Author + information
- Received January 21, 2013
- Revision received April 8, 2013
- Accepted May 8, 2013
- Published online September 10, 2013.
- Ruben L.J. Osnabrugge, MS∗,
- Darren Mylotte, MD†,‡,
- Stuart J. Head, MS∗,
- Nicolas M. Van Mieghem, MD§,
- Vuyisile T. Nkomo, MD, MPH‖,
- Corinne M. LeReun, MS¶,
- Ad J.J.C. Bogers, MD, PhD∗,
- Nicolo Piazza, MD, PhD†,# and
- A. Pieter Kappetein, MD, PhD∗∗ ()
- ∗Department of Cardiothoracic Surgery, Erasmus University Medical Center, Rotterdam, the Netherlands
- †Department of Interventional Cardiology at McGill University Health Center, Montreal, Quebec, Canada
- ‡Galway University Hospital, Galway, Ireland
- §Department of Interventional Cardiology, Erasmus University Medical Center, Rotterdam, the Netherlands
- ‖Division of Cardiovascular Diseases, Department of Internal Medicine, Mayo Clinic, Rochester, Minnesota
- ¶Self-employed biostatistician, Carrigaline, County Cork, Ireland
- #Department of Cardiovascular Surgery, German Heart Center, Munich, Germany
- ↵∗Reprint requests and correspondence:
Dr. A. Pieter Kappetein, Department of Cardiothoracic Surgery, Erasmus University Medical Center, P.O. Box 2040, 3000 CA Rotterdam, the Netherlands.
Objectives The purpose of this study was to evaluate the prevalence of aortic stenosis (AS) in the elderly and to estimate the current and future number of candidates for transcatheter aortic valve replacement (TAVR).
Background Severe AS is a major cause of morbidity and mortality in the elderly. A proportion of these patients is at high or prohibitive risk for surgical aortic valve replacement, and is now considered for TAVR.
Methods A systematic search was conducted in multiple databases, and prevalence rates of patients (>75 years) were pooled. A model was based on a second systematic literature search of studies on decision making in AS. Monte Carlo simulations were performed to estimate the number of TAVR candidates in 19 European countries and North America.
Results Data from 7 studies (n = 9,723 subjects) were used. The pooled prevalence of all AS in the elderly was 12.4% (95% confidence interval [CI]: 6.6% to 18.2%), and the prevalence of severe AS was 3.4% (95% CI: 1.1% to 5.7%). Among elderly patients with severe AS, 75.6% (95% CI: 65.8% to 85.4%) were symptomatic, and 40.5% (95% CI: 35.8% to 45.1%) of these patients were not treated surgically. Of those, 40.3% (95% CI: 33.8% to 46.7%) received TAVR. Of the high-risk patients, 5.2% were TAVR candidates. Projections showed that there are approximately 189,836 (95% CI: 80,281 to 347,372) TAVR candidates in the European countries and 102,558 (95% CI: 43,612 to 187,002) in North America. Annually, there are 17,712 (95% CI: 7,590 to 32,691) new TAVR candidates in the European countries and 9,189 (95% CI: 3,898 to 16,682) in North America.
Conclusions With a pooled prevalence of 3.4%, the burden of disease among the elderly due to severe AS is substantial. Under the current indications, approximately 290,000 elderly patients with severe AS are TAVR candidates. Nearly 27,000 patients become eligible for TAVR annually.
Aortic stenosis (AS) is the most common valvular heart disease in developed countries, and its impact on public health and health care resources is expected to increase due to aging Western populations (1,2). Each year, approximately 67,500 surgical aortic valve replacements (SAVR) are performed in the United States (3). Studies describing the prevalence of AS are scarce and report disparate results (3% to 23%) (4,5), and currently there is no systematic overview of population-based studies that have assessed the prevalence of AS.
The emergence of transcatheter aortic valve replacement (TAVR) has renewed interest in the epidemiology of AS. In particular, these data may be important to predict the number of TAVR candidates, service development, financial planning, and physician training. In addition, estimates of potential TAVR candidates at intermediate and low surgical risk are not available. Several factors must be considered when estimating the number of TAVR candidates: the percentage of patients with severe AS who are symptomatic; the proportion of patients with symptomatic severe AS who do not undergo SAVR and could thus be considered TAVR candidates; and the percentage of those patients referred for TAVR who actually receive a transcatheter valve.
Therefore, we sought to assess the prevalence of AS in the general elderly population (age ≥75 years) through a systematic review and meta-analysis of population-based studies. The second objective was to systematically estimate the number of elderly patients who are TAVR candidates in both the European countries and North America.
Studies were identified through a systematic search of MEDLINE and EMBASE in February 2012. Keywords included “valvular heart disease,” “heart valve disease,” “aortic stenosis,” “aortic valve stenosis,” “epidemiology,” “incidence,” “prevalence,” and “survey.” No time restrictions were applied. Reference lists of selected studies and (systematic) reviews were examined, and the related article feature in PubMed was used to maximize relevant study identification.
All titles and abstracts were screened independently by 2 investigators using the following criteria: 1) the publication was an original full-length manuscript in a peer-reviewed journal; 2) the publication reported numbers of AS cases and sample size or the prevalence of AS in the general elderly population (≥75 years of age); and 3) AS and AS severity was diagnosed with echocardiography (6,7). The definition of AS used in each study was extracted, as was other relevant information including study location, inclusion period, and patient characteristics. After excluding manuscripts on the basis of title and abstract, the remaining full-text manuscripts were carefully assessed and were evaluated according to the criteria. If overlap between studies existed, only the publication with the largest population was included. Disagreement on study inclusion was solved by consensus.
For each included study, the prevalence rate of AS and its 95% binomial confidence interval (CI) was calculated based on the numbers of subjects in the sample and the number of patients with AS. These rates were subsequently combined to produce a pooled prevalence rate of both AS and severe AS. Both fixed- and random-effects models were used, and results of the appropriate model are presented as Forest plots. The fixed-effects model was performed using the inverse variance method and the random-effects model with the DerSimonian and Laird method. Heterogeneity was assessed by the Cochran Q test and I2 statistics, derived from the inverse variance fixed-effects model (8). All analyses were performed with Stata SE version 12.0 (StataCorp, College Station, Texas).
Estimation of TAVR candidates
To estimate the number of elderly patients who could potentially be treated with TAVR under current indications, we performed a second literature search on clinical decision making in patients with severe AS. Specifically, we searched for studies that reported: 1) the percentage of patients with severe AS who experienced symptoms; 2) the percentage of patients with symptomatic severe AS who did not undergo SAVR and could thus be considered potential TAVR candidates; and/or 3) the percentage of those patients referred for TAVR who actually received a transcatheter valve. As TAVR is an approved therapy for patients at high operative risk, we also determined the proportion of elderly high-risk patients (The Society of Thoracic Surgery-Predicted Risk Of Mortality [STS-PROM] score ≥10%) undergoing SAVR (9), and the percentage of patients who would be considered TAVR-eligible. In anticipation of current and potential future trials in lower risk groups, estimates of the proportion of intermediate- and low-risk patients were also derived. For all studies, the point estimate and 95% binomial CI were calculated.
These data were combined to produce a pooled percentage estimate for each individual search. In each case, a fixed- or random-effects model was used and heterogeneity was assessed. To calculate national estimates of the number of patients with AS and TAVR candidates, we obtained population demographic data focusing on the elderly (≥75 years of age) for the following nations: Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Greece, Italy, the Republic of Ireland, Luxembourg, Norway, Poland, Portugal, Spain, Sweden, Switzerland, the Netherlands, the United Kingdom, Canada, and the United States (10–12). The annual number of new TAVR candidates was calculated using the number of people ages 75 years old in 2011 in the individual countries.
A flowchart was built in TreeAge Pro 2011 (TreeAge Software, Williamstown, MA). The probabilities in the flowchart were based on the pooled estimates from the systematic literature searches. Beta distributions were used and 10,000 Monte Carlo simulations were performed to estimate the number of elderly patients who are eligible to undergo TAVR, along with its 95% percentile CI.
To account for the heterogeneous nature of the studies, sensitivity analyses were performed. In particular, the proportion of patients receiving TAVR after referral for TAVR assessment was determined using European studies alone and then by combining European and U.S. studies. This analysis was performed to account for the different adoption of TAVR in the United States, where until recently TAVR was only used in the context of clinical trials. In a second sensitivity analysis, we varied the percentage of high-risk SAVR-eligible patients who undergo TAVR.
The systematic literature search yielded 1,523 studies. After the title and abstract were screened, 1,408 studies were excluded because they did not focus on the epidemiology of disease. After assessing full-text articles, another 109 studies were excluded because they were not performed in the general elderly population, AS was not assessed, or because it was not an original publication. After the inclusion of an additional study through cross-referencing, our final analysis consisted of 7 studies, with a total of 9,723 elderly patients (Fig. 1) (1,4,5,13–16). The characteristics of these studies are outlined in Table 1. The 7 studies reported the prevalence of AS in 9 study populations on 3 continents. The study periods ranged from 1989 to 2009. All studies had a cross-sectional character, and most were part of larger population-based cohort studies. In all 7 studies, echocardiography was used to diagnose AS, although definitions of AS and its severity were variable (Table 1).
The combined prevalence of AS in the elderly was reported in 6 studies and ranged from 2.6% to 22.8% (Fig. 2A) (4,5,13,15,16). The pooled prevalence was 12.4% (95% CI: 6.6% to 18.2%) using a random-effects model (I2 = 98.5%; Q = 337.70, p < 0.001). The prevalence of severe AS in the elderly was reported separately in 5 studies and ranged from 1.2% to 6.1% (Fig. 2B) (1,4,13,14,16). The pooled prevalence of severe AS was 3.4% (95% CI: 1.1% to 5.7%) using a random-effects model (I2 = 85.7%; Q = 27.99, p < 0.001).
These estimates of the prevalence of AS in patients ≥75 years old correspond to approximately 4.9 million elderly patients with AS in the European countries and 2.7 million in North America. If only symptomatic severe AS is considered, this translates to 1.0 million elderly patients in the European countries and 540,000 in North America. In 2011, 8.5% of the population in the 19 European countries was ≥75 years of age, and this number is expected to increase to 10.7% in 2025 and 16.6% in 2050 (11). In North America, similar increases in the population demographics of the elderly are expected (2025, 8.3%, and 2050, 11.8%) (10,12). These numbers correspond to approximately 1.3 million and 2.1 million patients with symptomatic severe AS in the 19 European countries in 2025 and 2050, respectively. In North America, there will be an estimated 0.8 million and 1.4 million patients with symptomatic severe AS in 2025 and 2050, respectively.
Estimates of TAVR candidates
The number of elderly patients who could potentially benefit from TAVR was estimated using the model outlined in Figure 3, with inputs from the systematic search and meta-analyses (Fig. 4). Seven studies reported the percentage of patients with severe AS who were symptomatic, resulting in a pooled estimate of severe symptomatic AS of 75.6% (95% CI: 65.8% to 85.4%) (Fig. 4A, Online Table 1). Of these patients with symptomatic severe AS, 40.5% (95% CI: 35.8% to 45.1%) did not undergo SAVR and thus could be considered candidates for TAVR (Fig. 4B, Online Table 2). Nine studies reported the percentage of patients referred for TAVR who actually received a transcatheter valve (Online Table 3). Three of these studies were performed in Europe, and 6 in the United States. The pooled percentage including both European and U.S. studies was 28.7% (95% CI: 22.8% to 34.6%) (Figs. 4C and 4D, respectively). The European pooled percentage was 40.3% (95% CI: 33.8% to 46.7%), whereas the U.S. pooled percentage was 24.4% (95% CI: 18.9% to 29.8%). In total, 12.3% of patients with symptomatic severe AS at prohibitive surgical risk are TAVR candidates.
To assess the proportion of elderly SAVR patients who was deemed to be at high surgical risk, we used a study that reported on all elderly SAVR patients in the United States between 1999 and 2007 (17). Among elderly patients undergoing isolated SAVR, 5.2% (95% CI: 4.9% to 5.4%) were at high risk (STS-PROM ≥10%), 15.8% (95% CI: 15.4% to 16.2%) at intermediate risk (STS-PROM 5% to 10%), and 79.1% (95% CI: 78.6% to 79.5%) at low risk (STS-PROM <5%). A recent study showed that in a group of operable patients with a EuroSCORE (European System for Cardiac Operative Risk Evaluation) ≥15, approximately 80% were treated with TAVR (18).
In 2011, there were 39,316,978 people ≥75 years of age in the European countries and 21,182,683 in North America (10–12). Combining these figures with the Monte Carlo simulations in the model (Fig. 3), we estimated that a total of 292,000 high- or prohibitive-risk elderly patients with symptomatic severe AS are candidates for TAVR. Specifically, there are 189,836 (95% CI: 80,281 to 347,372) TAVR candidates in the European countries and 102,558 (95% CI: 43,612 to 187,002) in North America. Annually there are 17,712 (95% CI: 7,590 to 32,691) new TAVR candidates in the European countries and 9,189 (95% CI: 3,898 to 16,682) in North America. The total and annual number of TAVR candidates in the individual countries is presented in Figures 5 and 6, respectively.
The intermediate surgical risk group comprises approximately 145,000 elderly patients with symptomatic severe AS. Specifically, there are 94,730 (95% CI: 40,574 to 171,896) patients at intermediate risk in the European countries and 50,733 (95% CI: 22,148 to 90,451) in North America. The low surgical risk group includes approximately 730,000 patients with symptomatic severe AS. Specifically, there are 477,314 (95% CI: 206,798 to 862,958) patients at low-risk in the European countries and 255,727 (95% CI: 108,549 to 460,026) in North America.
In the pre-specified sensitivity analysis that varied the proportion of patients receiving TAVR after referral for TAVR assessment according to study location (28.7%, 95% CI: 22.8% to 34.6% in Europe and the United States combined), we estimated that approximately 220,000 patients are TAVR candidates. Of these, 142,658 (95% CI: 61,065 to 263,795) candidates lived in the European countries and 76,962 (95% CI: 32,805 to 140,673) in North America.
In the sensitivity analysis varying the percentage of high-risk operable patients who would undergo TAVR, the total number of TAVR candidates was 277,570 (95% CI: 119,406 to 512,707) assuming that 50% would undergo TAVR whereas there were 302,865 (95% CI: 129,433 to 550,562) candidates if all the high-risk patients would undergo TAVR. Finally, we estimated that the total number of patients with symptomatic severe AS in the intermediate-risk category was 145,936 (95% CI: 62,802 to 263,340), and 733,861 (95% CI: 310,623 to 1,302,586) in the low-risk category.
The current study found that the prevalence of AS in the elderly (≥75 years of age) is 12.4%, and severe AS is present in 3.4%. Among elderly patients with severe AS, 75.6% are symptomatic, and 40.5% of these patients are not treated surgically. From those, 40.3% are potentially treated with TAVR. In total, 12.3% of the prohibitive risk group are TAVR candidates. Among patients undergoing SAVR for severe symptomatic AS, 5.2% are high risk and 80% of those are potential TAVR candidates. Based on these data, we estimated that there are currently approximately 190,000 and 100,000 TAVR candidates in the European countries and North America, respectively. Each year, approximately 18,000 new TAVR candidates emerge in the European countries and 9,000 in North America.
The prevalence of AS
Our estimates of the prevalence demonstrate that the overall burden of disease due to AS in the general elderly population is substantial. Population demographics clearly show that Western populations are aging, thereby further increasing the impact of AS. No effective medical therapy is available for patients with AS, and if not treated by intervention, the estimated 5-year survival of severe AS is only 15% to 50% (7). These data suggest that the treatment of AS in the elderly will have an increasing impact on public health and health care resource consumption in the future.
Based on echocardiographic diagnosis, we found that severe AS occurs in 12.4% of the general elderly (≥75 years of age) population. Previous autopsy series and a study based on aortic valve diagnoses in Medicare claims have reported AS prevalence estimates of 9.2% and 16%, respectively (19,20). Our pooled prevalence of AS (12.4%) is lower than the estimates from Medicare claims, but covered a lower age group and did not include diagnoses of aortic regurgitation. The methodological differences between studies are likely to account for the variability in AS estimates.
We explored heterogeneity by assessing the individual study characteristics, but the limited number of studies prevented separate analyses. The heterogeneity is reflective of different diagnostic definitions for AS, dissimilar recruitment methods, and varying study periods (Table 1). Study participation was only 50% to 60% in 2 studies, making their results vulnerable for selection bias (5,15). In 1 study, AS was diagnosed using clinically indicated echocardiography (1). That might have caused a lower prevalence rate of AS. Moreover, improvements of echocardiographic techniques and interobserver variability might have had an influence on the prevalence rates and heterogeneity.
The number TAVR candidates
Nearly 40.5% of all patients with symptomatic severe AS did not undergo SAVR (Fig. 4B). Possible explanations for the lower than expected rates of SAVR include excessive operative risk, advanced age, comorbidities, and patient preference (21,22). TAVR is a safe, effective, and less invasive treatment strategy for a highly selected proportion of the patients who do not undergo SAVR (23), represented by the 40.3% of patients who underwent TAVR (Fig. 4C). The treatment decisions reflect heart team discussions, in which (interventional) cardiologists and cardiac surgeons combine risk models with additional factors such as frailty, porcelain aorta, and vessel tortuosity (24).
The estimated large number of TAVR candidates has clinical, economic, and social implications. If the index admission costs (US $72,000) of the PARTNER (Placement of Aortic Transcatheter Valves) trial are applied (25), treating all TAVR candidates would represent a budget impact of $13.7 billion in the European countries and $7.2 in North America. At a price of $30,000, the total device turnover would be approximately $8.7 billion. Although TAVR is cost effective in the United States for patients at high and prohibitive risk (25,26), data from other countries show that, for intermediate-risk patients, the costs of TAVR at 1 year are considerably higher than the costs of SAVR (27). Importantly, cost is not the only factor that determines the adoption of novel technologies such as TAVR (28). Reimbursement strategies, physician training, and health care culture may be related to the dissemination of this costly technology.
Despite budgetary concerns, current clinical trials are evaluating TAVR for patients at intermediate surgical risk (NCT01314313 and NCT01586910) (9,29). If TAVR proves to be noninferior to SAVR in this population, we estimate that a further 145,000 patients would become TAVR eligible. Indeed, there is some evidence that suggests that TAVR is already being performed in these intermediate-risk patients (18,30). Thus, our estimates of the impact of positive outcomes in the ongoing trials are likely to be conservative. In the future, TAVR may even compete with SAVR in patients at low surgical risk (30,31), a group that comprises 730,000 severe AS patients in the European countries and North America combined.
TAVR learning curve analyses show increasing proficiency with evidence of plateau after the first 30 cases (32). In addition, governmental bodies mandate that each TAVR center performs at least 20 to 50 TAVR procedures per year (33–35). These requirements, combined with the figures from this study, are useful to estimate the number of TAVR centers and physicians who need to be trained in TAVR in the individual countries. For example, the 526 (95% CI: 224 to 965) new TAVR candidates per year in the Netherlands justify approximately 10 certified centers, assuming that each center performs 50 cases annually. Similarly, the 8,205 (95% CI: 3,470 to 15,139) new TAVR candidates per year in the United States suggest a requirement of approximately 165 certified TAVR centers.
The divergent standards of medical evidence required to introduce new therapies in Europe and the United States are likely to account for the difference in TAVR dissemination between the continents (36). Although the Edwards Sapien valve (Edwards Lifesciences, Inc., Irvine, California) and Medtronic CoreValve (Medtronic, Inc., Minneapolis, Minnesota) both received the Conformité Européenne (CE) mark in 2007, the U.S. Food and Drug Administration used trial data to approve the Edwards Sapien valve for patients at prohibitive and high surgical risk only in November 2011 and October 2012, respectively. Consequently, TAVR has been performed with greater frequency and for a wider range of indications in Europe than in the United States. The studies on decision making in patients with AS reflect the commercial use of TAVR in Europe, whereas the U.S. studies display decision making in a time when TAVR use was restricted to clinical trials. These differences in practice are likely to disappear after the commercialization of TAVR in the United States and were taken into account in our sensitivity analyses.
Although we systematically searched the literature, relatively few reports on the prevalence of AS in the general population were identified. Additional population-based studies that use a unified echocardiographic definition of AS are warranted. The current study, however, reflects all of the currently available evidence on the prevalence of AS.
The estimation of TAVR candidates is as accurate as the currently available inputs and assumptions from the literature. However, we used sensitivity analyses to assess the influence of uncertain parameters. In addition, we included measures of uncertainty in each step of the model to calculate confidence intervals, representing the likelihood of the final estimates.
This systematic review and meta-analysis of population-based studies found that the prevalence of AS and severe AS among the elderly is 12.4%, and 3.4%, respectively. The overall burden of disease due to severe AS in the general elderly population is substantial. Our model showed that under the current indications approximately 290,000 elderly patients at high or prohibitive surgical risk could potentially be treated with TAVR in Europe and North America, and that each year there are approximately 27,000 new TAVR candidates. These estimates have considerable clinical, economic, and social implications.
The authors would like to acknowledge the suggestions by Rachele Busca, MS, PharmD, and Liesl C. Birinyi-Strachan, BS, PhD, from Medtronic.
For supplementary tables and references, please see the online version of this article.
Dr. van Mieghem is a member of the steering committee of the SURTAVI trial sponsored by Medtronic. Dr. Nkomo is co-investigator of the PARTNER II trial, sponsored by Edwards Lifesciences. Ms. LeReun received fees from Medtronic. Dr. Piazza is a proctor and consultant for Medtronic; and a member of the steering committee of the SURTAVI trial. Dr. Kappetein is a member of the steering committee of the SURTAVI trial. All other authors have reported they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- aortic stenosis
- confidence interval
- surgical aortic valve replacement
- The Society of Thoracic Surgery Predicted Risk Of Mortality
- transcatheter aortic valve replacement
- Received January 21, 2013.
- Revision received April 8, 2013.
- Accepted May 8, 2013.
- American College of Cardiology Foundation
- Stewart B.F.,
- Siscovick D.,
- Lind B.K.,
- et al.
- Bonow R.O.,
- Carabello B.A.,
- Chatterjee K.,
- et al.
- Vahanian A.,
- Alfieri O.,
- Andreotti F.,
- et al.
- Kappetein A.P.,
- Head S.J.,
- Genereux P.,
- et al.
- ↵Statistics Canada. 2012. Available at: http://www.statcan.gc.ca/start-debut-eng.html. Accessed November 16, 2012.
- ↵EUROSTAT. 2012. Available at: http://epp.eurostat.ec.europa.eu/portal/page/portal/population/data/database. Accessed November 16, 2012.
- United States Census Bureau 2011. Available at: http://www.census.gov/popest/data/national/asrh/2011/index.html. Accessed November 16, 2012.
- Lindroos M.,
- Kupari M.,
- Valvanne J.,
- Strandberg T.,
- Heikkila J.,
- Tilvis R.
- Lorz W.,
- Cottier C.,
- Gyr N.
- Brennan J.M.,
- Edwards F.H.,
- Zhao Y.,
- et al.
- Wenaweser P.,
- Pilgrim T.,
- Kadner A.,
- et al.
- Iung B.,
- Cachier A.,
- Baron G.,
- et al.
- Reynolds M.R.,
- Magnuson E.A.,
- Lei Y.,
- et al.
- Reynolds M.R.,
- Magnuson E.A.,
- Wang K.,
- et al.
- Mylotte D.,
- Osnabrugge R.L.,
- Windecker S.,
- et al.
- Lange R.,
- Bleiziffer S.,
- Mazzitelli D.,
- et al.
- Webb J.G.
- ↵Centers for Medicare and Medicaid Services. 2012. Decision memo for transcatheter aortic valve replacement. Available at: http://www.cms.gov/medicare-coverage-database/details/nca-proposed-decision-memo.aspx?ncaid=257&ver=5&ncaname=transcatheter+aortic+valve+replacement+(tavr)&bc=aaaaaaaaiaaa&. Accessed November 30, 2012.
- Haute Autorité de Santé. 2011. Transcutaneous aortic valve implantation by the transfemoral or transapical route. Available at: http://www.has-sante.fr. Accessed November 30, 2012.
- College voor zorgverzekeringen. 2011. Transcatheter aortaklepvervanging. Available at: http://www.cvz.nl/hetcvz/zoeken?query=transcatheter+aortaklepvervanging. Accessed November 30, 2012.