Author + information
- Received May 31, 2005
- Revision received August 5, 2005
- Accepted October 17, 2005
- Published online March 7, 2006.
- Pilar Tornos, MD⁎ (, )
- Antonia Sambola, MD,
- Gaietà Permanyer-Miralda, MD,
- Arturo Evangelista, MD, FESC,
- Zamira Gomez, MD and
- Jordi Soler-Soler, MD, FESC, FACC
- ↵⁎Reprint requests and correspondence:
Dr. Pilar Tornos, Servei de Cardiologia, Hospital Universitari Vall d’Hebron, P Vall d’Hebron 119-129, 08035 Barcelona, Spain.
Objectives The purpose of this study was to compare postoperative outcome in two groups of patients with chronic severe aortic regurgitation (AR): those operated on early and those operated on late according to the guidelines.
Background The impact of earlier surgery for chronic severe AR as defined in guidelines has not been evaluated.
Methods A total of 170 patients with chronic severe AR submitted to aortic valve replacement were prospectively followed up. Patients were divided in two groups depending on the clinical situation at the time of surgery. Group A were 60 patients who were operated on following guidelines advice of earlier surgery, and group B were 110 patients who were operated on late with regard to guideline recommendations.
Results Follow-up was 10 ± 6 years (1 to 22 years). During follow-up 44 patients died, 7 patients (12%) from group A and 37 (37%) from group B (p = 0.001). The cause of death was non-cardiac in 11 patients, 2 (3%) in group A and 9 (8%) in group B. Cardiac deaths occurred in 33 patients, 5 (9%) from group A and 28 (28%) from group B (p = 0.002). Causes of death differed between groups A and B: heart failure or sudden death were significantly more frequent in group B (20 patients vs. 1 patient, p = 0.001). Overall survival in groups A and B was 90 ± 4% vs. 75 ± 8% at 5 years, 86 ± 5% vs. 64 ± 5% at 10 years, and 78 ± 7% vs. 53 ± 6% at 15 years, respectively (p = 0.009).
Conclusions Early operation as defined in the guidelines improves long-term survival in patients with chronic AR.
In recent years the natural history of chronic aortic regurgitation (AR) has been well described, and several studies have clearly shown the predictors of unfavorable outcomes (1–5). It has been shown that a decrease of the left ventricular ejection fraction (LVEF) at rest increases the probability of symptoms at a rate of 25% per year (5), an end-systolic diameter (ESD) >50 mm at a rate of 19% per year (5), and more recently that changes in LVEF and ESD predict clinical deterioration or death at an annual rate of 10% to 20% (6). On the other hand, postoperative outcome depends on preoperative symptoms, preoperative LVEF, and LV dilation (7–9). In the American College of Cardiology/American Heart Association (ACC/AHA) guidelines (10), surgery is considered a Class I recommendation in all symptomatic patients and in asymptomatic patients when LVEF at rest is lower than 49%, and a Class II recommendation in asymptomatic patients with ESD >55 mm. In the recommendations of the management of asymptomatic patients with valvular heart disease from the European Society of Cardiology (ESC) (11), surgery is recommended in severe AR with resting LVEF <50% or with enlarged ventricles end-diastolic diameter (EDD) >70 mm and ESD >50 mm.
The aim of these recommendations is to offer a favorable postoperative long-term course. Although guidelines have been available for some years, the real impact of earlier surgery as defined in guidelines has not been evaluated, and there is some concern about the real benefit of surgery in the asymptomatic patient (12–14).
In the present study we analyze the impact of adherence to guidelines recommendations regarding earlier surgery on the long-term postoperative outcome in a large series of patients who were operated on in a single institution for isolated chronic severe AR according to a predefined protocol established in 1982 (15), similar to the subsequent ACC/AHA and ESC guidelines, which were designed in the late 1990s.
A total of 170 patients with severe isolated chronic AR and without significant coronary heart disease were consecutively submitted to aortic valve replacement in our institution between 1982 and 2002 according to a predefined protocol (15) that advised surgery for all symptomatic patients and asymptomatic patients when ejection fraction (EF) <50% and/or ESD is between 50 and 55 mm. No patient was denied surgery because of preoperative very severe LV dysfunction. Data regarding preoperative New York Heart Association (NYHA) functional class, symptoms, and echocardiographic parameters were analyzed. Patients were divided in two groups depending on the clinical situation at the time of surgery. Group A (“early” surgery) included asymptomatic patients with moderate degrees of LV dysfunction (LVEFs between 45% and 50% and/or ESDs between 50 and 55 mm) and patients in NYHA functional class II. Group B (“too late” surgery) included patients either severely symptomatic (NYHA functional class III and IV) or with an LVEF <45% or an ESD >55 mm. All patients were operated on by the same surgical team. After surgery all patients were prospectively followed up by a single cardiologist (P. T.) that followed both groups of patients in an exactly similar manner, with the same clinical visits and drug therapy orientation in case they had heart failure or any other problem. Regular visits were scheduled at three months and yearly thereafter, and echocardiograms were performed after one year and every five years. This study was approved by an institutional review committee and all subjects gave informed consent.
Descriptive analysis was performed with mean values ± SD and range for continuous variables and absolute and relative frequencies of patients in each category for categorical variables. Differences between study groups were evaluated with the use of unpaired Student ttest for continuous variables and chi-square test for categorical variables. Paired ttest was used when appropriate. The probability of survival during follow-up was calculated by Kaplan-Meier analysis; the survival curves obtained were compared with the log-rank test (Mantel-Cox). Two Cox models were developed to assess the relative risk of cardiovascular death and death due to heart failure or sudden death. An additional analysis was also performed with the same methodology considering the two study groups only on the basis of NYHA functional class (I to II vs. III to IV), irrespective of the LVEF or LV dimensions of the patients.
A two-tailed value of p < 0.05 was considered significant. Statistical analysis was performed with the statistical package SPSS 11.0 (SPSS Inc., Chicago, Illinois).
The study population is 170 patients, 135 men and 35 women, ages from 14 to 78 years, with a mean age of 50 ± 14 years (median: 53 years). All patients over 45 years had a normal coronary angiography. Group A includes 60 patients: 26 (43%) were asymptomatic patients with moderate LV dysfunction, and 34 patients (57%) were in NYHA functional class II. Group B includes 110 patients: 44 had either an LVEF <45% or an ESD >55 mm with either no symptoms or moderate symptoms, and 66 patients had severe symptoms (NYHA functional class III or IV). Demographic characteristics of the whole series are shown in Tables 1 and 2.⇓All patients received an aortic prosthesis, 101 (59%) mechanical and 69 (41%) biological.
Postoperative clinical outcome
Hospital mortality was 12 patients (7%), 3 patients from group A (5%), and 9 (8%) from group B (p = 0.5). Patients from group A who died were two patients with renal insufficiency (one with previous renal transplant and one in hemodialysis program) and one patient who had infection of the sternal wound and secondary sepsis.
Follow-up was 10 ± 6 years (1 to 22 years) and only one patient was lost to follow-up, and another was lost after 10 years. During follow-up 44 patients died, 7 patients (12 %) from group A and 37 (37%) from group B, p = 0.001 (Table 3).The cause of death was non-cardiovascular in 11 patients (2 in group A and 9 in group B). Cardiovascular deaths occurred in 33 patients, 5 from group A (9%) and 28 (28%) from group B (p = 0.002). Causes of death differed between groups A and B: heart failure and sudden death were significantly more frequent in group B (20 patients vs. 1 patient, p = 0.001). Preoperative and one-year postoperative clinical and echocardiographic data from patients dying of heart failure and sudden death during follow-up are shown in Table 4.
The relative risk of cardiac death through follow-up was 2.91 in group B patients (95% confidence interval [CI] 1.1 to 7.5, p = 0.028). Furthermore, the relative risk of death due to heart failure or sudden death was 10.35 in group B patients (95% CI 1.39 to 77.26, p = 0.023).
Overall survival in groups A and B was in the best-case scenario (i.e., considering as alive the two patients with incomplete follow-up) 90 ± 4% and 75 ± 8% at 5 years, 86 ± 5% and 64 ± 5% at 10 years, and 78 ± 7% and 53 ± 6% at 15 years, respectively (p = 0.009) (Fig. 1).In the worst-case scenario, considering as death the two patients with incomplete follow-up, the difference in survival of the two patient groups was still significant (p = 0.028).
Postoperative echocardiographic outcome
At one year, group A postoperative echocardiograms showed an outstanding reduction in EDD (from 71 ± 7 mm to 53 ± 6 mm, p = 0.0001) and ESD (from 48 ± 6 mm to 38 ± 6 mm, p = 0.0001) and an increased LVEF (from 53 ± 7% to 57 ± 9%, p = 0.023). Group B patients also showed an improvement in EDD (from 75 ± 8 mm to 59 ± 12 mm, p = 0.0001), ESD (from 55 ± 10 mm to 44 ± 14 mm, p = 0.0001), and LVEF (from 42 ± 10% to 47 ± 16%, p = 0.0001). This reduction occurred in both groups in the first year postoperative study and remained stable thereafter. Echocardiographic data are shown in Table 5.
At the end of follow-up survivors from group A were in NYHA functional class I (46 patients, 92%) or II (4 patients, 8%), and patients from group B were in NYHA class I (52 patients, 72%), II (11 patients, 21%), or III (1 patient, 7%). Final EDD, ESD, and LVEF in survivors were significantly worse in group B (p = 0.021, 0.023, and 0.047, respectively).
Analysis on the basis of preoperative NYHA functional class
Data were additionally analyzed on the basis of categorizing the population in preoperative NYHA functional classes I to II (group A: n = 104) versus III to IV (group B: n = 66) irrespective of the LVEF or dimensions. The main study results then disclosed a more marked survival difference between groups than with our original classification: operative mortality was now significantly greater in group B (3% vs. 14%, p < 0.01). During follow-up, 15 patients in group A (15%) and 29 (51%) in group B died (p < 0.0001); deaths were cardiovascular in 10 (10%) from group A and in 23 (35%) from group B (p < 0.0001). With this new categorization, the relative risk of cardiac death in group B patients was 6.16. The 15-year survival was 76 ± 6% in group A and 38 ± 6% in group B (p < 0.0001); however, echocardiographic improvement was less marked in both groups: EDD at one year was 55 ± 7 mm in group A and 61 ± 15 mm in group B, ESD was 39 ± 9 mm in group A and 46 ± 16 mm in group B, and LVEF was 54 ± 11% in group A and 44 ± 18% in group B.
The current ACC/AHA/ESC guidelines on the management of AR are not based on randomized clinical trials but on observational studies that suggest that patients fulfilling these criteria (5) are very likely to develop symptoms in the near future and on studies that have shown that the postoperative course is significantly worse in patients who are operated on, either severely symptomatic or with severe degrees of LV dysfunction (5,7–9). In the present study we have analyzed the long-term postoperative outcome of a large series of consecutive patients with isolated AR who have been operated on with pre-defined surgical indications. Our surgical protocol included the operation of asymptomatic patients with either EF <50% or enlarged end-systolic dimensions (15). These indications are very similar to the subsequent ACC/AHA and ESC recommendations (10,11). The present study is not a formal validation of the guidelines because, strictly speaking, a validation should entail a different design and the application to the patients of exactly the same criteria as recommended; however, because our policy for the management of chronic AR was very close to the subsequently developed guidelines, we think that our assessment of its effectiveness adds strengths to the recommendations of the latter, particularly regarding the benefits of operation in selected asymptomatic patients.
We have considered that patients operated on “early” were those who were only moderately symptomatic or had a moderate degree of LV dysfunction at the time of surgery. All group A patients fulfilled these criteria. In contrast, patients considered to have been operated on “too late” were those with severe symptoms or with severely depressed LV function at the time of surgery. It might be wondered if group A patients include individuals who were operated on too early (i.e., before fulfilling the guidelines criteria for operation). This is not the case, because no patient was operated on that was asymptomatic and with LVEF >50% and ESD <55 mm. In fact, group A patients belong to a cohort of AR patients with an extended preoperative follow-up in our department, in whom meticulous symptoms assessment was performed in each visit. In contrast, most group B patients were sent to us already with clear reasons for surgery, either severe symptoms or severely depressed LVEF.
The results of the present series clearly reflect the long-term benefit of having surgery according to guidelines recommendations and validate the beneficial influence of guidelines adherence toward early surgery. Overall survival was statistically different between both groups. In group A, no patient died of heart failure during follow-up, and only one patient died suddenly. The remaining four patients died of causes that were unlikely to have been prevented by earlier surgery. Also, at the end of follow-up, LV dimensions and LVEF were normal in group A and statistically different from that of group B. Remarkably, in group B most cardiac deaths were either sudden or due to heart failure, and these death occurred in patients in whom no improvement of ventricular function occurred after surgery. Although a proportion of patients with poor preoperative prognostic markers had an acceptable outcome, failure of improvement shortly after surgery identified a subset of patients with particular poor prognosis, as previously described (9).
Our results could be considered obvious, because patients operated on in better conditions do better in the long term. One could then argue perhaps that all patients with severe AR should be operated on. There are several reasons to delay surgery until guideline recommendations are reached: natural history studies have clearly shown that asymptomatic patients with severe AR may remain asymptomatic and with preserved LV function for many years before reaching the proposed operative values (5,13,16–20), and our data have demonstrated that when patients are operated on as soon as these values occur, the postoperative outcome in relation to LV function is excellent.
Therefore, it seems wise to defer surgery at least until unfavorable predictors such as the ones reported in guidelines are present. That surgical indication should not await the development of symptoms is well-illustrated by the additional analysis of our population classifying the patients only on the basis of NYHA functional class regardless of LV function. When classified in this way, group B patients (NYHA functional classes III to IV) had a predictably higher operative mortality and poorer long-term survival than with our original classification, whereas group A patients, which now includes many with a markedly depressed ventricular function, had a less marked long-term recovery of hemodynamic parameters. Although this behavior might seem too obvious, it nevertheless emphasizes that our categorization of patients in “early” and “late” surgery was on the basis of sensible criteria.
Nevertheless, results in our group A patients reflect that patients operated on early cannot avoid some mortality, either perioperative or during follow-up, probably related not to LV dysfunction but to aortic disease, endocarditis, and complications of the anticoagulation, all of which are unlikely to benefit by earlier surgery. It might be wondered if newer prostheses or surgical advances could, in the future, decrease these late mortality figures. In fact, our hospital mortality rate can be considered high; however, 38% were operated on in the 1980s and, as previously mentioned, in no patient was surgery denied because of severe LV dysfunction. Although our study clearly indicates the benefits of early surgery, it has to be emphasized that the outcome in group B patients also indicates a benefit of surgery, both regarding quality of life and ventricular function, and reinforces previous opinions that surgery should be offered to all patients with severe aortic insufficiency, even with extreme degrees of dilation or LV dysfunction (21,22).
Certainly this is not a clinical trial, but a cohort study where effectiveness of having followed established recommendations for early surgery in real life is assessed. This study modality might not adjust for confounding as accurately as a clinical trial. In this particular instance, however, the number of confounding variables might not be so large as to forbid adequate adjustment in an observational study and, indeed, our study design has the advantage of including the whole population of patients with a very long follow-up, thus providing a comprehensive view of the problem. In summary, our results illustrate how, in real clinical practice the adherence to guidelines for early surgery have beneficial effect on the outcome of AR patients, stressing the importance of the routine implementation of guidelines.
The authors are indebted to Ana Martin for her help in following patients.
This work was partially supported by Red Tematica de Investigación Cooperativa del Instituto de Salud Carlos III, Ministerio de Sanidad y Consumo, Spain.
- Abbreviations and Acronyms
- American College of Cardiology/American Heart Association
- aortic regurgitation
- end-diastolic diameter
- ejection fraction
- European Society of Cardiology
- end-systolic diameter
- left ventricular/ventricle
- New York Heart Association
- Received May 31, 2005.
- Revision received August 5, 2005.
- Accepted October 17, 2005.
- American College of Cardiology Foundation
- Henry W.L.,
- Bonow R.O.,
- Borer J.S.,
- et al.
- Dujardin K.S.,
- Enriquez-Sarano M.,
- Schaff H.V.,
- Bailey K.R.,
- Seward J.B.,
- Tajik A.J.
- Borer J.S.,
- Hochreiter C.,
- Herrold E.M.,
- et al.
- Bonow R.O.,
- Lakatos E.,
- Maron B.J.,
- Epstein S.E.
- Borer J.S.,
- Bonow R.O.
- Fioretti P.,
- Roelandt J.,
- Sclavo M.,
- et al.
- Bonow R.O.,
- Rosing D.R.,
- Maron B.J.,
- et al.
- Corti R.,
- Binggeli C.,
- Turina M.,
- Jenni R.,
- Luscher T.F.,
- Turina J.
- Bonow R.O.,
- Carabello B.,
- de Leon A.C. Jr.,
- et al.
- Cheitlin M.D.
- Otto C.M.
- Klodas E.,
- Enriquez-Sarano M.,
- Tajik A.J.,
- et al.