Clinical Outcome of High-Risk Patients with Severe Aortic Stenosis and Reduced Left Ventricular Ejection Fraction Undergoing Medical Treatment or TAVI

Introduction Reduced left ventricular function in patients with severe symptomatic valvular aortic stenosis is associated with impaired clinical outcome in patients undergoing surgical aortic valve replacement (SAVR). Transcatheter Aortic Valve Implantation (TAVI) has been shown non-inferior to SAVR in high-risk patients with respect to mortality and may result in faster left ventricular recovery. Methods We investigated clinical outcomes of high-risk patients with severe aortic stenosis undergoing medical treatment (n = 71) or TAVI (n = 256) stratified by left ventricular ejection fraction (LVEF) in a prospective single center registry. Results Twenty-five patients (35%) among the medical cohort were found to have an LVEF≤30% (mean 26.7±4.1%) and 37 patients (14%) among the TAVI patients (mean 25.2±4.4%). Estimated peri-interventional risk as assessed by logistic EuroSCORE was significantly higher in patients with severely impaired LVEF as compared to patients with LVEF>30% (medical/TAVI 38.5±13.8%/40.6±16.4% versus medical/TAVI 22.5±10.8%/22.1±12.8%, p <0.001). In patients undergoing TAVI, there was no significant difference in the combined endpoint of death, myocardial infarction, major stroke, life-threatening bleeding, major access-site complications, valvular re-intervention, or renal failure at 30 days between the two groups (21.0% versus 27.0%, p = 0.40). After TAVI, patients with LVEF≤30% experienced a rapid improvement in LVEF (from 25±4% to 34±10% at discharge, p = 0.002) associated with improved NYHA functional class at 30 days (decrease ≥1 NYHA class in 95%). During long-term follow-up no difference in survival was observed in patients undergoing TAVI irrespective of baseline LVEF (p = 0.29), whereas there was a significantly higher mortality in medically treated patients with severely reduced LVEF (log rank p = 0.001). Conclusion TAVI in patients with severely reduced left ventricular function may be performed safely and is associated with rapid recovery of systolic left ventricular function and heart failure symptoms.


Introduction
Reduced left ventricular ejection fraction (LVEF) among patients with severe aortic stenosis importantly impacts prognosis in patients treated conservatively, and increases peri-operative risk in patients undergoing surgical aortic valve replacement (SAVR) [1][2][3][4]. Although mechanical relief of aortic outflow obstruction as accomplished by SAVR has been shown to compensate for the increased peri-operative risk in patients with decreased LVEF [5], severely impaired LVEF remains one of the principal reasons to defer SAVR [6]. Moreover, recovery of LVEF in response to SAVR remains variable and difficult to predict. Transcatheter Aortic Valve Implantation (TAVI) is a less invasive procedure, which is predominantly performed among patients previously managed by medical treatment. In patients not considered suitable candidates for SAVR, TAVI has been shown to reduce mortality and rehospitalization compared with a conservative strategy [7]. In addition, it has been suggested that TAVI is associated with favorable effects on LVEF recovery [8]. The safety and efficacy of TAVI in patients with reduced LVEF (#30%) vis-à-vis a conservative strategy has not been resolved. Therefore, we investigated clinical outcomes of high-risk patients with severe aortic stenosis undergoing medical treatment or TAVI stratified by LVEF in a prospective single-center registry.

Patient Population
High-risk patients with symptomatic, severe aortic stenosis deemed at increased surgical risk have been consecutively included in a prospective single center registry initiated in July 2007. Inclusion criteria involved (1) symptomatic, severe aortic stenosis with an echocardiographic mean gradient .40 mmHg or a calculated aortic valve area ,1 cm 2 ; (2) age $80 years in the presence of a logistic EuroSCORE .15%. Patients ,80 years of age were eligible if at least one of the following comorbid conditions was present: previous cardiac surgery, chronic obstructive pulmonary disease (forced expiratory volume during one second ,1.0), severe pulmonary hypertension ($60 mmHg), porcelain aorta, history of radiation therapy to the mediastinum, or frailty (BMI ,18 kg/m 2 ). Patients with severe aortic regurgitation due to degenerated aortic valve prosthesis were excluded.

Ethics Statement
The study complies with the Declaration of Helsinki and was approved by the Ethics Committee of the University of Bern, Switzerland (www.kek-bern.ch). All subjects gave informed written consent.

Evaluation and Treatment
After a comprehensive evaluation according to a standardized protocol including left and right heart catheterization, aortography, transthoracic (TTE) and transesophageal echocardiography (TEE), and CT angiography of the chest, abdomen and pelvis an interdisciplinary team of interventional cardiologists and cardiac surgeons reviewed the cases and formed a consensus on treatment allocation to medical treatment (1), SAVR (2) or TAVI (3) based on risk assessment, anatomical considerations and patient preference [9]. For the purpose of this analysis, we focused on patients allocated to medical treatment or TAVI. Allocation to medical treatment resulted from patient's refusal to undergo either SAVR or TAVI despite the recommendation for an intervention put forward by the heart team, comorbidities with poor prognosis, anatomical or technical reasons not allowing for a transcatheter approach in patients refusing to undergo SAVR, and exceedingly high estimated risk for intervention. Reasons for TAVI included refusal of SAVR, advanced age (.80 years) in the setting of a high surgical risk, or severe comorbidities. Patients with cross-over from medical treatment or TAVI to SAVR or from medical treatment to TAVI during the time of follow-up were excluded. Medical treatment encompassed percutaneous coronary intervention in case of significant coronary artery disease with limiting angina, as well as optimal medical treatment for comorbidities such as congestive heart failure, atrial fibrillation, and hypertension, and was not necessarily associated with an equally conservative strategy for treatment of non-cardiac disease manifestations. Conservative treatment did however not include balloon aortic valvuloplasty. Also, it was not used as a bridge to TAVI.
Latter was performed through a tranfemoral, transapical, or transsubclavian approach according to anatomical characteristics using either the Medtronic CoreValve Revalving system or the Edwards Sapien valve as previously described [10].

Data collection
Patients treated medically were included into this registry at the time of the in-hospital evaluation for a potential intervention, whereas the date of the intervention was considered the time of inclusion among patients undergoing TAVI. Follow-up was performed regularly at 1, 6, and 12 months during a clinic visit or by means of a standardized telephone interview. Furthermore, all patients were contacted within two months of data freezing for the purpose of the present analysis. Hospital records and municipal civil registries were consulted to ascertain vital status. Medical records, discharge letters, and documentations of hospitalizations were systematically collected and all suspected events were adjudicated by an unblinded clinical event committee consisting of cardiac surgeons and interventional cardiologists.

Definitions
Assessment of left ventricular ejection fraction at baseline was based on measurements from TTE using planimetry. All endpoint definitions were in accordance with the criteria suggested by the Valve Academic Research Consortium [11]. The definition of cardiovascular death involved any death due to a proximate cardiac cause or a death of unknown cause, as well as all procedure-related deaths and death caused by non-coronary vascular conditions such as cerebrovascular disease, pulmonary embolism, or other vascular disease. Peri-procedural myocardial infarction was determined as new ischemic symptoms or signs in the presence of elevated cardiac biomarkers (two or more postprocedure samples that were .6-8 hours apart with a 20% increase in the second sample and a peak value exceeding 10x the 99 th percentile upper reference limit (URL), or a peak value exceeding 5x the 99 th percentile URL with new pathological Q waves in at least two contiguous leads) within 72 hours after the index procedure. Major stroke encompassed a rapid onset of focal or global neurological deficit of $24 hours duration necessitating therapeutic intervention, or documentation of a new intracranial defect using MRI or CT-scan. The modified RIFLE (Risk, Injury, Failure, Loss, End-stage kidney disease) classification was used for the definition of kidney injury which was based upon changes in serum creatinine up to 72 hours after the procedure. Stage 1 was determined as an increase of serum creatinine to 150-200% (or an increase of $26.4 mmol/l), stage 2 required an increase of the baseline creatinine to 200-300%, and stage 3 was considered in case of an increase in creatinine of $300% with an acute increase of at least 44 mmol/l.

Statistical Analysis
All analyses were performed using SPSS Statistics Version 17.0. Continuous variables are presented as mean 6 standard deviation (SD) and were compared by means of a two-sided students T-test. Categorical data are expressed as frequency (percentages), and were compared using the chi-square and Fishers exact tests. Survival was estimated using the Kaplan Meier method. We performed an univariate analysis and included in addition to age and gender all variables with a p,0.1 into a cox multivariate regression model to adjust for potential confounders.

Baseline Characteristics
Among 452 patients with severe aortic stenosis at increased surgical risk, 10 patients died before treatment allocation, and 107 patients underwent SAVR leaving 335 patients allocated to medical treatment or TAVI as basis of the present study. After exclusion of two patients with cross-over from medical treatment to SAVR, one patient with cross-over from TAVI to SAVR, and five patients with cross-over from medical treatment to TAVI, 71 patients treated medically and 256 patients treated by TAVI remained for the purpose of the present analysis. The patient population was divided into four groups according to treatment strategy (medical treatment versus TAVI) and left ventricular ejection fraction (LVEF#30% versus LVEF.30%). Severely impaired LVEF (#30%) was observed in 25 (35%) patients treated medically (M30-), whereas 46 (65%) patients treated medically had normal or only moderately reduced LVEF (M30+). A total of 37 (14%) TAVI patients had severely diminished LVEF (T30-), whereas 219 (86%) TAVI patients showed normal or moderately reduced LVEF (T30+) (Figure 1). Baseline characteristics of the overall patient population undergoing medical treatment or TAVI stratified by ventricular function are summarized in Tables 1 and 2

Short-term Clinical Outcomes
Peri-procedural characteristics and short-term clinical outcome of patients undergoing TAVI are summarized in Tables 3 and 4. All-cause mortality at 30 days amounted to 6.8% and 5.4% among patients of the T30+ and T30-group (p = 1.0), respectively. There were no differences between patients of group T30+ and T30with regard to peri-procedural myocardial infarction (0% versus 2.7%, p = 0.15), major stroke (3.7% versus 5.4%, p = 0.64), access related complications, bleeding and renal failure. Two patients of the group T30-required valvular reinterventions, whereas no valvular reinterventions were performed among patients of the group T30+ (5.4% for T30-versus 0% for T30+, p = 0.02). Both patients underwent post-dilatation 13 days and 14 days after TAVI, respectively, due to severe aortic regurgitation. There was no significant difference in the incidence of the Valve Academic Research Consortium (VARC) combined safety end point [11] between the two groups (21.0% for T30+ versus 27.0% for T30-, p = 0.40) (Figure 2). Aortic regurgitation .grade 2+ as assessed by transthoracic echocardiography before discharge was found in Exercise intolerance as assessed by NYHA functional class was documented at 30 days and is illustrated in Figure 3. Whereas patients under medical treatment reported an increase in shortness of breath at 30 days (increase $1 NYHA class in 30.8% of M30and 19.3% of M30+), TAVI patients consistently noted an improvement in exercise tolerance that seemed to be particularly pronounced among patients of the T30-group (decrease $1 NYHA class in 95% of T30-and 77.3% of T30+).

Long-term Follow-Up
Crude and adjusted long-term survival of patients undergoing TAVI and medical treatment as a function of LVEF is shown in Figure 4. Patients undergoing TAVI experienced a favorable longterm course compared with patients under medical treatment.  recovery of LVEF from baseline to discharge (from 2564% to 34610%, p = 0.002) that continued to increase during mid-term follow-up (41613%) ( Figure 5).

Discussion
The main findings of the study are as follows.
1. Severely diminished left ventricular function among patients with severe aortic stenosis treated medically has an important impact on clinical outcome with dismal prognosis.
2. TAVI in patients with severely impaired left ventricular function may be performed safely and may not be associated with an increased peri-procedural risk.
3. Patients with severely diminished LVEF undergoing TAVI demonstrate a rapid improvement in LV function and may have a similar prognosis as compared to patients with normal or moderately reduced LVEF.
4. Among patients with severely diminished LVEF, TAVI disproportionally improves NYHA functional class as com-  pared with patients with normal or moderately reduced LVEF.
The observational data of the present single center experience demonstrate the favorable impact of TAVI on recovery of LVEF and long-term clinical outcome as compared with medical treatment. Moreover, the risk of peri-procedural complications in patients with severely impaired LVEF appears to be comparable to patients with normal or mildly reduced LVEF.
There are several limitations to be considered. First, the observational, non-randomized nature of the present analysis is susceptible to a selection bias attributable to the primary allocation of the selected treatment strategy. The decision whether to perform TAVI or medical treatment in patients with severely impaired LVEF was driven by the decision of the interdisciplinary heart team, by the feasibility based on anatomic and technical features, and by the final decision of the patient. Second, the threshold values for categorization of the overall patient cohort according to LVEF, although arbitrary, were based on previous clinical studies of patients undergoing SAVR indicating an association of impaired LVEF with adverse clinical outcome [1][2][3][4][5]. At last, we might have been unaware of unknown confounding factors influencing clinical outcome.
Our data highlight the dismal prognosis of patients with severe, symptomatic aortic stenosis undergoing medical treatment. The overall estimated rate of mortality after one year amounted to 55% and is in line with the results of the medical group of the PARTNER B cohort (Placement of AoRtic TraNscathetER Valve) trial [7]. In our series, stratification according to LVEF demonstrates comparable baseline characteristics but an excessive risk of mortality among patients with severely reduced LVEF.
The assessment of peri-procedural complications in the TAVI cohort using the VARC criteria revealed no differences between the two groups stratified according to LVEF. In particular, we observed no differences with regard to all-cause mortality, incidence of peri-procedural myocardial infarction or major stroke. The most frequently encountered periprocedural complications were bleeding events and vascular complications, which occurred with similar frequency in both groups. A higher rate of repeat valvular interventions among patients of the T30-group may be due to chance but needs further investigation. The overall incidence of adverse events was comparable with previous reports of patients undergoing TAVI [12][13][14] and suggests that TAVI in patients with severely reduced left ventricular function is not associated with an increased perioperative risk. This finding contrasts with data from the surgical literature showing an increased risk of adverse events of patients with reduced LVEF [1][2][3][4][5]. For instance, Sharony et al reported a 30-day mortality of 9.6% among patients with LVEF#40% in a series of 260 patients [3]. Several factors might explain the negligible role of a diminished LVEF during the peri-procedural phase of TAVI: the strategy of a pure percutaneous approach using local anesthesia and mild conscious sedation reduces the risk of unfavorable hemodynamics during the intervention and the need of vasoactive drugs. Positioning of the stiff wire in the left ventricle and the deployment of the bioprosthesis is considered to be easier in patients with an enlarged ventricle with low output  compared with patients with a small, hypertrophic and hypercontractile ventricle. Furthermore, TAVI provides the possibility of valve implantation without cardiac arrest and its sequelae like the need for prolonged ventilation, the risk of renal failure, infection and neurologic complications [15].
Our data show, that the combination of severe aortic stenosis and severely reduced left ventricular function is associated with a dismal prognosis if treated conservatively. In our cohort, more than half of the patients died within six months of evaluation for potential intervention and almost 80% died within one year. The present findings therefore suggest that severely impaired left ventricular function should not serve as a reason to deny transcatheter aortic valve implantation. The favorable periprocedural outcome was accompanied by a rapid recovery in LVEF already during the in-hospital phase and eventually translated into favorable long-term survival comparable to patients with normal or moderately reduced LVEF. Of note, the mean aortic transvalvular gradient in patients with LVEF#30 amounted to 35616 mmHg indicating that low-flow, low-gradient aortic stenosis was encountered relatively infrequently and a majority of patients may have maintained some contractile function. Since we did not routinely perform dobutamine stress echocardiography the issue whether contractile reserve plays an important role with respect to prognosis remains unanswered. Nevertheless, patients with LVEF#30% assigned to medical treatment exhibited similar transvalvular gradients (35621 mmHg) and were found to have a considerably higher mortality rate as compared to medically treated patients with LVEF.30%.
Among patients with severely impaired LVEF symptom status as assessed by NYHA functional class improved disproportionally along with a rapid recovery in systolic left ventricular function. These findings suggest that patients with severely reduced LVEF may substantially benefit from TAVI.