• aortic valve stenosis
• echocardiography
• epidemiology

The burden of aortic stenosis (AS) has been the subject of particular attention over the last decades due to growing evidence that, beyond its present impact in the elderly, AS is also likely to increase in the near future, with at least a predicted doubling of cases in the next 50 years.1

The analysis of the cumulative burden of AS performed by Owen et al 2 in the large cohort of elderly patients of the Cardiovascular Health Study (CHS) is therefore of particular interest. Previous analyses of the CHS cohort have led to major advances in the knowledge of the determinants and prognosis of calcific aortic disease, and the CHS presents unique features for the purpose of analysing the very long-term burden of AS. The CHS cohort was indeed prospectively designed for an accurate echocardiographic assessment of the aortic valve function, which was not current practice in a large patient sample in the late 1980s and early 1990s. The criteria used for the baseline evaluation of AS severity correspond to the contemporary criteria for high-gradient AS, which is the most frequent presentation of AS and for which there has always been strong indication for intervention in all guidelines. Finally, the CHS cohort was a population-based cohort of more than 5000 patients aged 65 or older, thereby avoiding selection bias inherent in hospital-based cohorts.

The main finding of the paper by Owen et al 2 is that the 20-year cumulative frequency of 3.7% for probable or definite clinically significant AS is somewhat higher than previously reported prevalence ranges from studies, which also included less severe AS presentations. When adding cases classified as possible AS, the 20-year cumulative frequency of AS increased to 4.7%. It is likely that the dynamic overview of the cumulative occurrence of AS in the overall range of an elderly population (age ≥65 at inclusion with a follow-up up to 25 years) can be considered as a more robust estimation of the true AS burden in the elderly than cross-sectional estimates taking into account simultaneously patients of different ages.

It is unlikely that the AS prevalence reported in the paper by Owen et al 2 might have been overestimated for methodological reasons. The use of hospital data for capture of events during follow-up may be subject to concerns to define clinically significant AS. However, the use of five different screening methods, the systematic review of hospitalisation records, and more importantly the availability of quantitative data on AS severity in more than 90% of cases considerably limit the risk of overdiagnosing AS, which would have occurred if using only hospitalisation discharge codes not taking into account AS severity. The use of the subdistribution survival method also precludes an overestimation of the cumulative frequency of AS since it takes into account the competitive risk of death. This is of particular importance in an elderly population with a very long follow-up. The competitive risk of death was considerable since 88% of patients died at the end of follow-up, and this would have resulted in an overestimation of the cumulative frequency of AS if using a more conventional cause-specific survival method. The respective burden of death and occurrence of AS during follow-up is clearly illustrated in figure 1, which also shows the increasing discrepancy between subdistribution and cause-specific estimates of the frequency of AS with increasing follow-up.

This higher cumulative frequency of clinically significant AS reported in the paper by Owen et al 2 raises many questions from public health and health economics perspectives. With population ageing, these numbers will markedly grow and the economic burden of the management and treatment of AS can only become greater. Providing the resources to afford such a high volume of patients is a major challenge that governments and healthcare providers surely have to anticipate and focus on already now. This necessary investment should not only concern the costly management of patients from diagnosis to treatment, but also aim to support basic and clinical research surrounding AS. It is also of utmost importance to take into consideration the vast disparities encountered across the world, even in industrialised countries, in terms of access to medical care and cost of treatment.3 Therefore, country-specific epidemiological studies, including cost and cost-effectiveness analyses, are warranted to anticipate the resources that will be required to face this challenge. It is also necessary to reduce procedure and hospital costs, for example, decrease in the price of implanted devices and initiatives for reducing length of hospital stay.4 An important lesson from the present study is that the true burden of AS in the community may be substantially underestimated.

Another striking finding of this analysis of the CHS cohort is the contrast observed in the rates of cumulative frequency and cumulated incidence between clinically significant AS and AS procedures. Approximately half of the patients who were hospitalised due to clinically significant AS did not undergo any intervention on the aortic valve. These findings are consistent with previous reports concluding to an underuse of surgery in the 2000s in patients referred to hospital, which corresponds to the follow-up time period of the CHS. The present findings suggest that underuse of intervention is even more frequent in the community. Although this can be expected, this is the first time that the underuse of surgery for AS is shown in a community setting. It is possible that certain patients did not fulfil class I indication for interventions or were contraindicated; however, it is unlikely that they accounted for half of the cases. Even in the absence of symptoms, patients with echocardiographic criteria corresponding to clinically significant AS are known to develop symptoms in the short term and should then undergo intervention.5 A recent survey showed that practices have improved in patients referred to hospitals, which can be attributed in particular to the dissemination of guidelines and the development of transcatheter aortic valve implantation.6

Some may argue that for the time being, knowing that the burden of incident and prevalent AS is higher than previously reported is futile, as a generalised and aggressive screening policy using systematic echocardiography to detect AS would not lead to any therapeutic decision that may significantly improve patient prognosis. The relationship between the risk factors for atherosclerosis and the development of AS is further confirmed in the long-term follow-up of the CHS cohort and is consistent with the current knowledge on the pathophysiology of AS. However, no strategy has been proven effective so far to slow the progression of AS. Indeed, the only treatment improving the prognosis of AS remains a timely intervention once the stenosis becomes severe and the patient symptomatic. Nevertheless, it is essential to bear in mind that basic and translational science researchers are currently making swift progress in the quest of deciphering the pathophysiology of the initiation and progression of valve calcification in AS. The bench-to-bedside research field has never been so dynamic, and the hope of identifying a therapeutic target within the complex pathophysiology of AS, and subsequently a pharmacological treatment, seems hopefully within reach.7 In this setting, quality epidemiological studies are essential to better capture the true burden of the disease and help identify risk subsets of the population who may benefit from echocardiographic screening and early pharmacological intervention that may suspend or slow down the natural history of AS in the future. Another interesting observation is the lower cumulative frequency of AS in the subgroup of patients of black ethnicity. Although consistent with previous reports,8 this result should nonetheless be cautiously interpreted as it may only be the reflection of a lower access of this subgroup of the population to medical care, thereby leading to underdetection of AS within the considered community, raising in the bigger picture a sensible issue in terms of public health. Indeed, collecting ethnicity-based data has been forbidden by ethics authorities in some countries, but such observations are valuable in the new era of personalised medicine to further identify subgroups at risk which may benefit from early screening and/or intervention.

It may be hoped that increased awareness towards valvular heart disease will decrease the number of undiagnosed cases1 and that improved guidelines implementation6 will improve patient care. These considerations, added to the findings from Owen et al 2 on a higher burden of AS than previously estimated, should be taken into account for healthcare planning and anticipating an increase in the consumption of healthcare resources.