OUP user menu

★ Editor's choice ★

Determinants of aortic sclerosis progression: implications regarding impairment of nitric oxide signalling and potential therapeutics

Aaron L. Sverdlov, Doan T.M. Ngo, Wai P.A. Chan, Yuliy Y. Chirkov, Bernard J. Gersh, John J. McNeil, John D. Horowitz
DOI: http://dx.doi.org/10.1093/eurheartj/ehs171 2419-2425 First published online: 6 July 2012


Aims Aortic valve stenosis (AS) and its precursor, aortic valve sclerosis (ASc), occur frequently in Western populations. Investigations to retard the progression of AS using statins have been unsuccessful. Development of ASc in humans is associated with increased aortic valve backscatter (AVBS) and poor tissue nitric oxide (NO) responsiveness. In an animal model, ramipril retarded AS/ASc development. We have now set out to identify factors associated with the progression of ASc in humans.

Methods and results At baseline and after 4 years, 204 randomly selected subjects (age 63 ± 6 years at study entry) underwent echocardiography with the determination of AVBS values, measurements of platelet NO responsiveness, plasma asymmetric dimethylarginine concentrations, lipid profile, high-sensitivity-C-reactive protein, routine biochemistry, and 25-hydroxy-vitamin D levels. During the study period, 68% of subjects had detectable AVBS progression. On multivariate analysis, higher calcium concentrations (β = 0.22; P = 0.004), poor platelet NO responsiveness (β = 0.18; P = 0.018), and increased arterial stiffness (β = 0.15; P = 0.044) were independent predictors of disease progression. The use of angiotensin-converting enzyme-inhibitors/angiotensin II receptor blockers (ACE-I/ARB) predicted the lack of disease progression (assessed categorically) in the overall cohort and in those without ASc at baseline (n = 159) (β = 0.8; P = 0.025 and β = 1.3; P = 0.001, respectively). No conventional coronary risk factors were associated with disease progression.

Conclusion This study of early aortic valve disease (i) demonstrates that disease progression occurs in the majority of the normal ageing population over a 4-year period; (ii) provides evidence of the importance of the NO signalling cascade in disease development and progression; and (iii) provides additional data linking ACE-I/ARB use with the retardation of ASc.

  • Aortic valve sclerosis/calcification
  • Nitric oxide
  • Endothelial function
  • ACE-inhibitors/angiotensin receptor blockers
  • Ageing
  • Inflammation


Aortic valve stenosis (AS) is now the most common valve disease in the Western world and its prevalence and incidence are rising.1,2 In general, AS reflects a progressive increase in calcium deposition within the aortic valve, leading to increased stiffness and progressive narrowing of the valve. The earliest clinically detectable stage of this process, aortic valve sclerosis (ASc), reflects abnormal aortic valve morphology in the absence of haemodynamic obstruction, but may progress to AS.3 A number of population studies have suggested that ∼25% of individuals aged >65 years have ASc, increasing to ∼50% at age > 80, despite a far lower prevalence of advanced AS among such populations.2,4,5

The development of ASc carries a number of pathophysiological and clinical implications. In two investigations,4,6 the presence of ASc was found to represent an independent marker of incremental risk of coronary events. Furthermore, we have demonstrated that ASc is also associated with impairment of platelet nitric oxide (NO) signalling.7 This provides a possible mechanism for the observed association with coronary event risk. In vitro studies suggest that the integrity of NO signalling is critical to the preservation of valve function8 and the prevention of calcification.9 This impairment of NO signalling may contribute to the development of ASc/AS.

The question of the rate of development of ASc has been addressed in few prospective studies to date. A clinical investigation in 50-year-old Norwegian males showed that only a small minority of subjects with soft heart murmurs progressed to aortic valve replacement over a 35-year follow-up period.10 Novaro et al.11 utilized serial echocardiography in a population aged >65 years and found progression in ∼44% of subjects over the 5-year follow-up period. Although male gender, advancing age, and African-American ethnicity were independently related to progression, there was no association with a history of diabetes, hypertension, or tobacco use, nor the presence of coronary heart disease or renal insufficiency. Endothelial and platelet function or other measures of inflammation/fibrosis were not evaluated in this study. Furthermore, it must be emphasized that the assessment of valve disease in this study was categorical rather than quantitative.

The application of ultrasonic backscatter to the aortic valve (AVBS), first utilized for the quantification of ASc and mild AS in our previous clinical investigations,12 represents a highly reproducible means for the evaluation of valve echogenicity. As such, it has facilitated both clinical7,13,14 and basic1517 studies and represents a technique of choice for ASc quantification.18,19

In the current study, we have utilized serial estimation of AVBS over a 4-year period to:

  1. determine the prevalence of detectable progression of aortic valve disease over that period;

  2. identify correlates of the extent of progression; and

  3. identify correlates of any progression.

The results therefore provide further insights into the factors modulating the earliest stages of aortic valve disease in the general population.


Study subjects

The North Western Adelaide Health Study (NWAHS)20 cohort participants were recruited by telephone to conduct the interviews, and the Electronic White Pages was used as the sampling frame. Within each household, the person who had his/her birthday last and was ≥18 years was selected for interview and invited to attend the clinic for a biomedical examination. To minimize potential bias due to differing probabilities of selection in the sample, the data were weighted by region (western and northern health regions), age group, and sex.

An initial cohort of 253 subjects was recruited prospectively as a substudy of the NWAHS.20 Selection criteria and baseline patient characteristics for this substudy have been published previously.7 Study personnel attempted to recall every subject 4 years after the initial evaluation to invite them to participate in the follow-up stage. A total of 204 subjects were recalled: of the remaining 49 subjects, 12 were lost to follow-up, 7 were deceased (5 due to cancers; for the remaining 2, the cause of death could not be ascertained), 6 had developed terminal illness or were receiving chemotherapy, and 24 declined to participate in the follow-up, citing personal reasons. The study was approved by the Ethics of Human Research Committee of The Queen Elizabeth Hospital.


Doppler echocardiography

Comprehensive transthoracic echocardiography was performed for all subjects with a commercially available system [Vivid 7 (GE Vingmed, Horten, Norway), with a 2.5 MHz phased array probe]. M-mode and two-dimensional (2D) echocardiograms with Doppler analysis were obtained for all subjects. Left ventricular (LV) diameters and wall thicknesses were measured from 2D-guided M-mode echocardiography. Mean and peak pressure gradients across the aortic valve were calculated with the modified Bernoulli equation, with continuous-wave Doppler recordings from the highest velocity available from any view. The aortic valve area was computed with the continuity equation with standard methods.

Ultrasound backscatter data analysis

Aortic valve backscatter values were obtained for all subjects with methods previously published.12 Briefly, 2D ultrasonic backscatter images of the aortic valves were obtained from standard parasternal long-axis views over three cardiac cycles with a zoom of 8 cm. Three consecutive scans were acquired for each study subject. Backscatter values from the blood pool in the LV outflow tract and aortic root were used as reference values. Calibrated backscatter values were obtained by subtracting the average blood pool value from the averaged backscatter values obtained from the aortic valves. Aortic valve sclerosis was quantitated on the basis of AVBS scores, and AVBS ≥ 16 dB was used as a definition of the presence of ASc.12 In the current series, inter-observer variability was 2.7 ± 1.8 dB (9.8 ± 7%). The progression of valve disorganization/ASc was defined as any positive change in AVBS scores. Valve morphology was categorized on the basis of visual assessment, as previously described.4,21

Comparisons between baseline and follow-up echocardiographic data were utilized in order to evaluate changes in valve structure and function.

Biochemical and physiological correlates

In all cases, parameters measured at study entry7 were utilized as potential correlates of subsequent disease progression.

These can be summarized as follows:

  1. Physiological measures

    1. Resting augmentation index, a measure of arterial stiffness, was determined by applanation tonometry (ShygmoCor, Atcor Medical, Sydney, Australia).22

    2. Platelet responsiveness to the NO donor sodium nitroprusside (SNP) was assessed using whole-blood aggregometry with a dual channel impedence aggregometer (model 560, Chrono-Log, Havertown, PA, USA).23

  2. Biochemical measures

    1. Plasma concentrations of asymmetric dimethylarginine (ADMA), a marker and mediator of endothelial dysfunction,24 were determined by high-performance liquid chromatography with the derivitization reagent AccQ-Fluor (Waters, Milford, MA, USA) after solid-phase extraction, as previously described.25

    2. High-sensitivity C-reactive protein (hs-C-reactive protein) concentrations, as a marker of systemic inflammatory activation, lipid profile, creatinine, serum calcium levels, and 1,25 dihydroxy cholecalciferol (vitamin D levels) were measured by a 125I radioimmunoassay (Immunodiagnostic Systems Ltd, Bolden, UK).

    3. C-terminal telopeptide of collagen type 1 and N-terminal peptide of procollagen I concentrations were measured as markers of collagen homeostasis.

    4. Creatinine clearance (CrCl) was calculated according to the Cockcroft–Gault equation and indexed for body surface area with the Dubois and Dubois formula.

Statistical analyses

All data are expressed as mean ± SD unless otherwise stated. All t-tests were two-sided. Normal distribution was tested for all continuous variables, and skewed data were normalized by either log or square root transformation. Comparisons between groups for normally distributed data were performed with non-paired t-tests, and comparisons for non-parametric data were made with the Mann–Whitney test. Comparisons between baseline and end-of-study continuous data parameters were performed with paired t-tests and tests for differential rates of progression and proportions were performed using χ2 tests. Correlations between transformed, continuous non-parametric data were made with linear regression. Correlation between baseline AVBS scores and change in scores was established. Determinants of the change in AVBS scores were evaluated utilizing univariate and then multivariate analyses. Variables selected for all multivariate backward regression analyses were on the basis of univariate significance (P ≤ 0.2). Included baseline variables to predict increase in AVBS scores were total cholesterol, calcium, ADMA, and hs-C-reactive protein concentrations; augmentation index, history of dyslipidaemia, and platelet NO responsiveness.

Changes in AVBS scores were also evaluated as a binary variable in a separate set of univariate and then multivariate analyses. Disease progression for these analyses was defined as any increase in AVBS scores. Variables selected for all binary logistic backward regression analyses were on the basis of univariate significance (P ≤ 0.2). Binary logistic backward regression analysis was performed to assess independent predictors of increase in AVBS scores in the entire cohort. Included baseline variables to predict increase in AVBS scores were history of hypercholesterolaemia; history of the use of angiotensin-converting enzyme-inhibitors/angiotensin II receptor blockers (ACE-I/ARB); calcium, LDL, total cholesterol, and hs-C-reactive protein concentrations; and body mass index (BMI). Furthermore, binary logistic backward regression analysis was also performed to assess independent predictors of increase in AVBS scores in those subjects without ASc at baseline. The baseline variables included in this analysis were history of statin use, history of the use of ACE-I/ARBs, history of hypertension, history of coronary events, calcium-phosphate product, vitamin D, N-terminal peptide of procollagen I, and hs-C-reactive protein concentrations. All analyses were performed with the SPSS version 17 software (SPSS, Chicago, IL, USA), and a P-value of <0.05 was considered to be statistically significant.


Subject characteristics

Subject characteristics both at baseline and at the 4-year follow-up are shown in Table 1. No subject had bicuspid aortic valve. Of the subjects receiving ACE-I/ARB at baseline, 52% were receiving ACE-I. Furthermore, during the course of study, in only three of the subjects was ACE-I/ARB therapy discontinued. Over the study period, there were significant increases in the proportion of subjects diagnosed with hypertension, dyslipidaemia, and diabetes, as well as a decrease in that of active smokers. More people were treated with ACE-I/ARB at the end of the study. Although the precise indications for ACE-I/ARB therapy were not ascertained, there was a strong correlation between the presence of systemic hypertension and the use of these agents (χ2: P < 0.001). At baseline, there was no significant difference in any of these parameters among subjects with and without ASc.7

View this table:
Table 1

Patient characteristics (n = 204)

Parameter (n = 204)Baseline (%)End of study (%)P-value
Age63 ± 6 years67 ± 6 years
Gender42.4% male
Diabetes24 (12%)33 (16%)0.013
Hypertension85 (42%)106 (52%)<0.001
Dyslipidaemia118 (58%)137 (67.5%)0.004
Smoking28 (14%)13 (6.4%)<0.01
Coronary disease24 (12%)28 (14%)0.26
Statin therapy65 (32%)71 (35%)0.24
ACE-I/ARB therapy69 (34%)83 (41%)0.008
BMI (kg/m2)28.2 ± 5.228.2 ± 5.20.37
Beta-blocker therapy20 (9.8%)22 (10.8%)0.53
Non-dihydropyridine calcium channel blockers7 (3.5%)10 (5.1%)0.1
Dihydropyridine calcium channel blockers16 (7.9%)19 (9.7%)0.2
Aspirin therapy48 (23.8%)48 (23.8%)0.87
Clopidogrel therapy9 (4.5%)11 (5.6%)0.48
  • ACE-I/ARB, angiotensin-converting enzyme-inhibitors/angiotensin receptor blockers; BMI, body mass index.

Baseline biochemical data have previously been published for this cohort.7 Table 2 summarizes both baseline and end-of-study biochemical data. Total cholesterol, HDL, C-terminal telopeptide of collagen type 1, and calcium levels, as well as CrCl increased marginally over the study period. As regards creatinine, this cohort had well-preserved renal function—no subject had CrCl <30 mL/min. Furthermore, it is possible that the apparent increase in CrCl resulted from a change in methodology of creatinine assay over the study period from Olympus AU5400 Chemistry-Immuno Analyzer (Olympus America, Melville, NY, USA) to Advia 2400 Chemistry System (Siemens Healthcare Diagnostics, Deerfield, IL, USA) and associated reagent kits. There were no significant changes in LDL, hs-C-reactive protein, or vitamin D levels.

View this table:
Table 2

Biochemical profile

Parameter (n = 204)BaselineEnd of studyP-value
Total cholesterol (mmol/L)4.9 ± 0.95 ± 1.10.02
LDL (mmol/L)2.8 ± 0.82.9 ± 10.75
HDL (mmol/L)1.3 ± 0.31.5 ± 0.4<0.001
Calcium level (mmol/L)2.2 ± 0.12.3 ± 0.1<0.001
Vitamin D level (mmol/L)72 ± 23.174.7 ± 26.60.29
CrCl (mL/min/1.73 m2)92 ± 21.698 ± 28.6<0.001
hs-C-reactive protein (mmol/L)3.5 ± 3.73.1 ± 3.80.14
CTx (median) (pg/mL)242 ± 143283 ± 1450.001
P1NP (median) (mcg/L)40.5 ± 19.843.4 ± 50.60.45
  • LDL, low-density lipoprotein; HDL, high-density lipoprotein; CrCl, creatinine clearance; CTx, C-terminal telopeptide of collagen type 1; P1NP, N-terminal peptide of procollagen I.

Baseline augmentation index, ADMA, and platelet NO responsiveness data have been published previously.7 In brief, baseline augmentation index was 27.2 ± 8.3% (normal range 15 ± 16%26), ADMA concentrations were 0.52 ± 0.08 μM (normal range 0.5 ± 0.08 μM25), and baseline platelet NO responsiveness was 33.9 ± 26.9% (normal range 54 ± 34%).

Changes in aortic valve backscatter

The distribution of AVBS scores at baseline and at the follow-up is summarized in Figure 1. Overall, AVBS increased markedly over the study period, with mean increase from 12.2 ± 4.4 to 14.2 ± 4.9 dB (P < 0.001). The implication of this change was that 68% of subjects had some disease progression, and while at baseline 17.6% of subjects had AVBS ≥ 16 dB, constituting the equivalent criterion of the presence of ASc,12 34.8% had ASc at the follow-up (P < 0.001). Progression occurred in 73.8% (n = 118) and 47.8% (n = 21) of subjects without and with ASc at study entry, respectively. There were 48 cases of de novo ASc development. Aortic valve backscatter changes are compared with those in conventional echocardiographic parameters (which also indicate some disease progression) in Supplementary material online, Table S1. No subject developed haemodynamically significant AS.

Figure 1

Mean change in AVBS over a 4-year period (P < 0.001).

Our previous studies have suggested that there is finite maximum for AVBS: for example, even in advanced AS, AVBS scores do not exceed 28 dB. In the current data set, the extent of increase in AVBS varied inversely with baseline AVBS (Figure 2; P < 0.001), providing further evidence of the existence of maximum values as well as the presence of the phenomenon known as regression to the mean.27 For these reasons, the baseline AVBS scores were not entered in the multivariate model, and progression data were also analysed as a dichotomous variable.

Figure 2

Relationship of change in AVBS with baseline AVBS values (P < 0.001).

Progression of aortic valve backscatter within the entire cohort

Aortic valve backscatter change as a continuous variable

Univariate correlates of increasing AVBS scores are presented in Table 3. These analyses raised the possibility of associations between dyslipidaemia, high hs-C-reactive protein, impaired NO responsiveness, and elevated plasma calcium concentrations with increasing AVBS.

View this table:
Table 3

Univariate correlates of increasing aortic valve backscatter in the entire cohort (β coefficients shown for continuous variable with P ≤ 0.2)

Baseline parameter (n = 204)β-CoefficientP-value
Age (years)0.83
Total cholesterol concentration (mmol/L)0.120.11
LDL concentration (mmol/L)0.120.103
HDL concentration (mmol/L)0.10.175
CrCl (mL/min)0.4
hs-C-reactive protein concentration (mmol/L)0.140.051
Vitamin D concentration (mmol/L)0.62
BMI (kg/m2)0.24
Augmentation index (%)0.090.2
Calcium concentration (mmol/L)0.140.045
ADMA concentrations (μM)0.10.17
Platelet SNP responsiveness (%)−0.1790.018
Systolic blood pressure (mmHg)0.85
CTx concentration (pg/mL)0.97
P1NP concentration (mcg/L)0.39
Statin therapy0.7
History of dyslipidaemia0.053
ACE-I/ARB therapy0.7
History of diabetes mellitus0.75
History of hypertension0.65
History of coronary disease0.7
  • LDL, low-density lipoprotein; HDL, high-density lipoprotein; CrCl, creatinine clearance; BMI, body mass index; ACE-I/ARB, angiotensin-converting enzyme-inhibitors/angiotensin receptor blockers; ADMA, asymmetric dimethylarginine; SNP, sodium nitroprusside; CTx, C-terminal telopeptide of collagen type 1; P1NP, N-terminal peptide of procollagen I.

On stepwise backward multiple linear regression analysis (Table 4), the independent predictors of increasing AVBS scores were lower platelet NO responsiveness, higher augmentation index, and plasma calcium concentrations.

View this table:
Table 4

Multivariate (backward stepwise multiple linear regression) correlates of increasing aortic valve backscatter

Platelet NO responsiveness (%)−0.1790.018
Plasma calcium concentration (mmol/L)0.2160.004
Augmentation index (%)0.1520.044
hs-C-reactive protein concentration (mmol/L)−0.1410.062

Aortic valve backscatter increase as a dichotomous variable

Univariate correlates of increase in AVBS scores are summarized in Table 5. None of the parameters tested were significantly related to disease progression on univariate analysis. However, the subjects in whom no progression occurred tended to be on ACE-I/ARB throughout the period of the study, had lower concentrations of total cholesterol, LDL, hs-C-reactive protein, and calcium, but higher BMI; and were less likely to have history of dyslipidaemia.

View this table:
Table 5

Univariate correlates of increase in aortic valve backscatter in the entire cohort

Baseline parameter (n = 204)P-valueOdds ratio95% confidence intervals
Age (years)0.390.980.93–1.03
Total cholesterol concentration (mmol/L)–1.80
LDL concentration (mmol/L)–1.90
HDL concentration (mmol/L)0.391.40.59–3.52
CrCl (mL/min)0.621.00.99–1.02
hs-C-reactive protein concentration (mmol/L)0.190.630.32–1.25
Vitamin D concentration (mmol/L)0.931.00.99–1.01
BMI (kg/m2)0.190.950.90–1.02
Augmentation index (%)0.691.010.97–1.04
Statin therapy0.690.880.46–1.67
History of dyslipidaemia0.21.490.81–2.73
ACE-I/ARB therapy0.080.570.31–1.07
History of diabetes mellitus0.711.190.47–3.04
History of hypertension0.280.720.39–1.31
Calcium concentration (mmol/L)–37.0
ADMA concentrations (μM)0.732.060.04–122.5
Platelet SNP responsiveness (%)0.481.210.94–2.36
CTx concentration (pg/mL)0.520.690.23–2.09
P1NP concentration (mcg/L)0.891.110.26–4.75
  • LDL, low-density lipoprotein; HDL, high-density lipoprotein; CrCl, creatinine clearance; BMI, body mass index; ACE-I/ARB, angiotensin-converting enzyme-inhibitors/angiotensin receptor blockers; ADMA, asymmetric dimethylarginine; SNP, sodium nitroprusside; CTx, C-terminal telopeptide of collagen type 1; P1NP, N-terminal peptide of procollagen I.

However, on multivariate (backward binary logistic regression; Table 6, ‘Entire cohort’) analysis, there was a negative relationship between disease progression and therapy with ACE-I/ARB (P = 0.025; β = 0.8).

View this table:
Table 6

Multivariate (binary multiple logistic regression analysis) correlates of increase in aortic valve backscatter in the entire cohort and in subjects without aortic valve sclerosis at baseline

ParameterP-valueβ-CoefficientOdds ratio95% confidence intervals
Entire cohort
 ACE-I/ARB treatment0.025−0.770.470.24–0.89
 History of hypercholesterolaemia0.1540.4871.70.9–3.28
 Baseline calcium concentrations (mmol/L)–34.8
Subjects without ASc at baseline
 ACE-I/ARB treatment0.001−–0.587
 hs-C-reactive protein concentration (mmol/L)0.020.972.61.17–6
 Vitamin D concentration (mmol/L)0.053−0.020.980.97–1.0
  • ACE-I/ARB, angiotensin-converting enzyme-inhibitors/angiotensin receptor blockers.

In view of the recently published results of Wakabayashi et al.,28 suggesting that ACE-I, but not ARB, might retard the progression of ASc, post hoc comparison between effects of ACE-I vs. ARB on progression rates was performed, which revealed no difference between these agents.

Progression of aortic valve backscatter in subjects without aortic valve sclerosis at baseline

In order to identify factors responsible for very early development of valvular disease, the cohort of subjects (n = 160) in whom baseline AVBS was <16 dB was analysed separately.

Univariate correlates of increases in AVBS scores in this cohort are listed in Supplementary material online, Table S2. Of note, there was a significant negative correlation between disease progression and therapy with ACE-I/ARB (P = 0.008). None of the other parameters tested were significantly related to disease progression on univariate analysis.

However, on multivariate (backwards binary logistic regression; Table 6, ‘Subjects without ASc at baseline’) analysis, there were negative relationships between disease progression and therapy with ACE-I/ARB (P = 0.001; β = 1.3) and lower hs-C-reactive protein (P = 0.02; β = 0.97) as well as a borderline negative relationship with higher vitamin D levels (P = 0.053; β = 0.17).


The current investigation represents the first evaluation of disease progression in humans, utilizing the highly reproducible technique of serial evaluation of AVBS. We have previously demonstrated the utility of this technique to categorize valve structure at baseline evaluation7 and to quantitate disease progression in a rabbit model.16 The current results indicate first and foremost that there is substantial progression of aortic valve disease in this cohort over a mean period of 4 years: increases in AVBS occurred in 68% of the subjects. The only comparable study in the literature11 utilized categorical rather than quantitative methodology: over a mean period of 5 years, new ASc developed in 44% of subjects.

It was found that there was an inverse relationship between the extent of AVBS progression and baseline AVBS. Although this would reflect, in part, ‘regression to the mean’,27 it also is consistent with the presence of a maximal value for AVBS, as suggested by our previous publication in AS.12 Therefore, we analysed our data both as regards the extent of increase of AVBS and also categorically, as the presence/absence of AVBS increase. Finally, in view of differing results, we explored the issue of de novo ASc development.

The principal determination of correlates of AVBS progression was performed via continuous data analysis. On this basis, elevated plasma calcium concentrations and impaired tissue NO responsiveness emerged as the strongest correlates of progression, with elevated augmentation index and hs-C-reactive protein as correlates of borderline significance.

Increased calcium levels have been previously linked with increased vascular and extravascular calcification.29 Furthermore, there have been a number of reports of aortic valve disease in association with hyperparathyroidism.30 Given that calcium deposition in arteries and valves may potentially be driven by similar process,31 these results should therefore stimulate interest in potential limitation of ASc/AS progression via the modulation of calcium uptake into extra-osseous tissues.

The finding that tissue resistance to NO, measured via platelet response to SNP, predicts ASc progression is consistent both with the known physiological role of NO in valve homeostasis32 and with the findings of our previous studies both in the rabbit model15,16 and in humans.7 In summary, NO is produced by both valve endothelium and matrix and inhibits calcific nodule formation in a cell culture model.9 The prevention of AS development in a rabbit model by ramipril was associated with the preservation of NO signalling,16 whereas platelet responsiveness to NO was inversely associated with AVBS values in a baseline evaluation of this cohort.7 The current data therefore provide further confirmation that ASc development is impeded by NO—this in turn provides a mechanistic basis for the known association between ASc and coronary risk.4

As regards the association between elevated augmentation index and ASc progression, augmentation index represents a measure of apparent arterial stiffness which has both fixed and variable components, the latter modulated largely by NO.33 The current data do not permit delineation of the bases for the association, which might include commonality of biochemical processes (e.g. including calcium uptake and/or NO signalling) and also the physical effect of increased wave reflection, associated with vascular rigidity, on the aortic valve.

As regards the correlation of hs-C-reactive protein levels with disease progression, this study is the first prospective population study to record such an association. Indeed, Novaro et al.11 found no significant association between these. From a theoretical point of view, a potential role for inflammatory activation in the progression of ASc is hardly surprising: ASc lesions include inflammatory infiltrates,3,34,35 and TXNIP, a key inflammasomal activator,36 has now been implicated in an animal model of AS.15,16 Skowasch et al.37 have demonstrated significant correlations between intravalvular and circulation hs-C-reactive protein in patients with advanced AS. Furthermore, the nexus between ASc and vascular events might reflect systemic inflammatory activation. More complete delineation of the role of inflammation in ASc is also relevant, given the recent finding of Miller et al.38

As a secondary analysis, the progression data were analysed on a categorical basis. The only significant finding was that such progression was less frequent in subjects receiving ACE-I/ARB therapy. Furthermore, the evaluation of data from subjects without ASc at baseline showed that this correlation was particularly prominent in this subgroup. The idea that the progression of aortic valve disease might be angiotensin-dependent is consistent with a number of pieces of experimental and clinical evidence. Experimentally, olmesartan retarded macrophage infiltration in valves of cholesterol-fed rabbits,39 whereas ramipril retarded AS development in rabbits treated with high doses of vitamin D.16 There is also substantial evidence of potentially increased angiotensin II formation in the presence of both ASc and AS.4042 As regards clinical data, no prospective trials have been undertaken to date. However, in human retrospective studies, Rosenhek et al.,43 utilizing echocardographic parameters, found a non-significant trend towards the retardation of AS progression with ACE-I therapy, whereas O'Brien et al.,44 in a study using CT-based calcification assessment, found lower rates of AS calcification among ACE-I treated subjects, after correction for comorbidities. In a more recent large study of ageing subjects with various degrees of AS conducted in Scotland, Nadir et al.45 demonstrated that the use of ACE-I/ARBs was associated with improved survival and cardiovascular outcomes.

The emergence of ACE-I/ARB therapy as a correlate of AVBS progression in categorical analyses may reflect a disproportionate impact in very early disease, as suggested by the subsidiary analysis (Table 6, ‘Subjects without ASc at baseline’), but may also reflect the interdependence of NO signalling and angiotensin II-related O2 production.46 For example, we have previously shown that ACE-I therapy potentiates platelet NO responsiveness.47 Furthermore, we have recently demonstrated that ramipril sensitizes platelets to NO, with a predominant effect in NO-resistant individuals.48 These observations suggest that the two ‘separate’ correlates may be different aspects of the same phenomenon. In a post hoc comparison between subjects who had never received ACE-I/ARB vs. those treated continuously with such agents during the study period, ACE-I/ARB therapy was associated with relative preservation of platelet NO responsiveness (ANCOVA: F = 4.5; P = 0.034).

It is important to emphasize that there was no evident association between hyperlipidaemia and ASc progression. These studies are consistent with our baseline data,7 and provide complementary information to that implicit in the results of interventional trials performed in patients with more advanced disease.4951

The unique methodology utilized in this study—determination of AVBS—has the advantage of considerable reproducibility12,1719 and therefore preferable to subjective scoring for ASc. AVBS values correlate well with both calcified volume on histology of aortic valve and calcium scoring on micro-CT in a rat model.17 Although this has not been evaluated in human studies in the valves, ultrasonic backscatter scores obtained from atherosclerotic vascular lesions correlate with histological features (volume of calcium) in a human study.52 Yet, the utility of AVBS for later progression is more doubtful. The other caveat related to this study is that few patients were diabetic and none had severe renal function impairment, conditions which may well predispose to rapid progression of ASc.53,54 Additionally, as most of hypertensive patients were treated with ACE-I/ARBs and vice versa, it is impossible to dissect the relative contribution of hypertension to aortic valve disease progression in this population cohort.

In conclusion, ASc progresses in the majority of an ageing cohort over a 4-year period. The current study has identified the elevation of plasma calcium concentrations and the impairment of tissue NO responsiveness as the main predictors of rapid ASc progression, and the utilization of ACE-I/ARB therapy as a predictor of lack of any progression, especially in very early stages of the disease. Furthermore, increased arterial stiffness and inflammatory activation were borderline correlates of progression, but hyperlipidaemia was not associated with progression in any way. These results provide a basis for the initiation of intervention studies directed at retardation of AS progression from its earliest stages.


This work is supported in part by the grants from the National Health and Medical Research Council of Australia and the Heart Foundation of Australia. A.L.S. and W.P.A.C. are recipients of the Cardiovascular Lipid Grants (Australia).

Conflict of interest: none declared.


We would like to thank the staff of the North West Adelaide Health Study Team for their help with patient recruitment. Our thanks to Ms G. Velissaris, Mr Matthew Chapman, and Mr Ronald Wuttke from the echocardiography department at the Queen Elizabeth Hospital for technical help. We would like to thank Associate Professor R. Wolfe for his invaluable advice on statistics.


  • These authors contributed equally to this work.


View Abstract