A few studies recently reported controversial results with transfemoral transcatheter aortic valve replacement (TF-TAVR) versus transapical transcatheter aortic valve replacement (TA-TAVR), often without adequate adjusted analysis for baseline differences. Data on patients who underwent TF-TAVR and TA-TAVR from the Observational Study of Effectiveness of avR–tavI procedures for severe Aortic stenosis Treatment study were analyzed with propensity score 1-to-1 matching. From a cohort of 1,654 patients (1,419 patients underwent TF-TAVR and 235 patients underwent TA-TAVR), propensity score matching resulted in 199 pairs of patients with similar operative risk (EuroSCORE II: TF-TAVR 8.1 ± 7.1% vs TA-TAVR, 8.4 ± 7.3%, p = 0.713). Thirty-day mortality was 8.0% after TA-TAVR and 4.0% after TF-TAVR (p = 0.102). Postoperative rates of stroke (TA-TAVR, 2.0% vs TF-TAVR 1.0%, p = 0.414), cardiac tamponade (TA-TAVR, 4.1% vs TF-TAVR 1.5%, p = 0.131), permanent pacemaker implantation (TA-TAVR, 8.7% vs TF-TAVR 13.3%, p = 0.414), and infection (TA-TAVR, 6.7% vs TF-TAVR 3.6%, p = 0.180) were similar in the study groups but with an overall trend in favor of TF-TAVR. Higher rates of major vascular damage (7.2% vs 1.0%, p = 0.003) and moderate-to-severe paravalvular regurgitation (7.8% vs 5.2%, p = 0.008) were observed after TF-TAVR. On the contrary, TA-TAVR was associated with higher rates of red blood cell transfusion (50.0% vs 30.4%, p = 0.0002) and acute kidney injury (stages 1 to 3: 44.4% vs 21.9%, p <0.0001) compared with TF-TAVR. Three-year survival rate was 69.1% after TF-TAVR and 57.0% after TA-TAVR (p = 0.006), whereas freedom from major adverse cardiovascular and cerebrovascular events was 61.9% after TF-TAVR and 50.4% after TA-TAVR (p = 0.011). In conclusion, TF-TAVR seems to be associated with significantly higher early and intermediate survival compared with TA-TAVR. The transfemoral approach, whenever feasible, should be considered the route of choice for TAVR.
Transcatheter aortic valve replacement through the transfemoral route (TF-TAVR) is favored over transapical TAVR (TA-TAVR) and other accesses because of its less invasive nature, which allows to be performed also with local anesthesia. However, TF-TAVR is a less direct procedure associated with a significant risk of vascular access complications, longer exposure to radiation, and increased amount of contrast agent compared with TA-TAVR. Furthermore, the transfemoral route may expose the patient to a higher risk of stroke secondary to manipulation and possible dislodgment of atherosclerotic debris from the ascending aorta and aortic arch. Although peripheral vascular complications are a major issue during TF-TAVR, transapical access also can be associated with life-threatening bleeding or the development of pseudoaneurysm at the ventricular apex. Furthermore, TA-TAVR can be associated with new hypokinesia or akinesia of the ventricular apex in a significant number of patients. A few recent meta-analyses summarizing the results of available observational studies showed an increased risk of early mortality with the transapical compared with the transfemoral approach. Despite this, the transapical route is still believed to offer a number of advantages over percutaneous peripheral approaches. We investigated this issue in a cohort of patients from the OBSERVANT (Observational Study of Effectiveness of avR–taVi procedures for severe Aortic stenosis Treatment) registry.
Methods
OBSERVANT is a national observational, prospective, multicenter, cohort study that enrolled consecutive patients with severe aortic valve stenosis undergoing TAVR or surgical aortic valve replacement at 93 Italian cardiology/cardiac surgery centers from December 2010 to June 2012. Details on the study design, patient eligibility criteria, and data collection modalities have been reported elsewhere. This study was coordinated by the Italian National Institute of Health and led in cooperation with the Italian Ministry of Health, the National Agency for Regional Health Services, Italian Regions, and Italian scientific societies and federations representing Italian professionals involved in the management of aortic valve stenosis. In the participating hospitals, both procedural treatments (surgical aortic valve replacement and/or TAVR) could be offered to patients with aortic valve stenosis (see Appendix for the complete list of executive working group, participating centers and investigators). The study protocol has been approved by the local ethics committee (ASL 2, Melegnano, Milano, Italy) of the coordinating institution (Policlinico San Donato, Milano, Italy), and the patients gave an informed consent to the scientific treatment of their data in an anonymous form. Data on demographic characteristics, health status before intervention, co-morbidities, and complete information on the type of intervention were collected into a standardized online datasheet on a password-protected website. An administrative follow-up has been set up for each enrolled patient through a record linkage with the National Hospital Discharged Records database for in-hospital events and with the Tax Registry Information System for information on survival. Collected data were stored and analyzed at the Italian National Institute of Health.
The study population included all consecutive adult patients requiring an aortic valve replacement for severe aortic valve stenosis (defined as an aortic valve area <1 cm 2 , maximum aortic velocity >4 m/s, or mean pressure gradient >40 mm Hg). Patients who underwent isolated TF-TAVR or TA-TAVR were included in this analysis.
Data auditing was performed by independent observers after specific standard operating procedures. They monitored the participating hospitals to assess the completeness of the enrolled cohort and compared the collected data with those of the original clinical records.
All-cause mortality within 30 days from intervention and up to 3 years was the primary outcome measurements of this study. Secondary outcome end points were in-hospital adverse events such as stroke, vascular complications, red blood cell transfusion, and acute kidney injury, as already reported. Other secondary outcome end points were major adverse cardiac and cerebrovascular events (MACCE) at 3 years. MACCE were defined as the composite end point including any of these adverse events: death from any cause, stroke, myocardial infarction, percutaneous coronary intervention, and/or coronary surgery.
Statistical analyses were performed using the SAS statistical package, version 9.4, (SAS Institute Inc., Cary, North Carolina). Continuous variables are presented as the mean ± SD and were compared using the analysis of variance test. Categorical variables are presented as counts and percentages and were compared with the chi-square test or Fisher’s exact test when appropriate.
As observational studies do not provide randomization, a propensity score matching method was used to select 2 groups of patients undergoing TF-TAVR and TA-TAVR, respectively, with similar baseline characteristics. The propensity score was estimated using a nonparsimonious logistic regression model with the treatment method as the dependent variable and all measured potential confounders as covariates. The following variables have been included: age, gender, previous percutaneous coronary intervention, previous balloon aortic valvuloplasty, previous cardiac surgery, previous operation on the aortoiliac arteries, dialysis, diabetes, pulmonary disease, previous myocardial infarction, peripheral artery disease, estimated glomerular filtration rate, critical preoperative state, unstable angina, neurologic dysfunction, pulmonary hypertension, chronic liver disease, active neoplastic disease, New York Heart Association class, left ventricular ejection fraction, coronary artery disease, frailty status, urgent procedure, and mitral valve regurgitation. Pairs of patients undergoing TF-TAVR and TA-TAVR having the same probability score (nearest neighbor method, caliper = 0.2×standard deviation [logitPs]) have been matched. The balance between the matched groups was evaluated with the t test for paired sample for continuous variables, the McNemar test for dichotomous variables, the Stuart-Maxwell test for categorical variables, and the analysis of the standardized differences before and after matching have been used ( Table 1 ). Differences in the outcome of propensity score–matched pairs were evaluated by the Kaplan–Meier method with the Klein–Moeschberger stratified log-rank test. Operative deaths were included in survival analyses. A p <0.05 was considered statistically significant.
| Variable | TF-TAVR (n=199) | TA-TAVR (n=199) | p-value | Standardized difference |
|---|---|---|---|---|
| Age (years±SD) | 81.5±6.2 | 81.2±6.6 | 0.721 | 0.04 |
| Men | 111 (55.8%) | 104 (52.3%) | 0.490 | 0.07 |
| BMI (kg/m2±SD) | 26.1±4.8 | 25.8±4.4 | 0.599 | 0.06 |
| Diabetes mellitus | 52 (26.1%) | 50 (25.1%) | 0.821 | 0.02 |
| eGFR (mg/min/1.73 m 2 ±SD) | 54.4±22.1 | 52.8±22.6 | 0.467 | 0.07 |
| Hemoglobin (gr/dL±SD) | 11.7±1.7 | 11.7±1.5 | 0.845 | 0.002 |
| Chronic dialysis | 5 (2.5%) | 5 (2.5%) | 1.000 | 0.00 |
| Smoking history | 37 (19.3%) | 39 (20.3%) | 0.763 | 0.03 |
| Neurologic dysfunction | 4 (2.0%) | 7 (3.5%) | 0.366 | 0.09 |
| Chronic liver disease | 7 (3.5%) | 8 (4.0%) | 0.796 | 0.03 |
| Active neoplastic disease | 6 (3.0%) | 4 (2.0%) | 0.527 | 0.06 |
| Peripheral arteriopathy | 95 (47.7%) | 96 (48.2%) | 0.888 | 0.01 |
| Pulmonary disease | 44 (22.1%) | 50 (25.1%) | 0.473 | 0.07 |
| Oxygen therapy | 11 (5.5%) | 7 (3.5%) | 0.346 | 0.10 |
| Pulmonary hypertension | 29 (15.0%) | 25 (13.0%) | 0.555 | 0.06 |
| Previous cardiac surgery | 57 (28.6%) | 50 (25.1%) | 0.431 | 0.08 |
| Previous op. on the aorta-iliac arteries | 16 (8.0%) | 20 (10.1%) | 0.450 | 0.07 |
| Previous BAV | 29 (14.6%) | 32 (16.1%) | 0.686 | 0.04 |
| Previous AMI | 54 (27.1%) | 53 (26.6%) | 0.904 | 0.01 |
| Previous PCI | 67 (33.7%) | 63 (31.7%) | 0.673 | 0.04 |
| Coronary artery disease | 51 (25.6%) | 53 (26.6%) | 0.816 | 0.02 |
| One-vessel disease | 20 (10.0%) | 34 (17.1%) | ||
| Two-vessels disease | 13 (6.5%) | 9 (4.5%) | 0.094 | 0.26 |
| Three-vessels disease | 18 (9.0%) | 10 (5.0%) | ||
| NYHA classes | ||||
| I | 8 (4.0%) | 9 (4.5%) | 0.722 | 0.08 |
| II | 56 (28.1%) | 60 (30.2%) | ||
| III | 107 (53.8%) | 99 (49.7%) | ||
| IV | 28 (14.1%) | 31 (15.6%) | ||
| Unstable angina pectoris | 2 (1.0%) | 5 (2.5%) | 0.256 | 0.12 |
| Critical preoperatve status | 4 (2.0%) | 7 (3.5%) | 0.818 | 0.09 |
| Frailty score (moderate-severe) | 35 (17.6%) | 38 (19.1%) | 0.710 | 0.04 |
| Urgent procedure | 3 (1.5%) | 5 (2.5%) | 0.479 | 0.07 |
| Logistic EuroSCORE I (%±SD) | 14.9±11.8 | 15.0±10.6 | 0.927 | 0.01 |
| Logistic EuroSCORE II (%±SD) | 8.1±7.1 | 8.4±7.3 | 0.713 | 0.04 |
Results
For the purposes of this study, 1,654 patients fulfilled the inclusion criteria and were the subjects of this analysis. Their baseline characteristics are summarized in Supplementary Table 1 . From this cohort, 1,419 patients (85.8%) underwent TF-TAVR and 235 patients (14.2%) underwent TA-TAVR. Significant differences in the baseline variables and operative risk were observed between TF-TAVR and TA-TAVR groups ( Supplementary Table 1 ). As expected, patients who underwent TA-TAVR had a significantly higher prevalence of peripheral artery disease and previous procedures on the aortoiliac arteries.
Propensity score matching resulted in 199 pairs of patients (15.3% of patients undergoing TA-TAVR were unmatched) whose baseline and echocardiographic characteristics are summarized in Tables 1 and 2 . The operative risk was similar in the study groups (EuroSCORE II: TF-TAVR 8.1 ± 7.1% vs TA-TAVR, 8.4 ± 7.3%, p = 0.713). Four covariates had a postmatching standardized difference >10%, which indicates an excellent covariate balance between the study groups. In particular, mean transvalvular gradient was greater among the patients who underwent TA-TAVR (49 ± 15 vs 46 ± 15 mm Hg, p = 0.057), but such a difference was unlikely of clinical significance given the comparable left ventricular ejection fraction in the study groups. Unstable angina was uncommon in this series and the difference between the study groups was not statistically significant. Furthermore, coronary artery disease had a standardized difference of 0.26 if considered in 4 classes (no coronary artery disease, 1 vessel, 2 vessels, and 3 vessels disease; p = 0.094), but its prevalence was similar in the study groups if considered as dichotomous covariate.
| Variable | TF-TAVR (n=199) | TA-TAVR (n=199) | p-value | Standardized difference |
|---|---|---|---|---|
| LVEF | ||||
| 30-50% | 67 (33.7%) | 64 (32.2%) | 0.721 | 0.08 |
| <30% | 6 (3.0%) | 9 (4.5%) | ||
| Mitral valve regurgitation | ||||
| Mild | 96 (48.2%) | 96 (48.2%) | 0.999 | 0.01 |
| Moderate | 52 (26.1%) | 51 (25.6%) | ||
| Aortic valve pattern | ||||
| Aortic valve area (cm 2 ±SD) | 0.68±0.28 | 0.67±0.24 | 0.406 | 0.06 |
| Peak gradient (mmHg±SD) | 76±21 | 79±22 | 0.241 | 0.12 |
| Mean gradient (mmHg±SD) | 46±15 | 49±15 | 0.057 | 0.20 |
| Annulus diameter (mm±SD) | 22.0±2.3 | 22.4±2.4 | 0.814 | 0.03 |
At 30 days, mortality was double after TA-TAVR than after TF-TAVR, although such difference did not reach statistical significance (TA-TAVR = 8.0% vs TF-TAVR = 4.0%, p = 0.102). Indeed, early mortality after TA-TAVR was similar to the expected rate as estimated by EuroSCORE II (observed/expected ratio = 0.95), whereas it was lower than the expected one after TF-TAVR (observed/expected ratio = 0.49). Postoperative stroke rate was rather low and not statistically different in the study groups, although cumulative incidence was still higher with the transapical approach. The transapical approach was associated with higher rates of cardiac tamponade and postoperative infection, whereas the rate of permanent pacemaker implantation was lower, but such differences did not reach statistical significance ( Table 3 ).
