The prevalence of calcific aortic valve disease increases with age. Mild changes, termed aortic sclerosis (Stage A disease) are present in about 25% of all adults over 65 years of age and are a marker for adverse cardiovascular events with about a 50% increased risk of mortality over 5 years even in the absence of valve obstruction. (8) These findings were confirmed in a study of 3944 subjects in the Heinz Nixdorf study. Aortic valve calcification scores in the third tertile were associated with a higher incidence of coronary (HR 2.11, 95% CI 1.28 to 3.81) and cardiovascular events (HR 1.67, 95% CI 1.08 to 2.58), even after adjustment for Framingham risk factors. (9) Although valve calcification did not provide additive predictive value over coronary calcium scores in this study, in clinical practice echocardiographical imaging avoids ionising radiation and is more widely available than CT imaging. Thus, it remains unclear if aortic sclerosis on echocardiography should be additive to risk models for cardiovascular risk.

In the Tromsř population-based study of 3273 participants, more severe calcific disease was present in 164 subjects with a marked increase in prevalence of significant valve obstruction with age (Figure 2). In adults over 50 years of age, the annual incidence rate of AS was 4.9% with an average annual increase in aortic transvalvular gradient of 3.2 mm Hg. (10) In these patients, who were treated with valve replacement as needed for severe symptomatic disease, outcomes were similar for patients with AS and the general population.

Although the phenotype of calcific aortic valve disease typically presents late in life, there is increasing evidence that there is an underlying genetic predisposition to this condition. In a genome-wide association study of 6942 participants, a single nucleotide polymorphism (SNP) in the locus for lipoprotein(a) (Lp(a)) was associated with the presence of aortic valve calcification as assessed by CT scanning. (11) In addition, genetically determined serum Lp(a) was associated with both incident AS and with AVR. (11) In a separate study using a Mendelian randomisation study design, plasma low-density lipoprotein cholesterol was associated with an increased risk of incident AS (HR per mmol/L, 1.51; 95% CI 1.07 to 2.14; p=0.02). (12) An SNP-based risk score for predisposition to elevated plasma lipids was also associated with aortic valve calcium and AS. (12) Taken together, these studies are strong evidence that an underlying genetic predisposition to elevated serum lipid levels is associated with the development of calcific valve disease. In the future, this might allow therapy to be targeted towards patients at highest risk of developing valve disease.

Echocardiographical and haemodynamical measures of AS severity have generated considerable confusion, in large part related to the naive notion that velocity, gradient and valve area should always match and fall into clear categories of mild, moderate or severe. Ideally, a simple single measure would allow accurate diagnosis of AS severe enough to require valve replacement. Unfortunately, it is not that easy. In clinical practice, evaluation should include measurement of aortic velocity, mean transaortic gradient and calculation of valve area by the continuity equation. Although adjusting valve area for body size makes sense conceptually, indexing aortic valve area (AVA) for body size significantly increased the prevalence of apparently severe AS by including larger patients with only mild-to-moderate disease. (13) In addition, indexing AVA did not improve the predictive accuracy for clinical events in patients with mild-to-moderate AS. (14) In an editorial, Professors Pibarot and Dumesnil suggest a practical approach to resolving apparent discrepancies between valve area and mean gradient as shown in Figure 3. (15)

Another measure of AS severity that accounts for body size is the ratio of the velocity proximal to and in the stenotic valve the velocity ratio with normal being close to 1.0 and with smaller numbers indicating more severe valve disease. For example, a velocity ratio of 0.25 indicates a valve area 25% of normal for that individual. The velocity ratio is most useful to help distinguish low-flow severe AS from moderate AS in patients with a small valve area but only a moderately increased velocity or gradient. In 435 patients in the Simvastatin and Ezetimibe in Aortic Stenosis study with an AVA<1.0 cm2, mean gradient ≤40 mm Hg and LVEF≥55%, aortic valve events occurred more often in patients with a velocity ratio <0.25 compared with those with a higher velocity ratio (57% vs 41%; p<0.001) but did not provide additive value to mean gradient for prediction of clinical events (Figure 4). (16)

Increased recognition of symptomatic AS in patients with only a moderate increase in transvalvular velocity or mean gradient has led to further studies on low-output low-gradient severe AS and the realisation that this condition can occur even with a normal LVEF. The new valve guidelines4 provide clarity to this discussion with new definitions of severe AS:

When symptom status is unclear in adults with AS, previous studies have suggested that measurement of serum B-natriuretic peptide (BNP) levels provides additional prognostic value. However, in a single centre study of 361 patients older than 70 years with at least mild AS, Nt-proBNP levels correlated only modestly with outcome on univariate analysis, but not on multivariate analysis when adjusted for age, sex and AS severity suggesting this parameter be used with caution in elderly patients with AS. (17)

Evaluation of risk in patients being considered for surgical or transcatheter valve replacement has largely been based on surgical risk scores, which may be not fully applicable to transcatheter procedures. The 2014 Valve Guidelines recommend that, in addition to surgical risk scores, factors such as frailty, other organ system involvement and procedural specific factors also be considered. A simple risk score for prediction of early mortality after TAVI has been proposed by the FRANCE-2 Investigators based on predictors of 30-day mortality in 3833 consecutive patients undergoing balloon-expandable (67%) or self-expanding (33%) TAVI (Figure 5). (21, 22)

The baseline severity of mitral regurgitation is an independent predictor of mortality after TAVI. (23) However, mitral regurgitation does improve after relief of AS in about half of the patients, and a decrease in MR severity is associated with better outcomes. Another potential variable for risk stratification of patients being considered for TAVI is the red cell distribution width (RDW) with a baseline RDW ≥15.5% and a greater increase in RDWover time both found to be independently predictive of mortality after transcatheter aortic valve replacement (TAVR). (24) It is likely that RDW serves as a marker of various comorbidities, including uraemia, malnutrition, iron deficiency and inflammation. The risk of TAVI might be further decreased by avoiding general anaesthesia for this procedure as has been piloted by a group in Germany. (25, 26) A series of 461 patients underwent TAVI with local anaesthesia only, with valve placement guided by fluoroscopy, rather than transoesophageal echocardiography (TOE), with a total combined safety end point, as defined by the Valve Academic Research Consortium consensus statement, of only 12.6%. Rates for specific complications were: death (5%), cerebral complications (2.1%), vascular complications (7.1%), life-threatening bleeding (4.8%), acute kidney injury (1.1%) and pacemaker implantation (12.8%).

Paravalvular aortic regurgitation after TAVR continues to be a clinical concern, given the association with poor clinical outcomes. (23) In 1432 consecutive patients in the German TAVI registry, in the 201 (15.2%) with more than mild paravalvular AR, 61 (31%) died within 1 year.20 Independent predictors of mortality in those with paravalvular AR were more than mild baseline mitral regurgitation, a higher systolic pulmonary artery pressure and male sex.

Bleeding continues to complicate TAVI procedures, with lifethreatening bleeding occurring in 13%, major bleeding in 9% and minor bleeding in 5% of the 250 patients undergoing TAVI in a recent series. (27) The only independent predictor of lifethreatening bleeding was a transapical access approach.

Transcatheter valves often have a lower transvalvular pressure gradient and large orifice area than a comparable sized surgical bioprosthetic valve. Whether these favourable haemodynamics translate into improved reverse LV remodelling after TAVI has not been established. In a study comparing 25 TAVI and 25 surgical AVR patients matched for gender and AS severity, the 6-month postprocedure decrease in LV volumes and mass, as measured by cardiovascular magnetic resonance, were similar in both groups with baseline myocardial scar and LV volumes being the strongest predictors of reverse remodelling. (28)

As with any newer technology, there is concern about the costs of TAVI for the healthcare system. Using a Markov model with a 10-year horizon, TAVI was cost-effective compared with surgical AVR despite higher procedural costs due to a reduced length and cost of hospital stay in this high-risk cohort. (29) However, these estimates could vary depending on long-term outcomes so that continued attention to this issue is needed.

In patients undergoing surgical AVR, current guidelines recommend a bioprosthetic valve in patients over 70 years because this valve type is durable in older adults and the risks of long-term anticoagulation with a vitamin-K antagonist are avoided. (1, 2) There was hope that the newer oral anticoagulants might make mechanical valves a more attractive option. Disappointingly, a randomised trial of dabigatran versus warfarin after mechanical mitral valve replacement had to be terminated prematurely due to an excess rate of thromboembolic and bleeding events in patients randomised to dabigatran. (31) Although patients with mechanical aortic valves were not included in this study, it now will require careful consideration whether other clinical trials are appropriate in this population. From a clinical point of view, warfarin currently remains the only accepted antithrombotic therapy for patients with a mechanical heart valve.

Early mortality after AVR in patients receiving a mechanical valve was higher than in those receiving a bioprosthetic valve (unadjusted surgical mortality rates of 1.04% vs 0.57%) in a retrospective analysis of over 66 000 adults over 65 years of age. (32) This difference was confirmed using a mixed effects model for 30-day mortality (adjusted OR, 1.18 (95% CI 1.09 to 1.28; p<0.001); relative risk, 1.16; Number needed to treat (NNT), 121). (32) However, longer-term outcomes were not different either in this study or in a younger population in another retrospective study. In a retrospective cohort analysis of 4259 patients aged 50–69 years who received either a bioprosthetic or mechanical AVR, there was no significant difference in 15-year survival or stroke. (33) As expected, patients with a mechanical valve had a lower likelihood of reoperation but a greater likelihood of major bleeding. (33)

Some patients, particularly elderly women, have a small aortic annulus with suboptimal haemodynamics after valve replacement, a condition called patient–prosthesis mismatch (PPM). Previous studies have suggested that PPM is associated with increased mortality, less regression of LV hypertrophy and a higher incidence of heart failure postoperatively, especially in patients with a low EF at baseline. However, the impact of PPM may depend on age at time of valve replacement. (34) In a series of 707 adults undergoing valve replacement for severe AS, PPM was present in 42%, defined as an indexed effective orifice area ≤0.85 cm²/m². PPM was present in 26% of patients <70 years of age and was associated with decreased survival and increased heart failure in those with LV systolic dysfunction over a median follow-up of 7.3 years. Conversely, although PPM was more common in those over 70 years of age (68%), PPM was only associated with reduced LV mass regression but not with mortality or heart failure in this older age group (Figure 6).

The promising results with the use of transcatheter valve implantation within a failing bioprosthetic valve-in-valve procedure are changing the clinical management of this challenging clinical situation. (35, 36) The first step in clinical evaluation is to distinguish bioprosthetic valve stenosis from PPM, as haemodynamics may be similar, yet stenosis will be improved by valve-in-valve TAVI, whereas PPM will be worsened with further reduction in the effective orifice area by the transcatheter valve. Looking at the time course of changes in valve haemodynamics is helpful because PPM will be evident on the baseline postoperative echocardiography. In contrast, prosthetic valve stenosis will show an increased transvalvular velocity and gradient over time. Direct visualisation of the valve leaflets on TOE or CT imaging also may clarify the diagnosis. In addition, an understanding of the dimensions and design of surgical bioprosthetic valves is needed for proper placement of a valve-in-valve TAVI (Figure 7). The Valve-in-Valve International Data Registry reported a 1-year survival of 83.2% in 459 patients undergoing this procedure. (36) The mean patient age was 77.6 years, 56% were men and all were high risk for repeat surgical AVR. Bioprosthetic valve dysfunction was predominantly stenosis in about 39%, regurgitation in 30% and combined valve dysfunction in the remainder. At 1-month follow-up, 7.6% had died and 1.7% suffered a major stroke, but 92.6% of the surviving patients had a good functional status.