Therapeutic decisions for patients with symptomatic severe aortic stenosis: room for improvement?☆ Free

Abstract

1 Introduction

The prevalence of aortic stenosis increases with age to up to 8% in the elderly [1]. Meanwhile the Western population increases to age during the last decades and this trend is expected to continue [2]. Therefore aortic stenosis constitutes a growing health burden.

While the treatment of asymptomatic patients with severe aortic stenosis remains debatable, both European and American guidelines on the management of valvular heart disease recommend that symptomatic patients have aortic valve replacement [3,4]. This recommendation is not only based on the survival advantage that can be expected after surgery but also on the improvement in functional class, even in elderly patients [3–5].

Recent literature suggests that a considerable proportion (33–60%) of patients with symptomatic severe aortic stenosis do not receive aortic valve replacement (AVR) [6–9]. We sought to confirm that many symptomatic patients remain unoperated and were interested in the reasons and the consequences of the decision to operate or not. The goal of our study therefore was to gain insight into decision making and survival in patients with severe symptomatic aortic stenosis.

2 Methods

2.1 Study design and data collection

A retrospective search in the echocardiography database of our department revealed 115 patients with severe aortic stenosis. An additional 140 patients were recruited from the echocardiography laboratories in the outpatient cardiology clinics of 7 hospitals in the Rotterdam region. All echocardiograms were made between October 2004 and December 2007. Patients had at least one of the following inclusion criteria: aortic valve area <1.0 cm², maximum aortic jet velocity >4.0 m/s, peak aortic gradient >64 mmHg or mean aortic gradient >40 mmHg. To avoid missing low-output aortic stenosis, patients were also included if the ratio between the velocity time integral over the aortic valve and the left ventricular outflow tract was >4.0.

Information was gathered on medical history, cardiovascular risk factors, and symptomatic status at the time of the echocardiogram. Asymptomatic patients were excluded from the eventual analysis. For all symptomatic patients, anticipated operative risk was calculated using the logistic EuroSCORE risk model (www.euroscore.org).

Treatment strategies and their reasons were retrieved from notes in the patients’ medical charts. Reasons for ‘conservative/medical treatment’ were classified in six main categories: (1) anticipated high operative risk (including advanced age or left ventricular dysfunction); (2) only mild symptoms; (3) stenosis non-severe; (4) patient preference; (5) decision not final yet; and (6) other, including ‘reason unclear’.

The study protocol was approved by the institutional review board, patient informed consent was waived (MEC 06–066, MEC 08–022). The authors had full access to the data and take responsibility for its integrity. All authors have read and agree to the manuscript as written.

2.2 Study population

Of the 255 patients that were initially identified, 73 asymptomatic patients were excluded plus 3 patients, of whom symptomatic status could not be retrieved, leaving 179 symptomatic patients in the study cohort. Mean age was 71 years, 50% were male.

During follow-up (mean 17 months, median 13.6, range 0.1–40) 76 patients (42%) underwent AVR or were scheduled for surgery (Fig. 1 ). There were 63 conventional aortic valve replacements, 9 percutaneous and 1 transapical valve implantations. Two patients were on a waiting list for AVR and one patient required a heart transplantation during follow-up. Medical treatment was given in 101 patients (56%). Two patients were lost to follow-up (99% completeness). Mean age of the surgical group was 68 years versus 73 years for the medically treated patients (p = 0.004). Predicted operative mortality according to the logistic EuroSCORE was 7.8% versus 11.3% (p = 0.009). More patient characteristics are given in Table 1 .

Fig. 1

Flow chart of main results.

Open in new tabDownload slide

Table 1

Patient characteristics.

Open in new tabDownload slide

2.3 Statistical analysis

Continuous data are presented as mean ± 1 standard deviation, and median. Categorical data are presented as proportions.

Chi-square testing was used for comparison of categorical variables. Continuous variables were compared using the Student’s t-test. A p value <0.05 was considered significant.

Survival curves were estimated by the Kaplan–Meier method. Differences in survival were not statistically assessed.

Statistical analyses were performed with SPSS for Windows (release 15.0; SPSS Inc., Chicago, Illinois).

3 Results

There was no 30-day mortality. During follow-up 12 patients died in the surgical group, versus 28 patients in the medical group. One- and 2-year survival was, respectively, 93% and 90% for the AVR group and for the conservative group 77% and 69% (Fig. 1).

Reasons for choosing non-surgical treatment were: operative risk deemed ‘too high’ (34%), symptoms regarded as ‘mild’ (19%), stenosis regarded as ‘non-severe’ (14%), and patient preference (9%). In 5% the decision to operate was still under consideration by cardiologist and/or patient. In 20 patients (20%) the reason behind decision making could not be retrieved accurately. Of the latter 20 patients, 11 were in NYHA class II, 6 were in NYHA III, and 3 were in NYHA class IV.

Of the 34 patients in whom the reason not to operate was ‘high risk’, the mean age was 75.7 years and the mean EuroSCORE was 11.6%. Eight of them had a history of malignancy or active malignancy (six of these patients eventually died during follow-up). Eighteen patients had a EuroSCORE <10% and only 9 of the 34 patients in whom the operative risk was deemed too high had a EuroSCORE >15%.

4 Discussion

Although treatment consensus seems to exist on symptomatic patients with severe aortic stenosis, it is not uncommon to diverge from these guidelines [6,8–10]. Advanced age and left ventricular dysfunction are known reasons to deny surgery in a symptomatic patient [6,11]. Instead of using patient characteristics to predict whether a patient gets AVR or not, our study was designed to investigate the decision making. Therefore it provides a different perspective: in our cohort an overestimation of operative risk, underestimation of symptoms, and misclassification of hemodynamic severity are common causes why symptomatic patients are denied AVR. Furthermore, we found that survival of the conservative group is not as pessimistic as reported by others [12,13].

4.1 ‘Overestimation’ of operative risk?

In a third of the patients who were treated conservatively, an anticipated high operative risk was the main reason not to go for AVR. This subgroup had a mean age of only 76 years, and only 9 of the 34 patients had a EuroSCORE >15%. Perhaps it is even more important that more than half (18 patients) had a relatively low operative risk with a EuroSCORE <10%.

From the literature it is known that remission of symptoms after starting medical treatment can be a reason to stay conservative and that patients who are treated conservatively are generally older and more often have impaired left ventricular function than surgically treated patients [6,7]. Yet, both remission of symptoms, advanced age, and depressed left ventricular function are debatable reasons not to operate on a symptomatic patient. Even elderly patients can be operated upon with acceptable morbidity and mortality, and can expect a considerable quality of life [5,11].

Note that 10 patients in the AVR group underwent a minimally invasive valve replacement. They were deemed not amenable for surgery. This indicates that even in a region with a tertiary center that uses new percutaneous and transapical techniques to replace the aortic valve, the majority of patients are treated conservatively.

Eight patients had either a malignancy in medical history or an active malignancy, risk factors which are not taken into account by the EuroSCORE. Another issue with risk models is that they do not score characteristics such as vitality or biological age. Furthermore there is a large variability between different risk models, and the one most commonly used (EuroSCORE) seems to overestimate the actual operative risk most [14]. Perhaps this adds to the large variance in treatment advice that exists among cardiologists which was already found by Bouma et al. [7].

4.2 Underestimation of symptoms?

Due to inactivity or gradual adjustment of daily activities to developing symptoms, patients with aortic stenosis often do not acknowledge the presence of symptoms or attribute them to the ageing process. Exercise testing is recommended in asymptomatic patients with aortic stenosis in order to exclude symptoms with more certainty [15–17], and up to 37% of patients previously considered asymptomatic have limiting symptoms when they are tested [17]. According to the European Heart Survey exercise testing is highly underused [3,18]. This could lead to an underestimation of the proportion of symptomatic patients treated medically that was reported by others and in the current study [6,8–10].

In this study, the classical aortic stenosis symptoms such as dyspnea, syncope, or angina were documented for several patients but regarded as mild or non-debilitating. Having only mild symptoms does not exclude a patient from being an AVR candidate [3,4]. It is furthermore known that even if symptoms are recognized, the resulting functional disability is often underestimated by physicians [19]. Symptomatic patients with severe aortic stenosis from our cohort suffer from both physical and emotional impairment hampering normal daily activities (unpublished data). These are clear reasons to assess symptomatic status accurately, and to reconsider a conservative approach when symptoms are present.

4.3 Underestimation of hemodynamic severity

As much as 14% of the symptomatic patients who were denied surgery were not referred because the stenosis was classified non-severe by the treating cardiologist during the initial assessment. According to the guidelines they should however have been classified as severe [3,4]. Since only patients with a severe stenosis are recommended to have surgery, these misclassified patients are at increased risk of left ventricular deterioration and sudden death [20].

Even if the stenosis severity is only just below the severe threshold, it can be disputed that watchful waiting is the best treatment. Peak aortic gradient increases 10–15 mmHg/year and aortic valve area decreases 0.1–0.12 cm²/year [21–23]. Given these progression rates, borderline patients will enter the severe category within a few months or at most a year later. Meanwhile left ventricular function will only get worse.

4.4 Survival in the conservative and in the surgical group

Survival in the medically treated group cannot easily be compared with the surgically treated group because the patients have quite different characteristics, which could account for a large part of the difference in survival. It is therefore questionable if, and to what extent, the survival of the total study group would have improved if more patients had aortic valve replacement.

From the survival curve of the non-AVR group it can be seen that a decline in survival already occurs in the first year after the echocardiogram (Fig. 2 ). Still, survival in the conservative group is not as bad as expected based on previous reports [12,13,20]. Perhaps improvement in medical treatment over the past years plays a role, but survival in the conservative group highly depends on referral strategy as well; if more high risk patients are operated upon, the patients with a really bad prognosis are left for conservative treatment, resulting in low survival in this category. Therefore the relatively good prognosis of our medically treated group could be a reflection of the conservative approach of the cardiologists in our region.

Fig. 2

Kaplan–Meier survival for the conservatively treated group and the AVR group.

Open in new tabDownload slide

Because of its dependence on referral, natural history of aortic stenosis is very difficult to study. If one would like to gain a clear view on natural history, theoretically all eligible patients should be excluded from having AVR, or they should be randomized to receive either surgical or conservative treatment. In practice this would be impossible and ethically incorrect.

4.5 Future prospects

Microsimulation methods can accurately estimate life expectancy for patients after AVR [24,25], but have yet to be developed for patients who are treated conservatively. Our department intends to develop these models, but this requires large datasets with extensive numbers of variables and some patient factors, such as vitality, will be difficult to grasp in a model.

5 Conclusion

A considerable proportion of patients with symptomatic severe aortic stenosis are not referred for surgery although theoretically they have an indication for aortic valve replacement. Often operative risk is estimated (too) high, and misclassification of both hemodynamic severity and symptomatic status occurs frequently.

Most patients who were treated conservatively were simply not referred to a surgical department. Referral to surgical departments should be encouraged in order to have more interdisciplinary team discussions between cardiologists and surgeons. Hopefully, this will result in better patient selection for surgery, possibly resulting in better survival of patients with severe symptomatic aortic stenosis.

Acknowledgements

The authors would like to thank the patients, cardiologists, echo laboratory workers and secretaries of the following participating hospitals for their kind cooperation: Havenziekenhuis, Rotterdam; St. Franciscus Gasthuis, Rotterdam; IJsselland Ziekenhuis, Capelle aan den IJssel; Vlietland Ziekenhuis, Vlaardingen; Albert Schweitzer Ziekenhuis, Dordrecht; Medisch Centrum Rijnmond Zuid, Rotterdam; and Erasmus University Medical Center, Rotterdam.

References