In a population-based study which followed up 953 subjects for 10 years, a high prevalence of calcific aortic valve disease (28%) associated with long-term exposure to raised cholesterol levels and active smoking was described.2 Intraleaflet haemorrhage (detected by immunohistochemistry at the moment of aortic valve replacement surgery) was frequently present in the valve leaflets of degenerative aortic stenosis (AS) and was associated with rapid progression of AS.3 In a small study of 164 patients with rheumatic AS (of whom 30 were treated with a statin), progression of AS was slower in patients receiving statins than in untreated patients (annual change of peak aortic velocity: 0.05±0.07 m/s/year vs 0.12±0.11 m/s/year, p=0.001).4 On the other hand in the ASTRONOMER trial, a randomised double-blind study, that allocated 269 patients to rosuvastatin 40mg daily or to placebo, statin treatment did not reduce progression of the disease in patients with AS.5

Based on the aortic jet velocity and the B-type natriuretic peptide (BNP) level, a risk score predicting outcome in patients with moderate-to-severe asymptomatic AS was derived and validated in an independent cohort: score=(peak velocity (m/s)×2)+(ln of BNP×1.5)+1.5 (if female sex). Event-free survival after 20 months was particularly poor (7%) for patients in the fourth quartile.6 In a separate study of patients with severe asymptomatic AS, event-free survival rates at 3 years were 49%, 33% and 11% for patients with peak aortic jet velocities between 4.0 and 5.0 m/s, 5.0 and 5.5 m/s or >5.5 m/s, respectively. In addition to the important implications for risk stratification, these data introduce us to the entity of “very severe aortic stenosis” based on a peak aortic jet velocity ≥5.0 m/s.7 In another study, receiver-operator curve analysis identified a peak aortic jet velocity ≥4.4 m/s, a left ventricular (LV) longitudinal myocardial deformation ≤5.9%, a valvular-arterial impedance ≥4.9mm Hg/ml/m2 and an indexed left atrial area ≥12.2cm2/m2 as factors associated with adverse outcomes in 163 patients with moderate to severe AS.8 Early elective surgery was performed on 102 patients with severe AS (valve area ≤0.75cm2, AV-velocity ≥4.5 m/s), and conventional treatment was used for 95 patients. Compared with conventional treatment, early surgery in patients with very severe AS was associated with improved long-term survival by decreasing cardiac mortality.9 However, this was not a randomised study and selection bias might have affected the results.

In asymptomatic patients with AS (n=135) and a normal exercise response, an exercise-induced increase in mean transvalvular gradient >20mm Hg was described as an independent risk predictor. These results thus suggest that exercise stress echocardiography may provide prognostic information additional to that obtained by standard exercise testing and resting echocardiography.10 Symptoms on treadmill exercise testing in 38 apparently asymptomatic patients with at least moderate AS were associated with a lower peak myocardial VO2, a lower peak stroke index during exercise and BNP levels.11 Increased valvuloarterial impedance (Z(va)) (which is calculated by dividing the estimated LV systolic pressure (systolic arterial pressure + mean transvalvular gradient) by the stroke volume indexed for the body surface area) is a marker of excessive LV haemodynamic load, and a value >3.5 successfully identifies patients with AS with a poor outcome.12 However, the clinical value of this measure remains to be fully determined.

The prognostic significance of mid-wall fibrosis and infarct patterns detected by late gadolinium enhancement was evaluated in 143 patients with aortic stenosis. Mid-wall fibrosis (HR=5.35; p=0.03) and ejection fraction (HR=0.96; p=0.01) were independent predictors of all-cause mortality and may provide a useful method of risk stratification.13 There is evidence of subclinical myocardial dysfunction early in the disease process despite normal left ventricular ejection fraction (LVEF). The myocardial dysfunction appears to start in the subendocardium and to progress to transmural dysfunction with increasing AS severity. Symptomatic patients with AS have more impaired multidirectional myocardial functions than asymptomatic patients.14 In patients with severe AS, impaired multidirectional LV strain and strain rate are present even with preserved LVEF, but a significant improvement occurs after aortic valve replacement (AVR).15 Lower average longitudinal strain is related to higher LV mass, concentric geometry and more severe AS.16 Inappropriately high LV mass was found in 58% of asymptomatic patients with severe AS and was related to cardiovascular events. Event-free survival for patients with appropriate and inappropriate LV mass, respectively, was 78% vs 56% at 1 year, 68% vs 29% at 3 years and 56% vs 10% at 5 years (all p<0.01).17 However, in patients with calcific AS and a normal LVEF the severity of stenosis was the most important correlate of symptomatic deterioration. Tissue Doppler measures of LV systolic and diastolic function and LV mass provide limited predictive information after accounting for the severity of stenosis.18

Severe aortic valve stenosis is a medical condition with limited short-term survival for patients over the age of 75 years, particularly those at high surgical risk. Patients with the highest surgical risk have the worst prognosis if AS is not treated.19 It has been confirmed, that patients screened but without the inclusion/exclusion criteria necessary to participate in a transcatheter aortic valve implantation trial do poorly and have extremely high mortality rates, especially in non-surgical groups: 274 such patients were treated medically or with balloon aortic valvuloplasty and had a mortality of 37.2% as compared with a mortality of 21.5% for 88 patients who underwent AVR (these latter patients were less symptomatic and had a lower EUROscore) during a median follow-up of about 1 year.20 In an observational study of 25 patients with severe AS presenting in cardiogenic shock, the use of an intra-aortic balloon pump improved the cardiac index from 1.77 to 2.18 and 2.36 l/min/m2 at 6 and 24h, respectively (p<0.001) and should thus be considered in this critically ill population while being evaluated for further interventions.21

Five-year survival in patients with low-flow/low-gradient aortic stenosis without contractile reserve was higher in patients undergoing AVR than in medically managed patients (54±7% vs 13±7%, p=0.001) despite a high operative mortality of 22%. Surgery should thus not be withheld in this subset of patients solely on the basis of lack of contractile reserve on dobutamine stress echocardiography.22

Measuring the degree of aortic valve calcification by multislice CT in patients with mild-to-moderate AS and an EF≥40%, showed that a threshold of 1651 arbitrary units provided 82% sensitivity, 80% specificity, 88% negative-predictive value and 70% positive-predictive value to diagnose severe AS. Performance was best in a subset of patients with low EF when the threshold correctly differentiated between patients with severe AS (the diagnosis was confirmed by mean gradient, natural history or dobutamine stress echocardiography) and those with non-severe AS in 46 of 49 cases. This method may be particularly useful for the evaluation of AS severity in difficult cases, such as patients with reduced EF and low or absent contractile reserve.23

On echocardiography approximately one-third of patients with severe aortic valve stenosis, based on aortic valve area <1.0cm2, have a non-severe mean pressure gradient (≤40mm Hg) despite apparently normal left ventricular function. Three hundred and thirty-three consecutive patients underwent cardiac catheterisation within 30 days after their index echocardiography. On invasive testing, 85 patients (26%) demonstrated inconsistent grading, with a significantly lower stroke volume and stroke volume index. However, 48% of inconsistently graded patients had a normal stroke volume index >35ml/m2. In the framework of current guidelines inconsistent grading of aortic valve stenosis is common, extends to cardiac catheterisation and is only partially explained by low stroke volume despite apparently normal left ventricular systolic function.24 In this SEAS substudy, aortic valve-related events, major cardiovascular events and cardiovascular death in patients with low-gradient “severe” aortic stenosis (aortic valve area <1.0cm2 and mean gradient ≤40mm Hg) were comparable to those of patients with moderate stenosis (aortic valve area 1.0–1.5cm2; mean gradient 25–40mm Hg).25 These results fuel the debate on the management of such patients. In severe AS, a low gradient is associated with a higher degree of interstitial fibrosis in biopsy samples and more late-enhancement MRI segments, decreased longitudinal function assessed by echocardiography and poorer clinical outcome despite preserved EF.26

Higher serum phosphate levels within the normal range were associated with aortic valve sclerosis and mitral and aortic annular calcification in a community-based cohort of older adults. In contrast, serum calcium, parathyroid hormone and 25-hydroxyvitamin D concentrations were not associated with aortic or mitral calcification. Phosphate may be a new risk factor for calcific aortic valve disease and warrants further study.27 The upregulation of the leukotriene pathway in human aortic valve stenosis and its correlation with clinical stenosis severity, taken together with the potentially detrimental leucotriene-induced effects on valvular myofibroblasts, suggests one possible role of inflammation in the development of AS.28 Mechanical properties of porcine aortic valve leaflets were evaluated: serotonin induced a decrease in the areal stiffness of the cusp, which was reversed by N-nitro-L-arginine-methyl ester or endothelial denudation. Endothelin-1 caused an increase in stiffness, but not in the presence of cytochalasin B. Changes in cusp stiffness were accompanied by aortic cusp relaxations to 5-hydroxytriptamine, which were reversed by endothelial denudation and by N-nitro-L-arginine-methyl ester. These data highlight the role of the endothelium in regulating the mechanical properties of aortic valve cusps and underline the importance of valve cellular integrity for optimal valve function.29 A reduced regenerative capacity of valvular endothelial cells due to senescence and decreased levels of endothelial progenitor cells might be, at least in part, a pathological link for the destruction of valvular endothelial cells, resulting in progression of degenerative AS.30 Direct in vivo evidence was provided that cathepsin S-induced elastolysis accelerates arterial and aortic valve calcification in chronic renal disease, providing new insight into the pathophysiology of cardiovascular calcification.31 In the low-density lipoprotein-receptor-deficient mouse, regular exercise training prevents aortic valve sclerosis by several mechanisms, including the preservation of endothelial integrity, a reduction in inflammation and oxidative stress, and inhibition of the osteogenic pathway.32 Recombinant apolipoprotein A-I Milano treatment reverses AS in an experimental rabbit model. The beneficial effects seem to be mediated by enhanced cholesterol removal and by reduced inflammation and calcification.33 Additional data indicate that reducing plasma lipid levels by genetic inactivation of the MTTP gene in hypercholesterolaemic mice with early aortic valve disease normalises oxidative stress, reduces proosteogenic signalling and halts the progression of aortic valve stenosis.34 Patients with AS and diabetes have worse diastolic LV dysfunction, predisposing to heart failure. It appears to result from greater myocardial fibrosis (documented with perioperative LV biopsies), more intra-myocardial vascular advanced glycation end-product deposition and higher cardiomyocyte Fpassive, which is related to hypophosphorylation of the N2B titin isoform.35

The controversy surrounding the Ross procedure is highlighted by four studies. In a randomised study 216 patients received either an autograft or a homograft aortic root replacement. At 10 years, four patients in the autograft group and 15 in the homograft group died. Actuarial survival at 10 years was 97% in the autograft group versus 83% in the homograft group.85 MRI of 45 patients at a median interval of 8 years postoperatively demonstrated minor autograft and homograft dysfunction in the majority of cases, associated with good ventricular function and exercise capacity.86 Another study compared the outcome of the Ross procedure (918 patients) with that of 406 mechanical valve recipients under optimal self-management anticoagulation treatment; there was no late survival difference in the first postoperative decade between the two groups. Survival in these selected young adult patients closely resembles that of the general population, possibly as a result of optimised anticoagulation self-management, timing of surgery and patient selection.87 A less optimistic picture was depicted by a study emphasising the broad spectrum of complex reoperations that may be required relatively often after the Ross procedure. The four most common indications for reoperation (n¼56) were isolated autograft (neoaortic) regurgitation in 20%, isolated pulmonary conduit regurgitation/stenosis in 16%, combined autograft regurgitation/dilatation in 14%, and combined autograft regurgitation and pulmonary conduit regurgitation/stenosis in 11%. Patients and family members considering the procedure should be informed of the potential for associated morbidity should reoperation be necessary.88

A 6 min walk test was found to be safe and feasible to carry out in patients with severe AS before AVR, and provides potentially important functional and prognostic information for clinical assessment and the Euroscore risk score. At 12 months, the rate of death, MI or stroke was 13% in patients walking <300 m as compared with 4% in those who walked ≥300 m (p=0.017).89 Physical quality of life 1 year after valve surgery was predicted by baseline physical quality of life and walk performance. Postoperative mental quality of life was predicted by depression, baseline mental quality of life and age, with age having a positive effect, suggesting that treating depression and modifying negative illness beliefs preoperatively, may improve outcome.90 Women referred for AVR were found to be older and more symptomatic. Although operative and long-term mortality were not increased, women remained in a more symptomatic stage.91 Patients undergoing renal transplantation requiring valve replacement have high mortality rates (approximately 20%/ year). Two-year survival estimates were comparable for patients receiving a tissue valve (61.5%) or a non-tissue-valve (59.5%, p=0.30).92

Patients aged 55–70 years undergoing AVR either with mechanical or bioprosthetic valves had similar 13-year rates of survival, thromboembolism, bleeding, endocarditis and major adverse prosthesis-related events. However, patients with bioprosthetic valves had a significantly higher risk of valve failure and reoperation.93 Using microsimulation of survival and valve-related outcomes from 5470 AVR procedures, it was found that bioprostheses may be implanted selectively in patients as young as 56 without significant adverse effects on life expectancy, although event-free survival remains significantly lower with bioprostheses for patients up to age of 63.94 Increasing numbers of the very elderly are undergoing AVR procedures. Late survival of 2890 consecutive elderly patients (≥70 years) who underwent AVR was influenced by age and preoperative comorbidities; the 33% in the lowest risk tertile had an overall survival similar to that of the age- and sex-matched general population. Structural deterioration of aortic bioprostheses was rare and there was no conclusive evidence that valve type affected survival in these patients.95 Also in octogenarians, survival after AVR is favourable even with concomitant bypass surgery and more than half of the patients survive for more than 6 years after their surgery. Median survival for patients undergoing isolated AVR was 6.8 years for those aged 80–84 years (n=419) and 6.2 years for those aged ≥85 years (n=156), similar to the life expectancy of the general population.96

Patient-prosthesis-mismatch (PPM) was identified in 40% of 645 patients after AVR in a study in which indexed effective orifice area was obtained by postoperative echocardiography and modelled as a continuous variable. After a median follow-up of 2.35 years, 92.1% of the patients were alive. Cardiac death among patients with a smaller indexed effective orifice area was significantly increased (HR=0.32, p=0.022).97 Among 2576 patients who survived AVR and after adjustment for other risk factors, severe PPM was associated with increased late overall mortality (HR=1.38; p=0.03) and cardiovascular mortality (HR=1.63; p=0.0006). Moderate PPM was a predictor of mortality in patients with LV ejection fraction <50% (HR=1.21; p=0.01), but not in patients with preserved LV function.98 The presence of PPM after AVR attenuates postoperative mitral regurgitation changes, mainly in patients with organic mitral regurgitation.99 In 564 patients receiving an aortic valve bioprosthesis, structural valve deterioration was diagnosed in 40 patients (7%). Stenosis-type structural valve deterioration (n=24) was found to be an early, PPM-related, and thus preventable, phenomenon. Regurgitation-type structural valve deterioration (n=16) is a time-dependent, non-specific wear of bioprosthetic valves, which is not related to PPM.100 In a multicentre series of 1006 mechanical and bioprosthetic mitral valves, PPM was not associated with worse early outcomes or worse mid-term survival.101

Elective mitral valve (MV) repair can be performed with a low operative mortality and good long-term outcomes in selected octogenarians with degenerative mitral disease, and is associated with better long-term survival than mitral replacement. Overall 90-day mortality of consecutive octogenarians who underwent MV repair was significantly lower (18.9%) than for MV replacement (31.6%). Adjusted 1-, 3- and 5-year survival for patients undergoing MV repair was 71±3, 61±4 and 59±4%, respectively, compared with 56±5, 50±6 and 45±6% for patients undergoing MV replacement (p=0.046). The survival benefit associated with surgery for non-degenerative disease is more questionable.102

Of 402 patients with atrial fibrillation-associated MV disease who underwent MV replacement with a mechanical prosthesis, 159 underwent a concomitant Maze procedure. At a median follow-up of 63.1 months, patients who had undergone the Maze procedure were at significantly lower risk of thromboembolic events (HR=0.26; p=0.041) and were at comparable risk of death and cardiac death as patients who underwent MV replacement alone.103

In 370 patients with ischaemic MR after adjusting for other risk factors and propensity score, the type of procedure (MV repair versus MV replacement) was not an independent predictor of either operative or overall mortality.104 One hundred and thirty-five patients with ischaemic heart disease and moderate ischaemic MR underwent isolated coronary artery bypass graft surgery. At 1 year, 57 patients had no or mild ischaemic MR, whereas 64 patients had failed to improve. Large extent (≥5 segments) of viable myocardium (OR=1.45; p<0.001) and absence (<60ms) of dyssynchrony (OR=1.49; p<0.001) were independently associated with improvement in ischaemic MR.105

Combined mitral and tricuspid valve repair in rheumatic disease showed satisfactory early results in 153 consecutive patients (mean age 46 years) who underwent combined mitral and tricuspid valve repair for rheumatic disease. However, long-term results were poor because of high mortality and a high number of valve-related reoperations. Survival-rate was 74.4% at 10 years and 57.0% at 15 years. At 20 years, the rate of freedom from reoperation was 48.5%.106

Despite the use of intravenous unfractionated heparin, the rate of early thromboembolism in a series of 300 consecutive mechanical valve replacements remained significant. Early thromboembolism within 30 days of surgery occurred in 22 patients (14.8%) after a mitral or double mechanical valve replacement and in two patients (1.3%) after an aortic mechanical valve replacement (p=0.005). Inappropriate anticoagulation on day 3 was significantly associated with early thromboembolism, suggesting that early effective anticoagulation is required after mitral mechanical valve replacement.107

A randomised controlled trial comparing an accelerated infusion with the conventional infusion of streptokinase was performed in 120 patients with a first episode of left-sided prosthetic valve thrombosis, recruited over a 2.5-year period at a single centre in India. The large patient number underlines the massive burden of prosthetic valve thrombosis in developing countries. Fibrinolytic therapy with streptokinase is less efficacious than previously believed, with a complete clinical response in 70 of 120 patients. The accelerated streptokinase infusion is no better than standard infusion for left-sided prosthetic valve thrombosis.108

Objective parameters to assess interventional risk and thus to identify patients at high risk who would benefit from percutaneous procedures are needed. For this, reliable risk scores that predict surgical mortality would be helpful. While the EuroSCORE still successfully discriminates high-risk patients undergoing surgical aortic valve replacement, it has become increasingly uncalibrated with absolute risk, resulting in over-estimation of 30-day mortality.109 The limitations of risk scores are also commented upon in a recent ESC position paper on risk assessment before interventions in patients with valvular disease.110

Adequate sizing of the aortic annulus is essential in order to assess the suitability of a patient for a transcatheter aortic valve implantation (TAVI) procedure and the choice of the prosthesis size. By CT an ellipsoid shape of the aortic valve annulus with a larger coronal than sagittal diameter (25.1±2.4 vs 22.9±2.0mm, p<0.001) was measured.111 2D imaging techniques underestimate aortic annulus diameters and 3D imaging techniques are now recommended for this purpose. 3D transoesophageal echocardiography (TOE) provides measurements of aortic annulus diameters similar to those obtained by CT.112 While measurements using transthoracic echocardiography, TOE and CT are close but not identical, a strategy based on TOE measurements provides good clinical results.113 The presence of LV dysfunction, male gender and larger body surface area are independent determinants of a larger aortic annular diameter.114 By CT, incomplete and non-uniform expansion of the CoreValve frame can be identified: undersizing and incomplete apposition is commonly present.115 Non-circular deployment of the prosthesis is found in 14% of patients. Moderate postprocedural aortic regurgitation is seen in 11% patients and is associated with larger aortic valve annulus, more calcified native valves and less favourable prosthesis deployment.111

To provide consistency across studies that can facilitate the evaluation of this new catheter-based treatment, and improve the quality of clinical research, the Valve Academic Research Consortium proposed standardised consensus definitions for important clinical end points in TAVI investigations.116 Retrospective examination of adherence to patient selection criteria identified an “off-label” use of TAVI beyond pre-market label indications in 42 of 63 patients.117 This study highlights the challenges encountered in the rollout phase of a new technology. The ability to offer either transfemoral or transapical aortic valve implantation, using a standardised approach, with the transfemoral approach as the first option, may expand the scope of the treatment of AS in high-risk patients and provide satisfactory 1-year results.118 Introduction of a TAVI service does not appear to have a negative effect on conventional surgical activity. One study reported a 37% increase in surgical AVR in the 2 years after introduction of TAVI in a dedicated centre, compared with an 8% increase nationally (p<0.001).119

An early single-centre experience established the feasibility of TAVI, both by the transfemoral approach (n=168), with a success rate of 94.1% and 1-year survival of 74%,120 and by the transapical approach (n=100), with a success rate of 97% and 1-year survival of 73%.121 Data for an extended follow-up period of 3 years have been reported, and no cases of structural valvular deterioration, stent fracture, deformation, or valve migration occurred.122

Patients with severe AS, considered unsuitable candidates for surgery (n=358), were randomly assigned to standard treatment (including balloon aortic valvuloplasty) or transfemoral TAVI using the Edwards SAPIEN valve in the PARTNER B trial. TAVI, as compared with standard treatment, significantly reduced the rates of death from any cause (30.7% vs 50.7%), the composite end point of death from any cause or repeat hospitalisation (42.5% vs 71.6%) and cardiac symptoms, despite a higher incidence of major strokes (5.0% vs 1.1%) and major vascular events (16.2% vs 1.1%).123 This study also nicely depicts the contemporary natural history of severe symptomatic AS.

The PARTNER A trial randomised 699 high-risk patients with severe AS to undergo TAVI or surgical AVR. Transcatheter and surgical procedures for AVR were associated with similar rates of survival at 1 year (24.2% vs 26.8%, respectively), although there were important differences in periprocedural risks, with vascular complications more common in the TAVI group (11.0% vs 3.2%, p<0.001) and more frequent major bleeding and new-onset atrial fibrillation with surgery.124 One-year survival in the SOURCE registry (n=1038) was 76.1% (72.1% for transapical and 81.1% for transfemoral TAVI). Interestingly, causes of death were mainly non-cardiac in 49.2% (cardiac in 25.1%, and unknown in 25.7%) with pulmonary complications (23.9%), renal failure (12.5%), cancer (11.4%) and stroke (10.2%) as the most common non-cardiac causes of death.125 These data reflect the importance of associated comorbidities. Several other multicentre registries (including the PARTNER EU registry, the German TAVI registry, the French FRANCE registry, an Italian and a Canadian registry) have confirmed the feasibility of the procedure in high-risk or unoperable patients with AS with good procedural success, haemodynamic results and mid-term outcomes.126–131

Mean transprosthetic gradients were lower for TAVI (10±4mm Hg) than for stented (13±5mm Hg) and stentless (14±6mm Hg) bioprostheses (p<0.001). Severe PPM was significantly lower with TAVI (6%) than with a bioprosthesis (24%; p=0.007).132 TAVI can be successfully carried out in most patients (34/35) with a small aortic annulus diameter <20mm, with severe PPM occurring in two patients only, and gradients remaining low in the other patients.133 It may also provide an interesting alternative to AVR in patients with depressed LV systolic function, where it is found to be associated with better LVEF recovery than conventional AVR (change in LVEF 14±15% vs 7±11%; p=0.005), although these patients were older and had more significant comorbidities. At 1 year, 58% of TAVI patients had a normalisation of LVEF (>50%) as opposed to 20% in the AVR group.134 Pre-procedural functional performance status (assessed by the Karnofsky index) predicts the in-hospital procedural success rates and MI and stroke rates after TAVI.135 Thirty days after TAVI quality of life and 6min walk distance improved significantly while BNP levels declined.136 Acute kidney injury occuring in 11.7% of patients after TAVI, is associated with a greater than fourfold increase in the risk of postoperative mortality. Hypertension, chronic obstructive pulmonary disease and blood transfusion are predictive factors of acute kidney injury.137 TAVI was systematically associated with some degree of myocardial injury in the majority of patients. The greater degree of myocardial injury seen with the transapical approach and baseline renal dysfunction is associated with less improvement in LVEF and a greater cardiac mortality at follow-up.138 Significant AR was reported to occur in 17.2% of patients and is associated with significantly higher in-hospital death rates (15.1% vs 6.7%), rates of low cardiac output and respiratory failure.139 For the CoreValve a greater likelihood of significant AR was found with a greater angle of LV outflow tract to ascending aorta (OR=1.24, p=0.001).140 New cerebral ischaemic lesions can be detected by diffusion-weighted MRI in between 68% and 84% of patients after TAVI.141–143 These lesions were usually multiple (1 to 19 per patient) and dispersed in both hemispheres in a pattern suggesting cerebral embolisation. These foci were not associated with apparent neurological events or measurable deterioration of neurocognitive function.

The rate of major stroke was in the range of 3.3% to 3.8%.124,141,143