Management of asymptomatic severe aortic stenosis: a systematic review and meta-analysis

Objectives The management of severe aortic stenosis mandates consideration of aortic valve intervention for symptomatic patients. However, for asymptomatic patients with severe aortic stenosis, recent randomised trials supported earlier intervention. We conducted a systematic review and meta-analysis to evaluate all the available data comparing the two management strategies. Methods PubMed, Cochrane and Web of Science databases were systematically searched from inception until 10 January 2022. The search key terms were ‘asymptomatic’, ‘severe aortic stenosis’ and ‘intervention’. Results Meta-analysis of two published randomised trials, AVATAR and RECOVERY, included 302 patients and showed that early intervention resulted in 55% reduction in all-cause mortality (HR=0.45, 95% CI 0.24 to 0.86; I2 0%) and 79% reduction in risk of hospitalisation for heart failure (HR=0.21, 95% CI 0.05 to 0.96; I2 15%). There was no difference in risk of cardiovascular death between the two groups (HR=0.36, 95% CI 0.03 to 3.78; I2 78%). Additionally, meta-analysis of eight observational studies showed improved mortality in patients treated with early intervention (HR=0.38, 95% CI 0.26 to 0.56; I2 77%). Conclusion This meta-analysis provides evidence that, in patients with severe asymptomatic aortic stenosis, early intervention reduces all-cause mortality and improves outcomes compared with conservative management. While this is very encouraging, further randomised controlled studies are needed to draw firm conclusions and identify the optimal timing of intervention. PROSPERO registration number CRD42022301037.


INTRODUCTION
Aortic stenosis is the valvular heart lesion with the highest clinical impact and mortality worldwide, accounting for half of valverelated deaths. 1 2 For patients with symptomatic severe aortic stenosis, European and American guidelines indicate that aortic valve replacement is advised as a class IB and class IA recommendation, respectively. 3 4 However, more ambiguity exists in the management of patients with asymptomatic severe aortic stenosis, where intervention is not recommended unless signs of adverse prognosis are present. These indicators include rapid

Key questions
What is already known about this subject?
⇒ According to the current guidelines and recommendations, patients with asymptomatic severe aortic stenosis are approached with the watchful waiting strategy unless signs of adverse prognosis or rapid disease progression are present. ⇒ The optimal timing of intervention has been debated as evidence suggests that when the adverse signs occur these may be irreversible and therefore it is already too late. ⇒ Recently, the largest randomised controlled trial was published comparing early intervention versus conservative management in this patient group.
What does this study add? ⇒ This systematic review and meta-analysis was conducted to critically assess the strengths and disadvantages of the two published randomised controlled trials on the subject. ⇒ Additionally, all observational studies in the literature were evaluated. ⇒ Within the limitations of the trials and the studies, early intervention appears to be associated with a favourable prognosis and lower all-cause mortality rates.
How might this impact on clinical practice? ⇒ The irreversible manifestations of severe aortic stenosis may occur before the presentation of symptoms or detectable myocardial impairment, which in turn may preclude worse prognosis and adverse outcomes. ⇒ The results of this meta-analysis may herald the beginning of change in the management of these patients. ⇒ Nevertheless, ongoing trials investigating this highrisk population are anticipated to shed more light into the matter and the identification of the optimal time of intervention.
progression of the valve stenosis, severe valve calcification, valve parameters reflecting the 'very severe' end of the spectrum of the disease, or signs of left ventricular impairment by echocardiographic data or blood biomarkers, although strong evidence for these is lacking. 3 4 However, many would argue that by the time the left ventricular myocardium shows direct or indirect signs of impairment, it is already too late, with irreversible damage established 5 and the 'therapeutic window' for intervention missed. Additionally, studies have shown that most patients with severe asymptomatic aortic stenosis exhibit symptoms in the first 2 years following the diagnosis. Many deaths however occur in the period of 'watchful waiting'. 6 7 These observations have caused much controversy among clinicians and highlight the need for improved risk stratification models and better individualised management strategies according to each patient's risk.
The optimal timing of aortic valve intervention in asymptomatic patients with severe aortic stenosis is yet to be ascertained. Undoubtedly, after the results of the Aortic Valve ReplAcemenT versus Conservative Treatment in Asymptomatic SeveRe Aortic Stenosis (AVATAR) trial, 8 many would argue whether the time for a change in clinical practice is fast-approaching. We therefore conducted a systematic review and meta-analysis of randomised and observational studies comparing early intervention versus conservative management of patients with asymptomatic severe aortic stenosis to synthesise and evaluate the current evidence.

METHODS
This systematic review and meta-analysis was conducted and reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines 9 and has been submitted and registered with PROSPERO (registration number: CRD42022301037). We performed a focused review and meta-analysis of all the randomised trials in the literature comparing aortic valve intervention with conservative management in patients with asymptomatic severe aortic stenosis. The primary endpoints of this meta-analysis included all-cause mortality, cardiovascular death and hospitalisation for heart failure. Additionally, using the same search results we separated the observational studies and performed a meta-analysis of all the published observational studies comparing outcomes, identified through the same search, between early intervention and conservative management in patients with asymptomatic severe aortic stenosis, with all-cause mortality being the primary endpoint.

Search strategy
PubMed, Cochrane and Web of Science databases were systematically searched from inception until 10 January 2022. The terms used for the search were 'asymptomatic', 'severe aortic stenosis' and 'intervention'. After removing duplicates, two investigators independently screened the remaining studies at the title/abstract level (VT, CG-C). Randomised controlled studies comparing mortality and clinical outcomes in patients with asymptomatic severe aortic stenosis treated with early intervention versus conservative management were included in the metaanalysis. Observational studies that compared mortality between early intervention and conservative management in patients with asymptomatic severe aortic stenosis were evaluated separately. The selected studies underwent full-text screening, performed by two independent investigators (VT, CG-C). The following inclusion and exclusion criteria were applied: Inclusion criteria ► Studies published in English and limited to humans. ► Studies with participants >18 years old. ► Randomised controlled studies and observational studies comparing mortality and clinical outcomes in patients with asymptomatic severe aortic stenosis treated with early intervention versus conservative management.
Exclusion criteria ► Studies published in language other than English. ► Studies including patients with symptomatic aortic stenosis. Conflicts were resolved by discussion with a third investigator (VSV), after which consensus was achieved. The primary outcomes of interest were all-cause mortality, cardiovascular death and hospitalisation for heart failure. The study selection process is shown in online supplemental figure 1.

Data analysis
The HR and 95% CI reported in each study were used for the meta-analysis. A random-effects model with inversevariance weights was used to combine the effect measures from all studies on a logarithmic scale. Statistical heterogeneity was assessed using I 2 statistics. Statistical analyses were conducted using the Review Manager (RevMan) software (V.5.3; Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014). Where appropriate funnel plots are also given. Statistical significance was defined as p<0.05.

Patient and public involvement
Given that this is a systematic review and meta-analysis of already published data, patients or the public were not involved in the design and conduct of this study.

RESULTS
The search yielded 1046 studies. After removing the duplicates and applying the inclusion and exclusion criteria, 71 articles underwent full-text evaluation. Out of these, two randomised controlled trials, the AVATAR trial 8 and the Randomized Comparison of Early Surgeryversus Conventional Treatment in Very SevereAortic Stenosis (RECOVERY) trial, 10 with a total of 302 patients, met the Valvular heart disease inclusion criteria and their data were extracted for metaanalysis. Of these patients, 151 were randomised to the early intervention group and 151 were randomised to the conservative management group. The two trials included patients with asymptomatic severe aortic stenosis of different aetiologies. The majority of the patients (85%) in the AVATAR trial had degenerative aortic valve stenosis. In contrast, only 33% of the patients had degenerative aortic stenosis in the RECOVERY trial and 61% had bicuspid aortic valve disease. Additionally, the inclusion criteria of the two trials were slightly different, as in the RECOVERY trial patients with very severe aortic stenosis (as evidenced by aortic valve area of <0.75 cm 2 , maximal velocity across the aortic valve (Vmax) >4.5 m/s or mean gradient >50 mm Hg) were included and an exercise test was performed only in selected cases. The main features of the two studies are summarised in table 1, and the baseline demographic and echocardiographic patient characteristics of the two trials are shown in table 2. The risk of bias assessment of these two trials is demonstrated in online supplemental table 1.
Despite the different endpoints of the trials, both demonstrated a reduced number of deaths from any cause in the group of patients treated with early surgery. More specifically, in the AVATAR trial there were 9 deaths from any cause in the early intervention group and 16 deaths in the conservative management group. Similarly in the RECOVERY trial there were 5 deaths in the early intervention group and 15 deaths in the conservative management group. There was a similar pattern in the number of hospitalisations for heart failure, with reduced number of hospitalisations in the early intervention groups in both trials compared with the conservative management groups. The raw number of events provided by each trial for all-cause death, hospitalisation for heart failure, myocardial infarction and stroke is depicted in online supplemental table 2.
The pooled meta-analysis showed a significant reduction in all-cause mortality in patients treated with the early intervention compared with those managed conservatively with the watchful waiting approach (HR=0.45, 95% CI 0.24 to 0.86; I 2 0%) (figure 1). Furthermore, the early intervention resulted in a significantly reduced risk of hospitalisation for heart failure (HR=0.21, 95% CI 0.05 to 0.96; I 2 15%) (figure 2). There was significant heterogeneity in the meta-analysis of cardiovascular mortality among the two studies, and while there was a trend towards improved survival this was mainly driven by the RECOVERY trial and did not reach statistical significance (HR=0.36, 95% CI 0.03 to 3.78; I 2 78%) (online supplemental figure 2).

Meta-analysis of observational studies
Additionally, eight observational studies, including 3496 patients, were analysed separately to assess how early intervention affects the risk of all-cause mortality in patients with asymptomatic severe aortic stenosis. 7 11-17 Table 3 summarises the main characteristics of each study, while the risk of bias assessment is demonstrated in online supplemental table 3. Of the total population, 1354 patients had an early aortic valve replacement and 2142 were managed conservatively with the watchful waiting approach. The metaanalysis from these studies showed that early intervention was associated with a 62% reduction in all-cause mortality (HR=0.38, 95% CI 0.26 to 0.56), although heterogeneity was high (I 2 77%) (figure 3). Online supplemental figure  3 shows the funnel plot for risk of all-cause mortality from these eight studies.

DISCUSSION
Our report represents the first meta-analysis of the two published randomised controlled trials assessing the impact of early intervention on the outcomes of patients with asymptomatic severe aortic stenosis. It demonstrates that all-cause mortality and hospitalisation for heart failure are significantly lower in patients randomised to early surgery compared with those treated with an initial conservative approach. Regarding cardiovascular mortality, while there was a trend towards improved survival in the early intervention group, this was mainly driven by the RECOVERY trial and was not found to be significantly different between the early and conservative groups. Additionally, a meta-analysis of eight observational studies showed a notable reduction in all-cause mortality in the early intervention group. This is in agreement with previous meta-analyses that were conducted even before the AVATAR trial was published, which showed a favourable outcome in those treated with early intervention versus conservative management. 18 The timing of intervention in patients with severe aortic stenosis remains contentious. The lack of large randomised controlled studies has been a 'deafening silence' in cardiology, with current practice often informed by observational studies, which despite their limitations timidly show a favour towards early intervention. Nonetheless, this needs to be balanced by the potentially longer need for a valve replacement, which in itself can come with risks of bleeding, endocarditis and failure. Several unanswered questions have long been discussed among clinicians. Current guidelines suggest intervention in patients with severe aortic stenosis when symptoms occur or in those without symptoms and left ventricular ejection fraction (LVEF) <50% or abnormal exercise test or Vmax >5 m/s or brain natriuretic peptide (BNP) more than three times the upper normal value or features of rapid disease progression. 3 19 20 However, it can be challenging to unmask these on many occasions, especially in patients with sedentary life or in the frail elderly population. Additionally, the presence of left ventricular impairment represents a high-risk feature that mandates intervention in these patients. Still, some argue that it may already be too late when the LVEF drops. 21 22 So far, there has been an ongoing debate with studies demonstrating a low risk of death in patients with asymptomatic aortic stenosis unless there is very severe aortic stenosis or signs of rapid progression or myocardial impairment. 23 24 Nevertheless, a retrospective observational study by Campo et al 16 demonstrated that almost half of the asymptomatic patients with severe aortic stenosis would have either died or undergone surgery within 2 years of the initial recommendation for the watchful waiting strategy. 15 This is in keeping with another retrospective registry by Pellikka et al, 25 stating that 75% of these patients will have died or have been operated on 5 years after the diagnosis. However, evidence suggests that, although this risk is low, it is still higher than the general population and this clinical entity may not be as benign as previously thought. 25 26 The two recently published studies, AVATAR and RECOVERY trials, have been enthusiastically awaited to answer some of these questions. Arguably, the two trials, while they do have differences as mentioned in table 2, both demonstrate a consistent benefit of the early intervention approach. However, there are essential points that need to be considered. The two trials have had some important differences in their study characteristics, with the RECOVERY trial including patients with more severe aortic stenosis. This could be one of the reasons why the RECOVERY trial demonstrated a more pronounced reduction in mortality and morbidity in favour of the early intervention strategy, as it is known that these patients are at higher risk of death and adverse prognosis. 27 28 One of the most notable things, however, is that rapid progression of aortic stenosis in the RECOVERY trial was defined as a change of 0.5 m/s per year, which is in contrast to what the current 4 and the previous guidelines, 29 on which the trial was based, defined as rapid progression of the disease. This, of course, raises the question as to whether there were participants in the conservative arm that fulfilled the criteria for intervention. Moreover, the patient population in the RECOVERY trial consisted mainly of patients with aortic stenosis secondary to the bicuspid aortic valve. Subsequently, participants of this trial were younger with low operative risk compared with the AVATAR trial. 10 On the other hand, the majority of the patients included in the AVATAR trial had aortic stenosis secondary to degenerative disease. However, the operative risk was low as it was in the RECOVERY trial. 8 Furthermore, while in the RECOVERY trial exercise test was used in selected cases only, all participants of the AVATAR trial underwent an exercise test to unmask symptoms or adverse signs before their inclusion. However, it is unclear if and how the exercise test was used in the AVATAR trial during the follow-up of these patients.
In addition to the above, whereas this meta-analysis showed a significant reduction in all-cause mortality in the early intervention group, this was not the case for cardiovascular death, where high heterogeneity was noted and no significant difference was demonstrated between the two management strategies. While the issues discussed above may have contributed to this discrepancy, perhaps one of the most important factors that should be considered is the follow-up period. The follow-up periods of the two studies are substantially different, with the median follow-up of the RECOVERY trial reaching 6.1 years, more than twice the AVATAR trial follow-up. On close examination, it is noted that the incidence of cardiovascular death between the two treatment groups

Valvular heart disease
in the RECOVERY trial becomes substantially different after approximately the first 2.5 years of the follow-up period. In the conservative group the cumulative incidence of cardiovascular death rises significantly, reaching a peak at 8 years after randomisation, while the intervention group demonstrates only a small rise. 10 Taking into consideration that the median follow-up of the conservative group in the AVATAR trial was just over 2.5 years (35 months), one could argue that a difference in the incidence of cardiovascular death between the two trials is expected. Previous data from a retrospective registry suggest that mortality between the two groups of patients (those treated with early surgery and those conservatively managed) begins to significantly differ after 1 year of follow-up. 13 It is important to note that both studies were performed at different time points in the past. Therefore, the inclusion criteria and characteristics of the study population may not reflect the most recent guidelines and recommendations. For instance, according to the latest European Society of Cardiology (ESC) guidelines, intervention should be considered in asymptomatic patients with low operative risk, LVEF >55% and Vmax ≥5 m/s or in asymptomatic individuals with LVEF <55% without another cause (class IIa recommendation). 3 Given the inclusion and exclusion criteria of the AVATAR and RECOVERY trials, some of the participants included in both studies unavoidably fall in these categories.
The meta-analysis from the observational studies yielded a categorically favourable outcome of the early intervention group compared with the conservative management. However, given their non-randomised design, the risk of bias is high. Therefore, conclusions should be drawn cautiously. Additionally, there was considerable heterogeneity in the meta-analysis from these studies that could be attributed to the different study designs and analysis methods, inclusion criteria and baseline patient characteristics, and different follow-up periods.  Finally, when putting the findings of these metaanalyses into clinical context, the baseline characteristics of the patient population included, and more specifically the age, should be taken into consideration. The participants of both trials and of some of the observational studies were relatively young, with a mean age of less than 70 years old. This may not fully reflect the cohort of patients usually seen in secondary or tertiary care, which may consist of older patients with more comorbidities and therefore higher operative risk. 30 Despite differences and limitations of the studies, all the data from this meta-analysis would support consideration of an earlier approach and highlight, however given the differences and limitations in the two studies, it would be prudent to allow the ongoing research in the field to complete. Nonetheless, this meta-analysis gives us an early insight as to what outcome we could expect. It is noteworthy that in a large registry that included more than 25 000 patients with aortic stenosis, the individuals with moderate aortic stenosis exhibited a surprisingly high risk of all-cause and cardiovascular deaths, which was similar to those with severe aortic stenosis. 31 Whether this finding means that some patients with parameters in the moderate spectrum of the disease are at high risk of death from the valvular disease alone or that, inevitably, comorbidities in this cohort play a much more significant role than previously thought remains unclear. Nevertheless, in the era of personalised medicine, the identification of the high-risk population and the decision regarding intervention should be made after careful individual risk assessment and evaluation not only of symptoms but also of indirect signs of myocardial decompensation and damage that could have a negative impact on the prognosis and outcome. Such parameters could include the presence of mid-wall late gadolinium enhancement of cardiovascular magnetic resonance, utilisation of left ventricular stroke volume in preference to LVEF or various biomarkers (eg, N-terminal pro b-type natriuretic peptide/BNP, highsensitivity troponin), which are associated with a worse outcome. [32][33][34][35] These clues could help refine the optimal timing of intervention for each patient. [35][36][37] The results of currently ongoing randomised trials investigating these parameters are greatly anticipated to help clarify the role of these parameters in the risk assessment of patients with severe aortic stenosis and the impact of early intervention on the outcome and prognosis. The main characteristics of these trials are summarised in table 4.

CONCLUSION
Asymptomatic severe aortic stenosis represents a unique entity that poses clinical dilemmas to physicians worldwide very frequently in clinical practice. This metaanalysis shows that the data from the two recent large randomised controlled trials and previous observational studies demonstrate a favourable outcome in the group of patients treated with early intervention rather than conservative management. Although this may herald the beginning of a change in the management of these patients, further randomised controlled studies are needed to draw firm conclusions and identify the optimal timing of intervention. Contributors VT designed the work, collected and analysed the data, and drafted the first manuscript. CG-C collected the data and substantially revised the first draft. AEGR, GA, PG, SP, LR, GPM, JR and MRD substantially revised the manuscript. VV conceived the idea, designed the work and substantially revised the manuscript. VT, CG-C and VSV had full access to the data and take responsibility for the integrity of the data and the accuracy of the data analysis. VT and VSV accept full responsibility for the work and the conduct of the study, had access to the data, and controlled the decision to publish.
Funding This work was partly funded by Wellcome Trust grants (220703/Z/20/Z). VT and CG-C are NIHR Academic Clinical Fellows. JR is part-supported by the NIHR Cambridge Biomedical Research Centre, the British Heart Foundation, HEFCE, the EPSRC and the Wellcome Trust. MRD is supported by the British Heart Foundation (FS/14/78/31020) and is the recipient of a Sir Jules Thorn Award for Biomedical Research 2015 (15/JTA).

Competing interests None declared.
Patient consent for publication Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data availability statement All data relevant to the study are included in the article or uploaded as supplementary information.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.
Open access This is an open access article distributed in accordance with the Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits others to copy, redistribute, remix, transform and build upon this work for any purpose, provided the original work is properly cited, a link to the licence is given, and indication of whether changes were made. See: https://creativecommons.org/ licenses/by/4.0/.