RESULTS
Our study population included 272 patients: 140 (51%) were men, and 132 (49%) were female. The average patient age was 85.6 ± 7.8 years. Most of the patients were Caucasian: 259 (95%) Caucasian vs. 8 (3%) Black vs. 3 (1%) American Indian vs. 2 (1%) Other. The average patient BMI was 28.2 ± 6. The majority of patients received the newer generation SAPIEN 3 valve: 14% SAPIEN vs 14% SAPIEN XT vs 56% SAPIEN 3 vs 5% CoreValve vs 8% CoreValve Evolut vs 3% Evolut Pro. A majority of the patients had hypertension, hyperlipidemia, and a smoking history: 246 (90%) HTN vs. 203 (75%) HLD vs. 145 (53%) smokers. 95 (35%) patients had diabetes mellitus type II. 59% of this sample did not have coronary revascularization procedures before TAVR. Average left ventricular ejection fraction, aortic valve area, mean gradient, and peak gradient before TAVR were as follows: 55 ± 13.3, 0.68 ± 0.21, 42.1 ± 16.7, and 67.3 ± 26.8. 70% of the sample had some degree of aortic insufficiency before TAVR. (Table 1).
Following TAVR, ejection fraction (p<0.001) as well as aortic valve effective orifice area (p<0.001) increased while both mean transvalvular (p<0.001) and peak transvalvular gradients (p<0.001) decreased. The ejection fraction increased by 3.5 percent, and the aortic valve effective orifice area increased by 1.25 cm squared. The mean transvalvular gradient decreased by 33.6 mmHg, and the peak transvalvular gradient decreased by 49.7 mmHg. (Table 2).
In receiver-operator curve (ROC) analysis, we determined the ELI cutoff value to be 1.34 (p<0.001), using all-cause mortality as the primary outcome. (Figure 1). Multivariate cox regression showed an increased cumulative survival time for nine years post-TAVR in patients with ELI > 1.34 (p=0.002). (Figure 2). All patients were assigned to one of two groups based on their calculated ELI.
Patient characteristics are shown in Table 3. One hundred and seventy-three patients had ELI > 1.34, and ninety-nine patients had ELI ≤ 1.34. There was no significant difference between both ELI groups in the following demographic and comorbid conditions: gender, race, age, BMI, hypertension, diabetes mellitus, hyperlipidemia, or smoking status. Additionally, there was no significant difference in prior coronary revascularization, aortic valve area, mean transvalvular gradient, peak transvalvular gradient, and left ventricular ejection fraction before TAVR between both ELI groups. 86 (61%) men compared to 87 (66%) women had elevated ELI > 1.34. Patients with ELI > 1.34 were 82.6 ± 8.2 years, while those with ELI ≤ 1.34 were 82.1 ± 6.5 years (p=0.634). BMI was 27.7 ± 6.1 in patients with ELI > 1.34 compared to 28.9 ± 5.7 for patients with low ELI (p = 0.055). Patients with ELI > 1.34 had ejection fraction of 55.8 ± 14.2 while patients with ELI ≤ 1.34 had ejection fraction of 53.5 ± 13.1 (p=0.09). SAPIEN valves were most commonly associated with ELI ≤ 1.34 (p=0.014): 62% of SAPIEN vs 35% of SAPIEN XT vs 29% of SAPIEN 3 vs 46% of CoreValve vs 35% CoreValve Evolut vs 44% CoreValve Evolut Pro (Table 3).
All-cause mortality was higher in patients with lower ELI (95% confidence interval, 5.59 (3.25 - 9.62); p< 0.001). Patients with ELI ≤ 1.34 had a fivefold increased risk of death following TAVR. 71 (71.7%) patients with ELI ≤ 1.34 compared to 53 (31.2%) patients with ELI > 1.34 were deceased within nine years following TAVR. There was no significant difference between ELI groups in the following clinical outcomes following TAVR: cerebrovascular events, congestive heart failure, prosthetic thrombosis, bacterial endocarditis, prosthetic regurgitation, myocardial infarction, new onset dysrhythmias, cardiac arrest requiring CPR, post-TAVR coronary intervention, TAVR revision or hospital readmission. (Table 4).
In total, one hundred and twenty-five patients died within nine years post-TAVR. Univariate analysis was performed on demographic and patient-related factors that may contribute to death following TAVR.  Hypertension was associated with increased mortality in this univariate analysis (95% confidence interval, 3.12 (1.21 - 8.03); p=0.021). While 48.4% of all patients had hypertension, 95.2% of deceased patients had hypertension. Each valve type was associated with different mortality rates within 9 years post-TAVR (p=0.027): 64.8% SAPIEN vs 53.3% CoreValve vs 51.7% SAPIEN XT vs 37.9% SAPIEN 3. Although not statistically significant, age and aortic valve area were other factors that were borderline. The age of the deceased patients was 83.3 ± 6.8 (95% confidence interval, 1.02 (0.99 - 1.06); p =0.07). The aortic valve area of the deceased patients was 0.70 ± 0.19 (95% confidence interval, 1.96 (0.88 - 4.34); p=0.098). The following factors were not associated with mortality post-TAVR: gender, race, BMI, diabetes mellitus, hyperlipidemia, smoking status, prior coronary revascularization, moderate/severe aortic insufficiency, mean transaortic gradient or left ventricular ejection fraction. (Table 5).
Multivariate Cox regression was performed on select variables with p values < 0.15. Increased age (hazard ratio, 1.037; 95% confidence interval 1.006 - 1.069, p=0.019), ELI <1.34 (hazard ratio, 1.783; 95% confidence interval 1.231 - 2.583, p=0.002) and SAPIEN valve (hazard ratio, 1.861; 95% confidence interval 1.025 - 3.382, p=0.041) were associated with increased mortality within 9 years after TAVR. Patients with ELI ≤ 1.34 had a 78% chance of all-cause mortality following TAVR. Unlike the univariate analysis, the following variables no longer were associated with mortality: hypertension (p=0.188), CoreValve (p=0.301), or SAPIEN XT (p=0.813). Although not statistically significant, the preoperative aortic valve area remained borderline (hazard ratio, 2.387; 95% confidence interval 0.961 - 5.931, p=0.061). (Table 6).