Survival in Patients With Severe Aortic Regurgitation and Severe Left Ventricular Dysfunction Is Improved by Aortic Valve Replacement
Results From a Cohort of 166 Patients With an Ejection Fraction ≤35%
Background— Aortic valve replacement (AVR) in patients with severe aortic regurgitation (AR) and left ventricular (LV) dysfunction entails a higher surgical risk. Though it may improve symptoms and LV size, it is not known whether it translates into a survival benefit.
Methods and Results— This retrospective cohort study included patients screened from our echocardiographic database between 1993 and 2007 for patients with severe AR and LV ejection fraction (EF) ≤35%. Charts reviews were conducted for clinical, pharmacological, and surgical information. Mortality data were obtained from the social security death index and analyzed as a function of AVR adjusted for the propensity score. Of the 785 patients with severe AR, 166 patients had severe LV dysfunction defined as an EF ≤35%: 69% of these were men, age 65±16 years, and LV EF was 23±8%. Kaplan–Meier analysis revealed that performance of AVR (n=53) was associated with a better survival (P=0.001). Adjusted for the propensity score, AVR was associated with a significantly lower mortality hazard (HR 0.59, CI 0.42 to 0.98, P=0.04).
Conclusions— There is a clear reluctance to offer AVR in a large number of patients with severe AR associated with LV dysfunction. However, the performance of AVR in these patients is associated with a mortality benefit supporting the current ACC/AHA guidelines.
The ACC/AHA Practice Guidelines recommend aortic valve replacement (AVR) for severe aortic regurgitation (AR) patients presenting with left ventricular (LV) dysfunction or cardiac symptoms.1 It has been shown that AVR can be performed in patients with AR and LV dysfunction with an acceptable mortality leading to an improvement in postoperative quality of life.2–4 However, there is a major reluctance to offer AVR to these patients. The Euro Heart Survey revealed that only 22% of patients with severe AR with ejection fraction (EF) between 30% to 50% received surgical intervention; and the AVR rate was only 3% in those with an EF <30%.5 This disinclination may stem from the clear association that exists, in severe AR patients, between lower preoperative LVEF and reduced rates of postoperative survival.6–12 However, no studies have evaluated the impact of AVR on survival in severe AR patients with severe LV dysfunction.13 We hypothesized that AVR would confer a survival benefit in these patients and tested this hypothesis using a large cohort of severe AR patients with an LVEF of ≤35%.
This is a retrospective cohort study from a large university medical center. This study was approved by our institutional review board. Our echocardiographic database, which contains records from 1993 to 2007, was searched for patients with severe aortic regurgitation defined by a jet height to LV outflow tract dimension ratio of ≥0.6 or a prominent holodiastolic flow reversal in the aortic arch or the abdominal aorta.14 Patients with prosthetic aortic valve were excluded. This yielded a total of 785 patients. Of these, the 166 patients with severe LV dysfunction, as defined by an EF ≤35%, formed the study cohort. Detailed chart reviews were then performed on these patients (both alive and dead) by senior medical residents to extract clinical, pharmacological, and surgical data.
Definition of Clinical Variables
Hypertension (HTN) was defined as blood pressure greater then 140/90 mm Hg, a history of hypertension, or treatment with relevant medications. Diabetes was defined as having a history of diabetes or being treated with antidiabetic medications. Renal insufficiency was defined as serum creatinine >2 mg/dL, and coronary artery disease was defined as having a history of angina, myocardial infarction or coronary revascularization, electrocardiographic presence of Q-waves, or a history of angiographic coronary artery disease.
Pharmacotherapy at the time of echocardiography or during follow-up was recorded. This was categorized into aspirin, warfarin, βblockers, angiotensin-converting enzyme inhibitors, statins, loop diuretics, and thiazide diuretics.
All patients had standard 2-dimensional echocardiographic examinations. LV ejection fraction was assessed by a level-3–trained echocardiographer and entered into a database at the time of the examination. Anatomic and Doppler measurements were performed according to the recommendations of the American Society of Echocardiography.14–15
The end point of the study was all-cause mortality. Mortality data were obtained from the National Death Index using social security numbers as of August 9, 2007. The patients who were alive were censored as of this date. The reliability of the data were ensured by a match between the social security numbers, names, and date of birth. In the lone patient in whom there was a mismatch, the patient was censored on the date of the echocardiogram.
Analysis was performed using StatView 5.01 (SAS Institute, Inc). Characteristics of patients with and without AVR were compared using the Student t test for continuous variables and χ2 test for categorical variables. Statistical tools used for survival analysis included the Kaplan–Meier method, Cox regression model, and propensity score analysis as described later. In all cases, time zero was defined as time of the earliest echocardiogram showing severe AR and EF ≤35% in our institution. A probability value of ≤0.05 was considered statistically significant.
Baseline Patient Characteristics
The baseline patient characteristics were as follows: age 65±17 years, 68% male, EF 23±8%, coronary artery disease in 53%, diabetes mellitus in 19%, and hypertension in 67%. During follow up, 53 patients underwent AVR. Over a follow-up of 3.7±3.9 years, there were 93 deaths.
Of the 53 patients who underwent AVR, 34 had mechanical valves and 19 bioprosthetic valves. The concomitant surgical procedures included the following: 17 (32%) had coronary artery bypass surgery, 9 (17%) had mitral valve repair, 7 (13%) had mitral valve replacement, and 5 (9%) had aortic root replacements. The aortic valve size was 25.6±2.1 mm. The median time from the echocardiogram to AVR was 3 days. The time to AVR after the index echocardiogram was <30 days in 43 (81%) patients, <6 months in 50 (94%) patients, and more than a year in 3 (6%) patients. The 30 day surgical mortality was 4% in the whole cohort.
AVR and Survival
Using Kaplan–Meier analysis with log-rank statistic, survival in patients who underwent AVR was significantly better than those managed medically (Figures 1 and 2⇓). One-year, 2-year, and 5-year survival rates among patients with AVR were 88%, 82%, and 70%, respectively, compared to 65%, 50%, and 37%, respectively in those who had no AVR (P=0.001).
AVR Versus No AVR Groups
The Table summarizes characteristics of the severe AR patients with severe LV dysfunction with and without AVR. The AVR group had a greater preponderance of males (87% versus 59%, P=0.0004), a higher EF (25±8% versus 22±8%, P=0.008), a lower prevalence of diabetes (9% versus 24%, P=0.03), and a lower prevalence of renal insufficiency (4% versus 38%, P<0.0001). They also had a greater use of cardiac medications such as coumadin (69% versus 23%, P<0.0001) and beta blockers (62% versus 40%, P=0.008).
Propensity Score Analysis
In view of covariate imbalance between the treatment and control groups, propensity score analysis was performed in an attempt to adjust for these differences. Probability of receiving AVR (propensity score) for each patient was modeled by using logistic regression conditioned on the covariate values for that individual including age, gender, EF, coronary disease, diabetes, hypertension, renal insufficiency, heart failure, chronic lung disease, stroke, and atrial fibrillation. The logistic regression model created was highly predictive of performing or not performing AVR with a C-statistic value of 0.86. Effect of AVR on survival was analyzed adjusting for the propensity score using the Cox regression model. By propensity score analysis, AVR was associated with a significantly lower mortality hazard (HR 0.59, 95% CI 0.42 to 0.98, P=0.04).
Time Varying Cox Regression
As described above, though the median time interval between the echocardiogram and AVR was only 3 days, 10 patients had delays >30 days. To account for this time varying Cox regression analysis was carried by treating this lag time in the nonAVR group. With this, the impact of AVR on survival remained unchanged (adjusted for propensity score) with a HR of 0.59 (95% CI 0.40 to 0.99, P=0.05).
Mechanical Versus Bioprosthetic AVR
The 34 patients receiving mechanical AVR were significantly younger than those receiving bioprosthetic AVR (age 53±16 years versus 66±16 years, P<0.001) with no statistically significant difference in their survival curves with or without adjustment for age (P=0.52 and 0.10, respectively).
Survival Benefit of AVR in Patient With EF <20%
Figure 3 shows the Kaplan–Meier survival curves in patients with EF ≤20% (n=78). Five-year survival rate was 35% in patients who did not undergo AVR (n=57) compared with 70% for the 21 patients who underwent AVR (P=0.02).
Potential Reasons for Not Offering AVR
Because of the retrospective nature of the study, it was not possible to infer precisely the reason for not performing AVR. Mostly, the reluctance to refer rested at the level of the patient, primary care physicians, and the cardiologists. Most of the medically managed patients were not even referred for surgical therapy. Logistic regression analysis showed higher age, female gender, low EF, and renal insufficiency to be independent predictors for not offering AVR.
Though ACC/AHA Guidelines unequivocally recommend AVR for patients with severe AR and severe LV dysfunction, there is a general reluctance to offer this treatment for this patient population.1,5 In our series, only 53 of the 166 patients who met these criteria received AVR. The reasons for rejection of surgical interventions varied, including old age, the presence of medical comorbidities, and perceived increased mortality and morbidity. In our study, nonsurgical management was associated with older age, the female gender, lower ejection fraction, smaller LV systolic and diastolic diameters, diabetes, renal insufficiency, and the presence of greater degrees of mitral and tricuspid regurgitation. Barriers for AVR seemed to exist at patient level as well as at the levels of primary care physicians, cardiologists, and the surgeons.
Benefit of AVR
Our study shows that in severe AR patients with severe LV dysfunction, AVR is an independent predictor of improved survival. AVR had significant survival benefit with 1-year, 2-year, and 5-year survival rates of 88%, 82%, and 70%, respectively, compared to 65%, 50%, and 37%, respectively, in the population who did not receive AVR surgery (P=0.001). The survival benefit was supported by propensity score which removes 85% to 90% of treatment bias.16–17 In the absence of a randomized study addressing AVR in this population, our data strongly support the ACC/AHA guidelines to offer AVR in those with severe AR despite severe LV dysfunction even in those with an EF <20%.
This finding is noteworthy because of the dearth of studies directly comparing the effects of AVR and medical management on the survival of severe AR patients with severe LV dysfunction. Chaliki et al found that severe AR patients receiving AVR with EF <35% (n=43) had lower survival rates and higher congestive heart failure rates than those with EF 35% to 50% (n=134) or EF >50% (n=273).3 Whereas patients with EF <35% in our study were observed to have a 70% survival rate at 5 years, those similarly situated in their study had a 49% 5-year survival. By referencing the poor natural history of conservatively managed severe AR patients and relying on survival information based on the age- and sex-matched 1990 U.S. white population, Chaliki et al concluded that patients with severe EF depression should not be denied AVR. Though their study did not directly compare survival rates among patients administered AVR with those managed medically, this article strengthens the validity of their conclusion by providing this data.
Chukwuemeka et al concluded that AVR should not be denied to severe AR patients with EF <40% on the basis of low LV EF.4 They reported survival rates of 96%, 79%, and 55% at 1, 5, and 10 years after AVR surgery, respectively, in the context of the poor natural history of AR. They did not compare severe AR patients who received AVR surgery against those who were medically managed.
Strengths of our Study
This study is the largest one so far to address this issue and the first to compare survivals with and without AVR. Our patients also are well characterized in terms of clinical, pharmacological, echocardiographic, and surgical data. We used robust statistical tools like propensity score analysis in addition to the standard Kaplan–Meier analysis. Propensity score analysis was used to correct covariate imbalances. Modeling based on propensity scores is estimated to remove up to 90% of inherent bias of a retrospective study.16–17
As an observational study, this article is prone to the inherent biases of a retrospective study. Though propensity score analysis used to attempt to remove effect of selection bias on survival, a prospective randomized study is the only way to answer this clinical question in unequivocal terms. However, such a prospective trial seems to be unlikely in the near future.
There is a clear reluctance to offer AVR in a large number of patients with severe AR associated with LV dysfunction. However, the performance of AVR in these patients is associated with a mortality benefit supporting the current ACC/AHA guidelines and should not be denied even to patients with an EF <20%.
Presented in part at American Heart Association Scientific Sessions 2008, November 8–12, 2008, New Orleans, La.
Bonow RO, Carabello BA, Chatterjee K, de Leon AC Jr, Faxon DP, Freed MD, Gaasch WH, Lytle BW, Nishimura RA, O'Gara PT, O'Rourke RA, Otto CM, Shah PM, Shanewise JS. ACC/AHA 2006 guidelines for the management of patients with valvular heart disease: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Develop Guidelines for the Management of Patients With Valvular Heart Disease). Circulation. 2006; 114: e84–e231.
Chaliki HP, Mohty D, Avierinos J, Scott CG, Schaff HV, Tajik AJ, Enriquez-Sarano M. Outcomes after aortic valve replacement in patients with severe aortic regurgitation and markedly reduced left ventricular function. Circulation. 2002; 106: 2687–2693.
Chukwuemeka A, Rao V, Armstrong S, Ivanov J, David T. Aortic valve replacement: a safe and durable option in patients with impaired left ventricular systolic function. Eur J Cardiothorac Surg. 2006; 29: 133–138.
Iung B, Baron G, Butchart EG, Delahaye F, Gohlke-Barwolf C, Levang OW, Tornos P, Vanoverschelde JL, Vermeer F, Boersma E, Ravaud P, Vahanian A. A prospective survey of patients with valvular heart disease in Europe: the Euro Heart Survey on valvular heart disease. Eur Heart J. 2003; 24: 1231–1243.
Bonow RO, Picone AL, McIntosh CL, Jones M, Rosing DR, Maron BJ, Lakatos E, Clark RE, Epstein SE. Survival and functional results after valve replacement for aortic regurgitation from 1976 to 1983: impact of preoperative left ventricular function. Circulation. 1985; 72: 1244–1256.
Zoghbi WA, Enriquez-Sarano M, Foster E, Grayburn PA, Kraft CD, Levine RA, Nihoyannopoulos P, Otto CM, Quinones MA, Rakowski H, Stewart WJ, Waggoner A, Weissman NJ. Recommendation for evaluation of the severity of native valvular regurgitation with two-dimensional and Doppler echocardiography. J Am Soc Echocardiogr. 2003; 16: 777–802.
Rosenbaum P, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983; 70: 41–55.