide information on how to interpret an individual’s aortic valve calcium score and their long-term risk for severe AS. These results help to identify persons at increased risk to develop severe AS and should be considered to identify persons most likely to benefit from clinical trial interventions targeting novel treatment strategies or novel therapeutic agents.

ABBREVIATIONS AND ACRONYMS

ARC

aortic ring calcium

AS

aortic stenosis

AVC

aortic valve calcification

AVR

aortic valve replacement

CAC

coronary artery calcium

CAVD

calcific aortic valve disease

CT

computed tomography

CVD

cardiovascular disease

Lp(a)

lipoprotein(a)