The aim of this study was to analyze procedural and postprocedural outcomes of patients aged ≥80 years treated by transcatheter aortic valve implantation (TAVI) or surgical aortic valve replacement (SAVR) as enrolled in the OBservational Study of Effectiveness of SAVR-TAVR procedures for severe Aortic steNosis Treatment (OBSERVANT) Study. TAVI is offered to patients with aortic stenosis judged inoperable or at high surgical risk. Nevertheless, it is common clinical practice to treat elderly (≥80 years) patients by TAVI regardless of surgical risk for traditional SAVR. OBSERVANT is a multicenter, observational, prospective cohort study that enrolled patients with symptomatic severe aortic stenosis who underwent SAVR or TAVI from December 2010 to June 2012 in 93 Italian participating hospitals. Information on demographic characteristics, health status before intervention, therapeutic approach, and intraprocedural and 30-day outcomes was collected. An administrative follow-up was set up to collect data on midterm to long-term outcomes. We reviewed baseline and procedural data of patients aged ≥80 years, looking for different early and late outcome after TAVI or SAVR. Patients treated by TAVI were sicker than SAVR because of higher rate of co-morbidities, advanced illness, frailty, and Logistic EuroSCORE. After propensity matching, early and midterm mortality were comparable between the 2 groups. However, patients treated by TAVI had higher rate of vascular complications (6.0% vs 0.5%; p <0.0001), permanent pacemaker implantation (13.4% vs 3.7%; p <0.0001), and paravalvular leak (8.9% vs 2.4%; p <0.0001). Patients who underwent SAVR had more frequent bleedings needing transfusion (63.2% vs 34.5%; p <0.0001) and acute kidney injury (9.6% vs 3.9%; p = 0.0010). In conclusion, patients aged ≥80 years treated by TAVI or SAVR had similar early and midterm mortality.
Transcatheter aortic valve implantation (TAVI) is considered the procedure of choice in patients with severe aortic stenosis (AS) judged inoperable or as alternative to surgical aortic valve replacement (SAVR) in patients at high risk. Although older age is not a contraindication to SAVR, it is a common clinical practice to refer patients aged ≥80 years to TAVI although the surgical risk is not prohibitive or high. Data on the impact of percutaneous versus surgical choice in older patients with severe AS are scant, especially if intermediate-to-low risk categories of patients are considered. The OBservational Study of Effectiveness of SAVR-TAVR procedures for severe Aortic steNosis Treatment study (OBSERVANT) is the first, single-nation, multicenter cohort study enrolling patients with severe AS treated by TAVI or SAVR. The aim of current OBSERVANT subanalysis was to assess both potential benefit and caveats of isolated TAVI choice compared to isolated SAVR in patients aged ≥80 years with severe AS, focusing on propensity-matched patients not at prohibitive risk for traditional SAVR.
Methods
OBSERVANT is a multicenter, observational, prospective cohort study that enrolled 7,618 consecutive patients with symptomatic severe AS who underwent SAVR or TAVI from December 2010 to June 2012 in 93 Italian participating hospitals (34 interventional catheterization laboratories and 59 cardiac surgeries). Information on demographic characteristics, health status before intervention, therapeutic approach, early (intraprocedural and 30 day) outcomes (including death, myocardial infarction, stroke, tamponade, shock, major vascular complications, permanent pacemaker [PPM] implantation, acute renal failure, blood transfusion, valve migration, emergent percutaneous coronary intervention, infections, and valve performance) was collected. An administrative follow-up was set up to collect data on midterm to long-term outcomes (including death, acute myocardial infarction, stroke, percutaneous coronary or surgical revascularization, or combined major adverse cardiac and cerebrovascular events [MACCE] which included all of them). This administrative follow-up was obtained through crosslinking with Regional Social Security Death and Events Master files. This approach guarantees a very low rate of lost to follow-up.
TAVI was performed indifferently with balloon-expandable Edwards Sapien XT device (Edwards LifeSciences, Irvine, CA) or self-expandable CoreValve device (Medtronic, Inc., Minneapolis, MN) according to single operator or center preference. The type of vascular access was selected in each case according to clinical judgment on the basis of the results of a multimodality imaging screening including aortic angiograms and angio–computed tomographic scans and operator experience.
Active endocarditis, chronic obstructive pulmonary disease, neurologic dysfunction, peripheral vascular disease, critical preoperative state, pulmonary hypertension, and unstable angina at baseline were defined according to EuroSCORE definitions. Frailty status was defined according to Geriatric Status Scale scoring. Hemodynamic performance of implanted prosthesis (transvalvular gradient, intraprosthesis and periprosthesis regurgitation) was evaluated according to echocardiogram criteria. To define postprocedural acute myocardial infarction, type 5 criteria from Universal Definition of Acute Myocardial Infarction was considered. Major vascular complications were defined as the need of surgical intervention or endoprosthesis, including aortic dissection but not the complications on the site of apical access which were counted separately. Stroke included only major stroke or coma status. Acute kidney injury included only the need of new dialitic treatment or continuous venous–venous ultrafiltration. Shock was defined as acute heart failure needing pharmacologic and/or mechanical support.
Statistical analyses were performed using the SAS statistical package version 9.2 (SAS Institute Inc., Cary, North Carolina). Continuous variables are presented as the mean ± standard deviations and were compared using the Student t test or Mann–Whitney test when appropriate. Categorical variables are presented as counts and percentages and were compared with the chi-square test or Fisher’s exact test when appropriate. The propensity score was estimated using a nonparsimonious logistic regression model with the treatment method as the dependent variable. The following variables have been included as covariates: age, gender, previous percutaneous coronary intervention, previous balloon aortic valvuloplasty, previous cardiac surgery, previous operation on the aorta, chronic dialytic treatment, diabetes, chronic obstructive pulmonary disease, previous myocardial infarction, peripheral arteriopathy, estimated glomerular filtration rate, critical preoperative state, unstable angina, neurologic dysfunction, pulmonary hypertension (systolic pulmonary arterial pressure >60 mm Hg), chronic liver disease, active neoplastic disease, New York Heart Association class, frailty score (Geriatric Status Scale ), left ventricular ejection fraction, coronary artery disease, urgency status, and mitral regurgitation. Patients who underwent TAVI and SAVR have been matched by propensity score using the nearest neighbor method (caliper = 0.2 × DS [logitPs]). Patients judged inoperable because of porcelain aorta, hostile thorax, need of emergent cardiac surgery, or need for concomitant coronary artery bypass were excluded from matching process. The analysis of the standardized differences after matching has been used to evaluate the balance between the matched groups; a standardized difference of <0.1 has been taken to indicate a negligible difference in the mean or prevalence of a covariate between treatment groups. The t test for paired sample (continuous variables), the McNemar test (dichotomous variables), and the Stuart–Maxwell test (categorical variables) were computed for all baseline covariate, and the same tests have been considered to compare periprocedural adverse events. The Kaplan–Meier method with the Klein–Moeschberger stratified log-rank test has been used to evaluate differences in the midterm to long-term outcomes (3-year mortality and MACCE).
Multivariable Cox regression on the matched cohort was used to establish the contribution of short-term outcomes on 3-year mortality adjusting for EuroSCORE to take into account potential confounders. The outcomes considered as predictors of mortality were those resulted associated with treatment (stroke, shock, PPM, major vascular complications, acute kidney injury, paravalvular leak). Presence of transfusions was not included in the model because in previous reports, it is associated with acute kidney injury. The results are expressed as hazard ratio (HR) with 95% confidence interval (CI) and p value. A p value of <0.05 was considered statistically significant in all the analysis.
Results
The flow algorithm of the study is presented in Figure 1 . This study included 2,820 patients aged ≥80 years who underwent either isolated TAVI or isolated SAVR. Baseline characteristics are presented in Table 1 . Patients who underwent TAVI were more symptomatic and sicker than who underwent SAVR. They had more co-morbidities such as diabetes mellitus, chronic obstructive pulmonary disease, chronic kidney disease, peripheral vasculopathy, pulmonary hypertension, cancer, neurologic dysfunction, or history of coronary artery disease. They presented more frequent multivessels coronary artery disease, left ventricular dysfunction, and mitral regurgitation. The frailty and Logistic EuroSCORE were also worse in patients who underwent TAVI ( Table 1 ).
| Variable | SAVR (n = 983) | TAVI (n = 1178) | p Value |
|---|---|---|---|
| Age (years) | 82.7 ± 2.9 | 84.8 ± 3.2 | <0.0001 |
| Women | 587 (59.8%) | 724 (61.5%) | 0.4644 |
| Body mass index (Kg/m 2 ) | 26.4 ± 4.1 | 25.6 ± 4.5 | <0.0001 |
| Smoker | 107 (11.4%) | 100 (8.8%) | 0.0472 |
| Diabetes mellitus | 148 (15.1%) | 284 (24.1%) | <0.0001 |
| Creatinine (mg/dL) | 1.1 ± 0.5 | 1.2 ± 0.7 | <0.0001 |
| Estimated glomerular filtration rate (mL/min/1.73m 2 ) | 61.2 ± 19.0 | 53.2 ± 19.8 | <0.0001 |
| Chronic dialytic treatment | 5 (0.5%) | 18 (1.5%) | 0.0207 |
| Albumin (mg/dL) | 3.7 ± 0.9 | 3.5 ± 0.8 | 0.0001 |
| Hemoglobin (mg/dL) | 12.2 ± 1.6 | 11.6 ± 1.6 | <0.0001 |
| Previous myocardial infarction | 53 (5.5%) | 191 (16.4%) | <0.0001 |
| Unstable angina pectoris | 23 (2.4%) | 34 (2.9%) | 0.5633 |
| Chronic obstructive pulmonary disease | 95 (9.7%) | 297 (25.5%) | <0.0001 |
| Active neoplastic disease | 9 (0.9%) | 33 (2.8%) | 0.0015 |
| Pulmonary hypertension | 62 (6.8%) | 209 (18.5%) | <0.0001 |
| Oxygen dependency | 11 (1.2%) | 57 (4.9%) | <0.0001 |
| Neurological dysfunction | 25 (2.6%) | 79 (6.8%) | <0.0001 |
| Chronic liver disease | 16 (1.7%) | 21 (1.8%) | 0.8530 |
| Peripheral arteriopathy | 150 (15.6%) | 292 (25.3%) | <0.0001 |
| Previous cardiac surgery | 35 (3.6%) | 153 (13.0%) | <0.0001 |
| Previous balloon aortic valvuloplasty | 22 (2.2%) | 185 (15.7%) | <0.0001 |
| Previous operation on the aorta | 22 (2.2%) | 51 (4.3%) | 0.0068 |
| Frailty class ≥ 2 | 89 (9.1%) | 299 (25.4%) | <0.0001 |
| Previous percutaneous coronary intervention | 93 (9.5%) | 338 (28.7%) | <0.0001 |
| Critical preoperative state | 18 (1.8%) | 41 (3.5%) | 0.0169 |
| New York Heart Association functional class | <0.0001 | ||
| I | 144 (14.6%) | 35 (3.0%) | |
| II | 398 (40.5%) | 337 (28.6%) | |
| III | 361 (36.7%) | 644 (54.7%) | |
| IV | 73 (7.4%) | 143 (12.1%) | |
| Logistic EuroSCORE I | 8.0 ± 5.7 | 14.9 ± 11.8 | <0.0001 |
| Number of narrowed coronary arteries | <0.0001 | ||
| 0 | 843 (85.8%) | 825 (70.0%) | |
| 1 | 89 (9.1%) | 207 (17.6%) | |
| 2 | 30 (3.1%) | 75 (6.4%) | |
| 3 | 21 (2.1%) | 71 (6.0%) | |
| Mitral valve regurgitation | <0.0001 | ||
| None | 449 (45.7%) | 195 (16.6%) | |
| Mild | 378 (38.5%) | 632 (53.7%) | |
| Moderate | 141 (14.3%) | 320 (27.2%) | |
| Severe | 15 (1.5%) | 31 (2.6%) | |
| Aortic valve | |||
| Area (cm 2 ) | 0.7 ± 0.2 | 0.6 ± 0.3 | <0.0001 |
| Peak gradient (mm Hg) | 84.9 ± 23.1 | 81.3 ± 22.6 | 0.0004 |
| Mean gradient (mm Hg) | 52.6 ± 15.4 | 50.1 ± 14.9 | 0.0001 |
| Annulus diameter (cm) | 21.3 ± 2.1 | 22.0 ± 2.2 | <0.0001 |
| Procedural characteristics | |||
| Urgent procedure | 48 (4.9%) | 27 (2.3%) | 0.0011 |
| General anestesia | 970 (99.3%) | 404 (34.3%) | <0.0001 |
| Left ventricular ejection fraction (%) | <0.0001 | ||
| ≤ 30 | 745 (75.8%) | 814 (69.1%) | |
| 30 – 50 | 138 (14.0%) | 305 (25.9%) | |
| > 50 | 15 (1.5%) | 51 (4.3%) | |
Of 1,178 patients who underwent TAVI, 624 (53%) received a CoreValve prosthesis. Most patients who underwent TAVI had a transfemoral access (87.5%).
The propensity score generated 415 pairs of patients with similar baseline characteristics as confirmed by a standardized difference <0.1 in all the considered baseline variables ( Figure 2 ). Baseline characteristics of matched population are summarized in Table 2 .
| Variable | SAVR (n = 415) | TAVI (n = 415) | p Value |
|---|---|---|---|
| Age (years) | 83.7 ± 2.6 | 83.7 ± 2.9 | 0.9130 |
| Women | 249 (60.0%) | 257 (61.9%) | 0.5697 |
| Body mass index (Kg/m 2 ) | 26.2 ± 4.1 | 25.9 ± 4.3 | 0.3679 |
| Smoker | 41 (10.4%) | 40 (10.1%) | 0.9068 |
| Diabetes mellitus | 80 (19.3%) | 77 (18.6%) | 0.7948 |
| Creatinine (mg/dL) | 1.1 ± 0.6 | 1.1 ± 0.4 | 0.6599 |
| Estimated glomerular filtration rate (mL/min/1.73m 2 ) | 58.9 ± 19.2 | 57.5 ± 18.7 | 0.2711 |
| Chronic dialytic treatment | 1 (0.2%) | 3 (0.7%) | 0.3173 |
| Albumin (mg/dL) | 3.7 ± 0.8 | 3.5 ± 0.8 | 0.0301 |
| Hemoglobin (mg/dL) | 12.2 ± 1.6 | 11.7 ± 1.6 | 0.0004 |
| Previous myocardial infarction | 36 (8.7%) | 30 (7.2%) | 0.4461 |
| Unstable angina pectoris | 13 (3.1%) | 10 (2.4%) | 0.4913 |
| Chronic obstructive pulmonary disease | 73 (17.6%) | 63 (15.2%) | 0.3573 |
| Active neoplastic disease | 8 (1.9%) | 7 (1.7%) | 0.7963 |
| Pulmonary hypertension | 42 (10.9%) | 51 (13.2%) | 0.3051 |
| Oxygen dependency | 8 (1.9%) | 8 (1.9%) | 1.0000 |
| Neurological dysfunction | 15 (3.6%) | 18 (4.3%) | 0.6015 |
| Chronic liver disease | 7 (1.7%) | 8 (1.9%) | 0.7815 |
| Peripheral arteriopathy | 72 (17.3%) | 72 (17.3%) | 1.0000 |
| Previous cardiac surgery | 21 (5.1%) | 21 (5.1%) | 1.0000 |
| Previous balloon aortic valvuloplasty | 12 (2.9%) | 17 (4.1%) | 0.3173 |
| Previous operation on the aorta | 13 (3.2%) | 10 (2.4%) | 0.4142 |
| Frailty class ≥ 2 | 62 (14.9%) | 63 (15.2%) | 0.9230 |
| Previous percutaneous coronary intervention | 55 (13.3%) | 49 (11.8%) | 0.5176 |
| Critical preoperative state | 10 (2.4%) | 10 (2.4%) | 1.0000 |
| New York Heart Association functional class | 0.9324 | ||
| I | 21 (5.1%) | 22 (5.3%) | |
| II | 149 (35.9%) | 150 (36.1%) | |
| III | 201 (48.4%) | 194 (46.7%) | |
| IV | 44 (10.6%) | 49 (11.8%) | |
| Logistic EuroSCORE I | 9.9 ± 6.9 | 9.9 ± 6.4 | 0.8727 |
| Number of narrowed coronary arteries | 0.7696 | ||
| 0 | 336 (81.0%) | 342 (82.4%) | |
| 1 | 48 (11.6%) | 49 (11.8%) | |
| 2 | 18 (4.3%) | 15 (3.6%) | |
| 3 | 13 (3.1%) | 9 (2.2%) | |
| Mitral valve regurgitation | 0.7868 | ||
| None | 116 (28.0%) | 111 (26.7%) | |
| Mild | 214 (51.6%) | 211 (50.8%) | |
| Moderate | 75 (18.1%) | 85 (20.5%) | |
| Severe | 10 (2.4%) | 8 (1.9%) | |
| Aortic valve | |||
| Area (cm 2 ) | 0.7 ± 0.2 | 0.7 ± 0.3 | 0.0628 |
| Peak gradient (mm Hg) | 84.8 ± 22.5 | 83.9 ± 19.9 | 0.5752 |
| Mean gradient (mm Hg) | 52.7 ± 15.3 | 52.4 ± 13.5 | 0.8203 |
| Annulus diameter (cm) | 21.1 ± 2.1 | 22.0 ± 2.1 | <0.0001 |
| Procedural characteristics | |||
| Urgent procedure | 17 (4.1%) | 16 (3.9%) | 0.8575 |
| General anestesia | 410 (99.0%) | 143 (34.5%) | <0.0001 |
| Left ventricular ejection fraction (%) | 0.7614 | ||
| ≤ 30 | 315 (75.9%) | 306 (73.7%) | |
| 30 – 50 | 83 (20.0%) | 91 (21.9%) | |
| > 50 | 8 (1.9%) | 9 (2.2%) | |
The early outcomes of the matched population according to the type of treatment are reported in Table 3 . Midterm mortality and MACCE are reported in Figure 3 . We failed to find statistically significant differences in 30-day, mid-to-long term mortality, and in MACCE at follow-up. As expected, procedure-specific complications arose after the 2 techniques: patients who underwent TAVI had higher rate of vascular complications and PPM implantation; patients who underwent SAVR had more frequent bleedings needing transfusion, periprocedural major stroke, acute renal failure, and shock. Related to the performance of the prosthesis, transcatheter aortic valve had a higher rate of paravalvular leak compared to SAVR but lower transprosthesis gradient ( Table 3 ).
