Abstract: Background. Aortic stenosis patients with left ventricular dysfunction are at increased risk for morbidity and mortality following surgical aortic valve replacement. There are few published data regarding the outcomes of patients with severe aortic stenosis and left ventricular (LV) dysfunction undergoing transcatheter aortic valve implantation (TAVI) and possible predictors of LV recovery. Aims. To compare the baseline characteristics and outcomes between patients with normal LV function and those with LV dysfunction and to assess the predictors of LV recovery after TAVI. Methods. We enrolled 505 consecutive patients with severe aortic stenosis who underwent TAVI between November 2007 and January 2010. Patients were stratified according to LV function as follows: normal LV function (ejection fraction [EF] >50%), moderate LV dysfunction (EF 35%-50%) and severe LV dysfunction (EF ≤35%). The baseline characteristics and clinical outcomes, up to 6 months, were subsequently compared among the 3 patient subgroups. Univariable and multivariable logistic regression analyses were used to identify independent predictors of LV recovery. Results. Normal LV function was identified in 324 patients (64%) and LV dysfunction in 181 patients (36%); in those with LV dysfunction, 111 patients (22%) had moderate LV dysfunction and 70 patients (14%) had severe LV dysfunction. As compared to patients with normal LV function, those with severe LV dysfunction were more likely to be male, had higher STS and logistic EuroSCORE, more coronary artery disease/previous coronary artery bypass surgery, higher NT-proBNP levels, lower mean transaortic valve gradients, and smaller aortic valve areas. No significant difference in 30-day mortality was observed between the LV function subgroups. The 6-month mortality, however, was 2-fold higher in patients with severe LV dysfunction (27% vs 15%, respectively; P=.03). Recovery of LVEF to more than 50% was observed in 15% of patients with baseline EF ≤35%. Baseline EF was the strongest independent predictor of LV recovery after TAVI (odds ratio, 85; 95% confidence interval, 19-380; P<.001). Conclusions. Despite a similar periprocedural outcome, patients with aortic stenosis and severe LV dysfunction exhibit a significantly increased 6-month mortality after TAVI. Survivors with LV dysfunction, however, show a significant potential for LV function recovery.
J INVASIVE CARDIOL 2014;26(3):132-138
Key words: LV function, ejection fraction, TAVI, TAVR
Patients with left ventricular dysfunction undergoing surgical aortic valve replacement (SAVR) are particularly challenging.1,2 More specifically, the postoperative mortality has been reported to be high, especially in those patients without myocardial contractile reserve.3 Other studies, however, revealed that a substantial proportion of patients without contractile reserve may nonetheless benefit from valve replacement4 by prolonging survival and improving left ventricular function.
Patients undergoing transcatheter aortic valve implantation (TAVI) exhibit many comorbidities, including left ventricular dysfunction.5 A recent study by Clavel et al showed that TAVI resulted in better recovery of left ventricular (LV) function than SAVR, although there was no difference in the 30-day mortality. There is a paucity of literature about the impact of LV dysfunction on the mortality of patients undergoing TAVI. A number of observational clinical studies demonstrated a potential postoperative diastolic and systolic LV improvement after TAVI.6-8
The aims of this study were to investigate clinical differences and compare mortality rates up to 6 months between patients undergoing TAVI with normal LV function and those with moderate (ejection fraction [EF], 35%-50%) and severe (EF ≤35%) LV dysfunction. Furthermore, we sought to evaluate potential predictors for LV function recovery after TAVI.
Between June 2007 and June 2010, a total of 505 patients underwent TAVI at the German Heart Center Munich. Patients considered to be at high surgical risk for SAVR were referred for TAVI. A dedicated team of cardiac surgeons, cardiologists, and anesthesiologists discussed and reached consensus that TAVI was in the best interest of the patient. To determine the suitability for TAVI, several diagnostic modalities were performed: transesophageal echocardiography, multislice computed tomography (MSCT), and angiography. Our approach to prosthesis type and size selection, and vascular access route has been described in detail elsewhere.9-11 Briefly, patients were considered if the aortic valve annulus diameter ranged from 18-27 mm. Access route strategy is based on minimal invasiveness. Thus, we consider first the transfemoral, followed by the subclavian, transapical, and transaortic approaches.
Our research protocol was approved by the local ethics and research committee. Written informed consent was obtained from all patients. No extramural funding was used to support this work. The authors are solely responsible for the design and conduct of this study, all study analyses, the drafting and editing of the manuscript, and its final contents.
Device description and procedure. Details of the Medtronic CoreValve (Medtronic CoreValve) and Edwards Sapien (Edwards Lifesciences, Inc) devices and technical aspects of the procedure have been previously published. Procedures were performed in a hybrid surgical suite. The patient received either general anesthesia or local anesthesia (with mild sedation) depending on patient characteristics. Transfemoral vascular access was obtained either percutaneously under angiographic control and use of 2 “pre-closing” devices (10 Fr Prostar XL and 6 Fr Proglide; Abbott Vascular) or by means of surgical cutdown.
Transapical access was gained through a left anterolateral minithoracotomy. Techniques of subclavian and transaortic approach were previously described.11
Study groups and endpoints. Three groups were identified according to baseline LV function. Group 1 had normal LV function (EF >50%), group 2 had moderate LV dysfunction (EF 35%-50%) and group 3 had severe LV dysfunction (EF ≤35%). The baseline and procedural characteristics were compared among all 3 groups.
The primary endpoints were all-cause mortality at 30 days and at 6 months. Prespecified secondary endpoints included postoperative acute kidney injury (defined as need for hemodialysis or increase in postoperative creatinine level more than 1 mg/dL) and New York Heart Association functional class up to 6-month after TAVI.
Echocardiographic data. Patients underwent echocardiographic evaluation at baseline, at discharge, and at 6 months after TAVI. The echocardiographic protocol contained measurements of the aortic annulus, the interventricular septum, the aortic valve peak and mean gradient, aortic valve area, and the left ventricular ejection fraction using commercially available ultrasound systems (Siemens; Acuson Sequoia 512). LVEF was estimated according to Simpson’s rule. The simplified Bernoulli equation was used to calculate the transaortic gradients and the aortic valve area was calculated by the continuity equation. All echocardiographic examinations were performed by experienced echocardiographers.
Statistics. Continuous variables were presented as mean ± standard deviation or median and interquartile range (IQR). One-way analysis-of-variance (ANOVA) was used to compare means across groups; post hoc pairwise comparison was done with Bonferroni’s correction. Categorical variables were analyzed by the Pearson chi-square test or, when indicated, by the Fisher’s exact test.
The impact of baseline data and Doppler echocardiographic variables on 6-month survival were assessed using a Cox proportional hazards model. Those variables with a P-value <.20 were incorporated into the multivariable model. To assess the impact of transapical access on mortality, we entered this into the univariable and multivariable analysis.
Second univariable and multivariable logistic regression analyses were performed to identify the independent predictors of postprocedural LV function recovery. Clinically relevant variables with a value of P≤.05 on individual analysis were included in the multivariable model.
Kaplan-Meier estimator curves were used to examine the distribution of 6-month mortality in the three groups. Log-rank test was used to compare differences between stratified LV function groups.
Furthermore, based on previous TAVI studies, we stratified all patients according to LV ≤50% and >50%. A Kaplan-Meier survival curve was performed to analyze survival differences in the two groups before and after the adjustment for significant baseline characterisctics. A two-sided P<.05 was considered to be significant.
All data were analyzed with SPSS software version 18 for Windows (SPSS, Inc).
Baseline characteristics of the study population. Baseline characteristics are summarized in Table 1. Among the 505 patients included in the study, 70 patients (14%) had severe LV dysfunction (EF ≤35%), 111 patients (22%) had moderate LV dysfunction (EF 35%-50%), and 324 patients (64%) had normal LV function (EF >50%). Compared to patients with normal LV function, those with severe ventricular dysfunction were more likely to be male, more likely to have prior cardiac surgery, exhibit higher levels of creatinine (mg/dL) and NT-proBNP (ng/L), and had 2-fold higher logistic EuroSCORE and STS scores.
Echocardiographic data. The echocardiographic data are shown in Table 2. Patients with LV dysfunction had significantly smaller aortic valve area, lower transvalvular aortic mean gradients, and larger aortic valve annulus and left ventricular end-diastolic diameter (LVEDD) than patients with normal LV function.
Procedural data and outcomes. Periprocedural parameters are defined in Table 3. Procedural success was achieved in 488 patients (96.8%). A cardiopulmonary bypass (CPB) and reanimation were more often needed for patients with severe LV dysfunction than for those with normal LV function (4.3% and 7.1% vs 2.8% and 3.6%, respectively; P=NS). Patients with severe LV dysfunction were more likely to be implanted via transfemoral than transapical approach (78.6% vs 58%, respectively; P<.001).
Mortality. The overall mortality was 18.4% at a median follow-up of 6 months (IQR, 4-10 months). Among the three groups, 30-day all-cause mortality was similar. The 6-month mortality rate, however, was nearly 2-fold higher in patients with severe LV dysfunction than those with normal LV function (26.9% vs 15%, respectively; P=.03) (Table 4). The Kaplan-Meier survival curves stratified according to LV function are illustrated in Figure 1.
Univariable and multivariable analysis of 6-month mortality. Univariable analysis was performed to assess associations between preoperative risk factors and 6-month mortality. The results of the analyses are shown in Table 5. Univariable analyses showed that survival was adversely affected by the following variables: severe LV dysfunction (HR, 1.8; 95% CI, 1.07-3.02; P=.02), peripheral vascular disease (HR, 1.6; 95% CI, 1.02-2.5; P=.03), serum creatinine (HR, 1.4; 95% CI, 1.01-1.8; P=.04), NT-proBNP (HR, 1.1; 95% CI, 1.001-1.03; P=.03), logistic EuroSCORE (HR, 1.1; 95% CI, 1.01-1.03; P=.04), and STS score (HR, 1.1; 95% CI, 1.04-1.12; P<.001). On the multivariable analysis, only STS score was statistically significant (HR, 1.1; 95% CI, 1.02-1.15; P=.01).
Hemodynamical and clinical improvements. The clinical, hemodynamical, and EF changes for patients alive at 6 months are shown in Figures 2 and 3. From those with severe LV dysfunction, 15% developed an LVEF >50% and 66% an LVEF 35%-50% postoperatively. There is a significant attrition in the transaortic mean gradient at discharge and this remained lower after 6-month follow-up. The NYHA functional class increases significantly after the procedure in all groups (Figure 3).
Predictors of LV recovery. Predictors of LV recovery are summarized in Table 6. On the univariable analyses, male gender (P=.01), baseline EF ≤35% (P<.001), a smaller AVA (P=.01), a mean transaortic gradient ≤40 mm Hg (P=.02), and low postoperative CK-MB (P=.01) and troponinT levels (P=.01) were predictors of LV recovery after TAVI. On the multivariable analysis, however, only baseline EF ≤35% remained a strong predictor of LV recovery (P<.001).
The principal findings of this study are: (1) patients with LV dysfunction undergoing TAVI were more likely to be male, have larger ventricular dimensions and more comorbidities (including higher STS and logistic EuroSCORE); (2) TAVI can be performed with similar periprocedural outcomes in patients with severe LV dysfunction; and (3) patients with severe LV dysfunction had 2-fold higher 6-month mortality as compared to patients with normal LV function with a larger potential for LV recovery. Additionally, a baseline EF ≤35% was a strong predictor of LV recovery after TAVI.
The data reported here confirm results from other previous studies with the conclusion that patients with LV dysfunction had more comorbidities including impaired renal function, coronary artery disease, higher STS and logistic Euro-SCORE, and higher preoperative NT-proBNP level.7,8,12
Comparing the annulus measurement between patients with and without LV dysfunction, we found that patients with severe AS and LV dysfunction had significantly larger aortic annular dimensions associated with increased LV volumes and LV sphericity. A recently published study made the same conclusion and confirmed in the multivariable analysis that LV dysfunction was an independent determinant of aortic annular diameter in severe AS patients.19 These findings may influence prosthesis type and size selection in patients undergoing TAVI.
Although patients with LV dysfunction presented with more comorbidities, they had 30-day mortality rates similar to patients with normal LV function. The 6-month mortality, however, was nearly double in patients with severe LV dysfunction (EF ≤35%) than in patients with normal LV function (EF >50%) (Table 4). In our study, patients with EF ≤50% had a higher 6-month mortality rate than those with normal LV function (EF >50%). Even after adjustment for baseline characteristics, 6-month mortality was higher in those with EF ≤50% (Figure 4). In contrast to our study, Ewe et al found that 6-month and 12-month survival was similar between TAVI patients with normal LV function (EF >50%) and those with LV dysfunction (EF ≤50%).20 They found, however, that patients with LV dysfunction (EF <50%) had significantly more major adverse cardiac cerebrovascular events (MACCE) than patients with normal LV function (EF >50%) (20% vs 7%; P<.05).20 Similarly, Van der Boon demonstrated that immediate and long-term outcomes after TAVI did not differ between patients with EF ≤35% and those with EF >35%.8 Pilgrim et al did not observe differences in survival in TAVI patients stratified according to baseline LVEF.7
Comparing patients with LV dysfunction undergoing transapical (TA) versus transfemoral (TF) access, there were no significant differences in 3-month and 6-month mortality between the two groups. However, patients with LV dysfunction undergoing TA access had more cardiac depression, requiring more catecholamines or cardiopulmonary bypass support (Supplemental Table 1). Severe LV dysfunction was a univariable predictor of 6-month mortality (HR, 1.8; 95% CI, 1.07-3.2; P=.02); in the multivariable analyses, however, only a trend toward LVEF ≤35% was detected (HR, 1.8; 95% CI, 0.9-3.79; P =.09) and STS score remained the only independent significant predictor of mortality. Of note, STS score included the baseline LVEF as well. This pattern closely matched that reported in the recently published data from the United Kingdom registry, which showed severe LV dysfunction (EF ≤35%) to be a significant predictor of 1-year mortality in the univariable and multivariable analyses.21 Thus, our data demonstrate that higher mid-term mortality in patients with EF ≤35% is mainly related to other comorbidities as expressed by the STS score, but also to the impaired LV function itself.
Left ventricular recovery after TAVI. Postoperatively and 6 months after implantation, the survivors showed marked hemodynamic and clinical improvements with transvalvular gradient decrease. This study adds, hereby, to the mounting body of evidence that patients with severe AS and LV dysfunction benefit from TAVI. Similar results were obtained in previous studies.7,8,12,22-24 A study by Clavel et al showed marked LV improvement in patients following TAVI vs SAVR.22 In other studies, a postoperative diastolic function improvement was found immediately after TAVI due to the reduction in LV filling pressure.6 In addition, Ewe et al observed that an LV mass regression occurred in all patients due to the improvement in LV hemodynamics post TAVI.
The current study revealed potential predictors of LV recovery in patients with severe AS and severe LV dysfunction undergoing TAVI, ie, baseline EF ≤35%, lower aortic valve area (AVA), higher mean transaortic valve gradient pressure, lower myocardial injury, and male gender were univariable predictors of LV recovery after TAVI. In the multivariable analysis, however, only baseline EF ≤35% remained a strong independent predictor of LV recovery.
The Clavel et al study revealed that a lower baseline LVEF, female gender, and absence of larger AVA were assosciated with a marked LV improvement.22
Similar to our study, baseline EF ≤35% was a strong predictor of LV improvement. A possible explanation of the difference in gender prediction was that, in our study, 61% of all patients who presented with severe LV dysfunction were males. In a recent study by Rodés-Cabau et al, the authors made the same conclusion that a larger myoacardial injury (CK-MB increase of >26 µg/L and cardiac troponin T of >0.48 µg/L) was associated with a deterioration of the LV function following TAVI.25
Since only patients with LV dysfunction have the potential for a clinically significant improvement, an EF ≤35% turns out to be an independent predictor of LV function recovery. In addition, the fact that more than 80% of the survivors after TAVI with a baseline EF ≤35% show an improved LV function at 6 months should encourage the treatment of this patient population.
Clinical implications. The current study adds to the mounting evidence that TAVI is a safe and feasible procedure in patients with calcified severe AS and additional severe LV dysfunction deemed to be at high risk for SAVR. Despite higher 6-month mortality rate in those with LV dysfunction, the survivors showed marked hemodynamical and clinical improvements, including LV recovery. Thus, TAVI should not be denied in this patient group.
Whether the absence or presence of contractile reserve in those patients with severe LV dysfunction undergoing TAVI impacts clinical outcomes and LV recovery should be the focus of future studies.
Study limitations. The small sample size of patients limits the conclusions regarding the ability to present statistically powerful results of the different subgroup analyses presented in this study. Therefore, these results will have to be confirmed by larger studies. The current study is retrospective, so we cannot rule out possible confounding variables not included in the Cox hazard or logistic regression.
None of our patients with LV dysfunction had a dobutamine stress echocardiography to assess LV contractile reserve. However, previous studies demonstrated LV recovery after surgical aortic valve implantation regardless of existence of cardiac contractile reserve.22,26 Thus, this is unlikely to have affected the statistical significance of our results of LV recovery after TAVI.
This study highlights the safety and feasibility of TAVI regardless of baseline LV function. Because of additional comorbidities, patients with severe aortic stenosis and LV dysfunction have worse 6-month survival outcomes than patients with normal LV function. The survivors, however, showed marked LV recovery and hemodynamical and clinical improvements.
- Rahimtoola SH. Severe aortic stenosis with low systolic gradient: the good and bad news. Circulation. 2000;101(16):1892-1894.
- Monin JL, Quere JP, Monchi M, et al. Low-gradient aortic stenosis: operative risk stratification and predictors for long-term outcome: a multicenter study using dobutamine stress hemodynamics. Circulation. 2003;108(3):319-324.
- Monin JL, Monchi M, Gest V, Duval-Moulin AM, Dubois-Rande JL, Gueret P. Aortic stenosis with severe left ventricular dysfunction and low transvalvular pressure gradients: risk stratification by low-dose dobutamine echocardiography. J Am Coll Cardiol. 2001;37(8):2101-2107.
- Quere JP, Monin JL, Levy F, et al. Influence of preoperative left ventricular contractile reserve on postoperative ejection fraction in low-gradient aortic stenosis. Circulation. 2006;113(14):1738-1744.
- Leon MB, Smith CR, Mack M, et al. Transcatheter aortic-valve implantation for aortic stenosis in patients who cannot undergo surgery. N Engl J Med. 2010;363(17):1597-1607.
- Goncalves A, Marcos-Alberca P, Almeria C, et al. Acute left ventricle diastolic function improvement after transcatheter aortic valve implantation. Eur J Echocardiogr. 2011;12(10):790-797.
- Pilgrim T, Wenaweser P, Meuli F, et al. Clinical outcome of high-risk patients with severe aortic stenosis and reduced left ventricular ejection fraction undergoing medical treatment or TAVI. PLoS One. 2011;6(11):e27556.
- van der Boon RM, Nuis RJ, Van Mieghem NM, et al. Clinical outcome following transcatheter aortic valve implantation in patients with impaired left ventricular systolic function. Catheter Cardiovasc Interv. 2012;79(5):702-710. Epub 2012 Jan 10.
- Elhmidi Y, Bleiziffer S, Piazza N, et al. Incidence and predictors of acute kidney injury in patients undergoing transcatheter aortic valve implantation. Am Heart J. 2011;161(4):735-739.
- Lange R, Bleiziffer S, Mazzitelli D, et al. Improvements in transcatheter aortic valve implantation outcomes in lower surgical risk patients a glimpse into the future. J Am Coll Cardiol. 2012;59(3):280-287.
- Bleiziffer S, Mazzitelli D, Opitz A, et al. Beyond the short-term: clinical outcome and valve performance 2 years after transcatheter aortic valve implantation in 227 patients. J Thorac Cardiovasc Surg. 2012;143(2):310-317.
- Ben-Dor I, Maluenda G, Iyasu GD, et al. Comparison of outcome of higher versus lower transvalvular gradients in patients with severe aortic stenosis and low (<40%) left ventricular ejection fraction. Am J Cardiol. 2012;109(7):1031-1037. Epub 2012 Jan 17.
- Carroll JD, Carroll EP, Feldman T, et al. Sex-associated differences in left ventricular function in aortic stenosis of the elderly. Circulation. 1992;86(4):1099-1107.
- Buonanno C, Arbustini E, Rossi L, et al. Left ventricular function in men and women. Another difference between sexes. Eur Heart J. 1982;3(6):525-528.
- Milavetz DL, Hayes SN, Weston SA, Seward JB, Mullany CJ, Roger VL. Sex differences in left ventricular geometry in aortic stenosis: impact on outcome. Chest. 2000;117(4):1094-1099.
- Rohde LE, Zhi G, Aranki SF, Beckel NE, Lee RT, Reimold SC. Gender-associated differences in left ventricular geometry in patients with aortic valve disease and effect of distinct overload subsets. Am J Cardiol. 1997;80(4):475-480.
- Regitz-Zagrosek V, Brokat S, Tschope C. Role of gender in heart failure with normal left ventricular ejection fraction. Prog Cardiovasc Dis. 2007;49(4):241-251.
- Tzemos N, Therrien J, Yip J, et al. Outcomes in adults with bicuspid aortic valves. JAMA. 2008;300(11):1317-1325.
- Ng AC, Yiu KH, Ewe SH, et al. Influence of left ventricular geometry and function on aortic annular dimensions as assessed with multi-detector row computed tomography: implications for transcatheter aortic valve implantation. Eur Heart J. 2011;32(22):2806-2813. Epub 2011 Jul 23.
- Ewe SH, Ajmone Marsan N, Pepi M, et al. Impact of left ventricular systolic function on clinical and echocardiographic outcomes following transcatheter aortic valve implantation for severe aortic stenosis. Am Heart J. 2010;160(6):1113-1120.
- Moat NE, Ludman P, de Belder MA, et al. Long-term outcomes after transcatheter aortic valve implantation in high-risk patients with severe aortic stenosis: the U.K. TAVI (United Kingdom Transcatheter Aortic Valve Implantation) registry. J Am Coll Cardiol. 2011;58(20):2130-2138.
- Clavel MA, Webb JG, Rodes-Cabau J, et al. Comparison between transcatheter and surgical prosthetic valve implantation in patients with severe aortic stenosis and reduced left ventricular ejection fraction. Circulation. 2010;122(19):1928-1936.
- Van Linden A, Kempfert J, Blumenstein J, et al. Transapical aortic valve implantation off-pump in patients with impaired left ventricular function. Ann Thorac Surg. 2011;92(1):18-23.
- Gotzmann M, Lindstaedt M, Bojara W, Ewers A, Mugge A. Clinical outcome of transcatheter aortic valve implantation in patients with low-flow, low gradient aortic stenosis. Catheter Cardiovasc Interv. 201;162(2):238-245.e1. Epub 2011 Jul 18.
- Rodes-Cabau J, Gutierrez M, Bagur R, et al. Incidence, predictive factors, and prognostic value of myocardial injury following uncomplicated transcatheter aortic valve implantation. J Am Coll Cardiol. 2011;57(20):1988-1999.
- Tribouilloy C, Levy F, Rusinaru D, et al. Outcome after aortic valve replacement for low-flow/low-gradient aortic stenosis without contractile reserve on dobutamine stress echocardiography. J Am Coll Cardiol. 2009;53(20):1865-1873.
From the 1Clinic for Cardiovascular Surgery, German Heart Center, Munich, Germany, and the 2Department of Cardiology, McGill University Health Center, Montréal, Canada.
Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Lange is a consultant for Medtronic and Edwards; Dr Piazza is a consultant and proctor for Medtronic CoreValve; Dr Bleiziffer is a proctor for Medtronic. The remaining authors report no conflicts of interest regarding the content herein.
Manuscript submitted October 8, 2013, provisional acceptance given October 15, 2013, final version accepted November 11, 2013.
Address for correspondence: Yacine Elhmidi, MD, Clinic for Cardiovascular Surgery, German Heart Center, Munich, 80636, Germany. Email: email@example.com