Outcomes Among Patients With Heart Failure With Reduced Ejection Fraction Undergoing Transcatheter Aortic Valve Replacement: Minimally Invasive Strategy Versus Conventional Strategy

Abstract: Objectives. To investigate the effect of TAVR technique on in-hospital and 30-day outcomes in patients with aortic stenosis (AS) and reduced ejection fraction (EF). Background. Patients with AS and concomitant low EF may be at risk for adverse hemodynamic effects from general anesthesia utilized in transcatheter aortic valve replacement (TAVR) via the conventional strategy (CS). These patients may be better suited for the minimally invasive strategy (MIS), which employs conscious sedation. However, data are lacking that compare MIS to CS in patients with AS and concomitant low EF. Methods. In this retrospective study, we identified all patients with low EF (<50%) undergoing transfemoral MIS-TAVR vs CS-TAVR between March 2011 and May 2018. Our primary endpoint was defined as the composite of in-hospital mortality and major periprocedural bleeding or vascular complications. Results. Two hundred and seventy patients had EF <50%, while 154 patients had EF ≤35%. Overall, a total of 236 patients were in the MIS group and 34 were in the CS group. Baseline characteristics between the two groups were similar except for Society of Thoracic Surgeons (STS) score (MIS 8.4 ± 5.1 vs CS 11.7 ± 6.8; P<.01). There were no differences between the two groups in incidence of the primary endpoint (MIS 5.5% vs CS 8.8%; odds ratio for MIS, 0.60; 95% confidence interval, 0.16-2.23; P=.45). Conclusions. In patients with severe AS and reduced EF, MIS was not associated with adverse in-hospital or 30-day clinical outcomes compared with CS. In these patients, MIS may be a suitable alternative to CS without compromising clinical outcomes.

J INVASIVE CARDIOL 2019;31(3):73-78. Epub 2018 December 15.

Key words: anesthesia, aortic stenosis, conscious sedation, heart failure, transcatheter aortic valve replacement

Transcatheter aortic valve replacement (TAVR) is a viable treatment for patients with severe symptomatic aortic stenosis (AS) who are otherwise poor candidates for surgical aortic valve replacement (SAVR). About 25% of patients with AS also have concomitant low ejection fraction (EF) due to systolic dysfunction, and may present a significant management dilemma to clinicians.^1,2 Up to 80%-85% of TAVR procedures performed in the United States are via the conventional strategy (CS), which employs general anesthesia and mechanical ventilation.³ Patients with low EF have depressed myocardial contractility, and may be at increased risk of adverse effects from hemodynamic shifts conferred by the use of general anesthesia and mechanical ventilation. The minimally invasive strategy (MIS) for TAVR is conducted in a standard cardiac catheterization laboratory with conscious sedation followed by transthoracic echocardiography (TTE) to confirm valve positioning following its deployment.^3,4 As MIS eliminates the use of general anesthesia and mechanical ventilation, it may be better suited for patients with low EF. However, to our knowledge, there are no studies comparing the choice of TAVR technique in patients with low EF. Using our institutional TAVR registry, we sought to investigate this important clinical issue further. 

This was a retrospective observational study of patients undergoing TAVR at our institution. In the beginning of our TAVR experience, we performed TAVRs via the CS, whereby the procedure was performed in a hybrid operating room under general anesthesia with transesophageal echocardiography (TEE) guidance; this technique is described in detail elsewhere.^5,6 In December 2013, we adopted the MIS,  whereby the procedure was performed in a standard cardiac catheterization laboratory with conscious sedation followed by TTE to confirm valve positioning following its deployment. At our institution, in patients with low EF, we use aortic valve calcium score instead of dobutamine stress echocardiography to distinguish severe from pseudo-severe AS in patients with low-flow, low-gradient AS. This method has been shown to be useful to identify true severe AS.⁷ We have described our TAVR program/registry, as well as the procedural details of CS and MIS, previously.⁸ 

As part of our institutional TAVR registry, patient information including demographic information (age, race, and sex), comorbidity information, laboratory data, and echocardiographic data were obtained from the electronic health record and entered into an encrypted database.⁹ Data on intraprocedural characteristics, hospital stay, complications, and follow-up exams were also collected. For the purposes of this study, vascular and bleeding complications were defined as recommended by the Valve Academic Research Consortium (VARC)-2.¹⁰ 

Patient selection. We examined all patients with low EF (<50%, as defined by a prior publication)¹¹ who underwent transfemoral MIS-TAVR or CS-TAVR at our institution between March 2011 and May 2018. Patients were stratified by technique (MIS or CS). We also performed a subgroup analysis among patients with very low EF (<35%). 

Outcomes. Acknowledging that the greatest effect of technique selection would likely be on intraprocedural and in-hospital outcomes, we defined our primary endpoint as a composite of in-hospital mortality and major periprocedural bleeding or vascular complications. A composite endpoint was chosen because rates of in-hospital mortality and major bleeding/vascular complications were low. Secondary outcomes included incidence of minor bleeding or vascular complications, acute kidney injury, stroke, need for permanent pacemaker (PPM) placement, intensive care unit (ICU) and total postprocedural length of stay, 30-day mortality, and 30-day readmissions (all cause and heart failure related). 

Statistical analysis. Categorical variables are displayed as frequencies and percentages, and continuous variables as mean ± standard deviation. The student’s t-test was used to compare continuous variables, and the Chi-square test or Fisher’s exact test (as appropriate) for categorical variables. All statistical analyses were two-sided, and a P-value of <.05 was considered significant. 

Univariate logistic or linear regression models were generated to study the effect of TAVR technique on outcomes. Multivariate logistic or linear models adjusting for covariates were generated if the results of the univariate analyses were statistically significant (P<.05). Covariates incorporated in the multivariate models included demographics (age, sex, race), body mass index (BMI), New York Heart Association (NYHA) class, Society of Thoracic Surgeons (STS) score, preprocedural echocardiographic and laboratory values, and the comorbidities listed in Table 1. Unadjusted and adjusted odds ratios (ORs) and parameter estimates are presented with 95% confidence intervals (CIs). SPSS version 25 (IBM) was used for all analyses. 

Of the 1029 patients who underwent TAVR during the study period, a total of 270 patients (26.2%) had EF <50%. Two hundred and thirty-six patients underwent transfemoral MIS-TAVR, while 34 patients underwent CS-TAVR. Patients in both groups were overall older, and predominantly white men (Table 1). Baseline characteristics between the two groups were similar except for prevalence of coronary artery disease (MIS 58.5% vs CS 79.4%; P=.03), prior coronary artery bypass grafting (MIS 33.1% vs CS 61.8%; P<.01), prior myocardial infarction (MIS 28.8% vs CS 55.9%; P<.01), and STS score (MIS 8.4 ± 5.1 vs CS 11.7 ± 6.8; P<.01). Of the 270 patients in our group, a total of 154 patients (57%) had very-low EF (≤35%), with 133 undergoing MIS and 21 undergoing CS. Baseline characteristics among the two groups were similar, except prior myocardial infarction (MIS 27.1% vs CS 57.1%; P=.01) (Table 1).

Outcomes in overall population. In the overall population, the two groups did not differ in the primary endpoint (MIS 5.5% vs CS 8.8%; OR for MIS, 0.60; 95% CI, 0.16-2.23; P=.45). There was no difference between the two groups in rates of stroke (MIS 0.4% vs CS 2.9%; OR for MIS, 0.14; 95% CI, 0.01-2.30; P=.17), minor bleeding (MIS 10.2% vs CS 5.9%; OR for MIS, 1.81; 95% CI, 0.43-8.03; P=.43), minor vascular complications (MIS 7.6% vs CS 14.7%; OR for MIS, 0.50; 95% CI, 0.17-1.39; P=.18), acute kidney injury (MIS 9.3% vs CS 20.6%; OR for MIS, 0.40; 95% CI, 0.16-1.02; P=.05), or need for PPM (MIS 10.6% vs CS 2.9%; OR for MIS, 4.06; 95% CI, 0.53-31.01; P=.18). 

Patients in the MIS group had both shorter median ICU length of stay (MIS 31.6 hours [IQR, 24.6-52.5 hours] vs CS 85.2 hours [IQR, 52.5-123.6 hours]; P<.001) and total median hospital stay (MIS 66.5 hours [IQR, 59.0-135.8 hours vs CS 115.5 hours [IQR, 88.7-209.6 hours]; P<.001) (Table 2). In adjusted analysis, MIS was not associated with early discharge home the day after TAVR (MIS 20.8% vs CS 2.9%; adjusted OR for MIS, 5.83; 95% CI, 0.60-57.19; P=.13) (Table 2). There was no difference in 30-day all-cause mortality (MIS 4.2% vs CS 5.9%; OR for MIS, 0.71; 95% CI, 0.15-3.38; P=.67), all-cause readmissions (MIS 5.9% vs CS 14.7%; OR for MIS, 0.37; 95% CI, 0.12-1.09; P=.07), or heart failure readmissions (MIS 3.8% vs CS 8.8%; OR for MIS, 0.41; 95% CI, 0.12-1.60; P=.20). 

Very-low EF group. In the very-low EF group (EF ≤35%), patients in the MIS group had similar outcomes to CS patients, including the primary outcome (MIS 6.0% vs CS 9.5%; OR for MIS, 0.61; 95% CI, 0.12-3.08; P=.55) and secondary outcomes (Table 2). However, in adjusted analysis, the MIS group did not have shorter ICU length of stay (adjusted parameter estimate for MIS, -29.2; 95% CI, -60.1 to 1.80; P=.07) or total length of stay (adjusted parameter estimate for MIS, -22.09 (95% CI, -75.72 to 31.51; P=.42).

In this study of patients with severe symptomatic AS and concomitant low EF, we showed that MIS-TAVR had similar clinical outcomes compared with CS-TAVR, along with a reduction in hospital length of stay. Up to 25% of patients with AS have concomitant left ventricular (LV) systolic dysfunction,^1,2 and this presents a frequent dilemma to clinicians considering AVR therapies for such patients. A secondary analysis of the PARTNER trial found no statistically significant difference in major clinical outcomes including 1-year mortality between patients with low EF (<50%), and normal EF,¹¹ while other work has reported worse outcomes from TAVR in patients with reduced EF.¹²

Patients with already depressed myocardial function may not have the ability to tolerate or compensate appropriately for the hemodynamic shifts caused by intubation, mechanical ventilation, and the use of general anesthesia. Positive pressure from mechanical ventilation reduces preload (by way of increased intrathoracic pressure), which may prompt the need for periprocedural fluid/inotropic support.^13-15 Moreover, commonly used anesthetics in general anesthesia have vasodilatory effects and further depress myocardial function, thus also increasing the likelihood of needing vasoactive support during the procedure.¹⁶ MIS-TAVR does not utilize mechanical ventilation or general anesthesia, thus reducing invasiveness, as well as the risk of hemodynamic insult in these at-risk patients. 

We found the two techniques to be similar to each other in terms of mortality and complications. However, patients in the MIS group were able to leave the ICU and the hospital earlier than the CS group. We have also shown this in our prior work;⁸ this may be due to the fact that the patients receiving general anesthesia and mechanical ventilation required in CS may need more time to recover prior to safe discharge. We have also seen significant cost savings due to reduced resource utilization — approximately $16,000 per procedure performed via the MIS,⁸ with other centers reporting similar results.¹⁷ 

Patients with very-low EF. In our cohort, patients with very low EF (≤35%) had similar outcomes compared with the overall group up to 30 days post procedure, except shorter length of stay. Patients with severely depressed myocardial function may be more susceptible to hemodynamic insult from general anesthesia than those with moderately reduced EF (35%-50%). Notably, this group formed the majority (57%) of our overall study population, and thus represented a significant proportion of patients at risk. Our analysis shows that the MIS approach may be as safe as CS-TAVR in this vulnerable population.

It must be noted that the MIS may not be suitable for all patients. There is always a risk of conversion to general anesthesia due to various factors, such as poor patient tolerance, procedural complications, and operator comfort. Our conversion rates have been low (~3.4%),⁸ while rates reported in the literature range from 3.4%-20.0%.¹³ Moreover, the MIS may not suitable for patients with pre-existing pulmonary conditions, such as severe chronic obstructive pulmonary disease, obstructive sleep apnea, or those at risk for aspiration, who may need a protected airway.

Study limitations. We acknowledge several limitations to our study. First, this was a non-randomized, retrospective, observational study with potential for selection bias between the two groups. Perhaps the largest bias may arise from the fact that the CS group was sicker than the MIS group, as reflected in the STS scores (CS 11.7 ± 6.8 vs MIS 8.4 ± 5.1; P<.01). Ideally, we would do propensity-score matching to reduce these biases, but the small number of patients in the CS group would have meant a significant reduction in the total sample size after matching (~30 in both groups). The two groups are only different in a few aspects, which may attenuate some of this bias. Second, this is a single-center study from a high-volume institution with considerable operator expertise in MIS, and thus this may preclude generalizability of our results to other centers. Other centers without our expertise and experience in MIS may not feel comfortable with this technique, and may instead opt for CS. Third, all CS cases in this study were conducted during the early learning-curve period of our experience, and operators conducting MIS were therefore more experienced, thus possibly leading to bias in our results as well. Lastly, in our analysis of 30-day outcomes, we found that our ORs were skewed, with wide CIs, which was likely due to the low event rates at 30 days. 

Our study demonstrates that in patients with severe AS and reduced EF, the use of MIS-TAVR is not associated with adverse in-hospital or 30-day clinical outcomes compared with CS-TAVR. Use of MIS reduces risk of adverse hemodynamic effects, and may be a suitable alternative to CS without compromising clinical outcomes.

From the ¹Department of Internal Medicine, University Hospitals Cleveland Medical Center, Cleveland, Ohio; ²The Valve & Structural Heart Disease Intervention Center, Harrington Heart and Vascular Institute, University Hospitals Cleveland Medical Center, Cleveland, Ohio; and ³School of Medicine, Case Western Reserve University, Cleveland, Ohio.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Li reports personal fees from Medtronic and Abbott. Dr Attizzani is a consultant and proctor for Edwards Lifesciences and Medtronic; consultant for St. Jude Medical, Inc. and Abbott Vascular. Dr Kalra is a consultant for Medtronic and Philips. The remaining authors report no conflicts of interest regarding the content herein.

Manuscript submitted September 17, 2018, final version accepted October 2, 2018.

Address for correspondence: Ankur Kalra, MD, FACP, FACC, FSCAI, University Hospitals Cleveland Medical Center, 11100 Euclid Ave, Mailstop LKS 5038, Cleveland, OH 44106. Email: kalramd.ankur@gmail.com