Many patients have iliofemoral vessel anatomy unsuitable for conventional transfemoral (TF) transcatheter aortic valve implantation (TAVI). Safe and practical alternatives to the TF approach are, therefore, needed. This study compared outcomes of alternative nonfemoral routes, transapical (TA), direct aortic (DA), and subclavian (SC), with standard femoral access. In this retrospective study, data from 3,962 patients in the UK TAVI registry were analyzed. All patients who received TAVI through a femoral, subclavian, TA, or DA approach were eligible for inclusion. The primary outcome measure was survival up to 2 years. Median Logistic EuroSCORE was similar for SC, DA, and TA but significantly lower in the TF cohort (22.1% vs 20.3% vs 21.2% vs 17.0%, respectively, p <0.0001). Estimated 1-year survival rate was similar for TF (84.6 ± 0.7%) and SC (80.5 ± 3%, p = 0.27) but significantly worse for TA (74.7 ± 1.6%, p <0.001) and DA (75.2 ± 3.3%, p <0.001). A Cox proportional hazard model was used to analyze survival up to 2 years. Survival in the SC group was not significantly different from the TF group (hazard ratio [HR] 1.22, 95% confidence interval [CI] 0.88 to 1.70, p = 0.24). In contrast, survival in the TA (HR 1.74, 95% CI 1.43 to 2.11; p <0.001) and DA (HR 1.55, 95% CI 1.13 to 2.14; p <0.01) cohorts was significantly reduced compared with TF. In conclusion, TA and DA TAVI were associated with similar survival, both significantly worse than with the TF route. In contrast, subclavian access was not significantly different from TF and may represent the safest nonfemoral access route for TAVI.
Transcatheter aortic valve implantation (TAVI) has expanded rapidly as an alternative to surgical aortic valve replacement, with well >250,000 patients treated worldwide since 2002, and trial data extending into high, intermediate, and even low-risk operable patients. The transfemoral (TF) approach is generally accepted as the default access route for TAVI. However, small vessel caliber and peripheral vascular disease preclude TF access in a significant cohort of patients. In the most recent trial investigating the new Edwards SAPIEN III valve, 36% of patients required nonfemoral access. With the miniaturization in TAVI delivery systems, including the 14Fr CoreValve Evolut R, it is commonly estimated that 10% to 15% of patients will still have unsuitable femoral access. Clinicians need to understand the relative outcomes of alternative nonfemoral access routes to guide optimal treatment of these patients. Transapical (TA), subclavian (SC), and most recently direct aortic (DA) access have developed as the principal alternatives to TF. Although femoral access is routinely gained percutaneously by the operator, all nonfemoral access routes mandate a surgical cutdown to the apex of the heart, SC artery, or to the ascending aorta through a right-sided mini-thoracotomy or central sternotomy. Previous studies have reported worse outcomes, including reduced survival, after nonfemoral TAVI. However, very few data exist comparing the different nonfemoral access routes. The aim of this study was to compare morbidity and mortality associated with different nonfemoral access routes after TAVI.
Methods
The UK TAVI registry is a large prospectively collected database that includes 100% of patients who underwent TAVI in any of the 33 centers performing TAVI procedures in the United Kingdom. Detailed information about the design of the database was published previously. It includes 3,980 patients who underwent TAVI from January 2007 to December 2012. All centers use the same database as recommended by the National Institute for Cardiovascular Outcomes Research. Data are collected at each hospital, electronically encrypted, and transferred online to the National Central Cardiac Audit database. This algorithm allows for linkage to the National Health Service Central Register, with all-cause mortality tracked for patients in England and Wales by unique National Health Service number up to July 2013.
Data collected include patient demographics, indications for TAVI, procedural characteristics, and adverse outcomes including complications up to the time of hospital discharge. Survival was monitored long term.
The primary outcome measure was long-term survival up to July 2013. Secondary outcome measures were procedural and in-hospital complications (including stroke, major vascular complications, bleeding, tamponade, pacemaker implantation, and renal replacement therapy) and in-hospital, 30-day, and 1-year mortality.
Statistical analyses were performed using SPSS 20.0 (IBM Corporation, New York, New York) and Stata 12.1 (StataCorp, College Station, Texas). Continuous variables are presented as the median and interquartile range (twenty-fifth to seventy-fifth percentile). The chi-square and Kruskal-Wallis test were used as appropriate. For survival analysis, a Kaplan-Meier curve was computed and a log-rank p value was calculated. A Cox proportional hazard model was applied for the primary outcome measure, corrected for EuroScore, valve type, presence and severity of coronary artery disease, access route, heart rhythm, occurrence of postprocedural aortic regurgitation as assessed by echocardiography, and year of implantation. Covariates that were already incorporated into the EuroScore (New York Heart Association and Canadian Cardiovascular Society class, age, gender, critical preoperative state, recent myocardial infarct, kidney failure, diabetes, extracardiac arteriopathy, previous cardiac surgery, chronic lung disease, LV function, and pulmonary hypertension) were not included as separate covariates into the Cox proportional hazard model. A sensitivity analysis was performed to confirm that survival with SC access was not significantly different from a TF approach. Patients were matched 1:4 for the variables LogEuroScore, body mass index, previous cardiac surgery, valve type, extracardiac arteriopathy, and year of implant, and then a conditional logistic regression analysis was performed. A 2-sided p value <0.05 was considered significant.
This study complies with the Declaration of Helsinki. The data were collected as part of a mandatory UK national cardiac audit, and all patient-identifiable fields were removed before analysis. The National Institute for Cardiovascular Outcomes Research, which includes the UK TAVI registry, has support under section 251 of the National Health Service Act 2006. Ethical approval was not required under research governance arrangements for the analyses.
Results
In total, 3,980 patients were registered in the UK TAVI database. Three patients who underwent TAVI through a carotid approach and 15 patients with missing information on access route were excluded. Finally, 2,828 patients who underwent TAVI through TF, 761 patients with TA, 185 patients with DA, and 188 patients with SC access were included. Table 1 summarizes baseline characteristics among the study groups.
| Variable | Subclavian (n=188) | Transapical (n=761) | Direct aortic (n=185) | Femoral (n=2828) | p-value |
|---|---|---|---|---|---|
| Age (years) | 83 (78-86) | 82 (77-86) | 84 (77-88) | 83 (77-87) | 0.18 |
| Male | 123 (65%) | 425 (56%) | 90 (49%) | 1451 (51%) | <0.0001 |
| Logistic EuroScore | 22 (14-34) | 20 (14-31) | 21.2 (15-33) | 17 (11-26) | <0.0001 ∗ |
| Body-mass-index (kg/m 2 ) | 26 (23-29) | 26 (23-29) | 25 (23-29) | 26 (23-30) | 0.005 ∗ |
| Creatinine | 108 (86-136) | 104 (86-132) | 100 (81-136) | 101 (82-130) | 0.03 ∗ |
| Diabetes | 45 (23%) | 167 (22%) | 39 (21%) | 629 (23%) | 0.8 |
| Smoker | <0.0001 ∗ | ||||
| Ex-smoker | 109 (61%) | 395 (54%) | 106 (59%) | 1334 (50%) | |
| Current smoker | 9 (5.0%) | 25 (3.0%) | 10 (6.0%) | 56 (2.1%) | |
| Atrial fibrillation | 32 (17%) | 163 (22%) | 30 (16%) | 591 (21%) | 0.01 |
| Neurological disease | 35 (19%) | 127 (17%) | 28 (15%) | 409 (15%) | 0.42 |
| Previous MI | 52 (28%) | 166 (22%) | 43 (23%) | 622 (22%) | 0.08 † |
| Previous PCI | 45 (24%) | 154 (20%) | 33 (18%) | 606 (21%) | 0.52 |
| Previous cardiac surgery | 63 (33%) | 324 (42%) | 44 (24%) | 830 (29%) | <0.0001 |
| Coronary artery disease | 94 (51%) | 406 (55%) | 88 (48%) | 1155 (42%) | <0.0001 |
| Left ventricular ejection fraction (LVEF) | 0.3 | ||||
| >50% | 99 (53%) | 466 (62%) | 105 (57%) | 1714 (61%) | |
| 30-49% | 68 (36%) | 229 (30%) | 60 (32%) | 818 (29%) | |
| <30% | 19 (10%) | 57 (8%) | 20 (11%) | 272 (10%) | |
| Aortic gradient (mmHg) | 75 (60-98) | 74 (61-90) | 76 (63-94) | 75 (61-92) | 0.17 |
| Aortic valve area (cm 2 ) | 0.7 (0.5-0.8) | 0.64 (0.5-0.8) | 0.60 (0.5-0.8) | 0.65 (0.5-0.8) | 0.19 |
| Aortic annulus (mm) | 23 (21-25) | 22 (21-24) | 23 (21-24) | 23.0 (21-24) | <0.0001 |
| Pulmonary hypertension | 28 (15%) | 70 (10%) | 28 (15%) | 401 (15%) | 0.007 ∗ |
| Chronic lung disease | 52 (28%) | 200 (27%) | 71 (39%) | 738 (27%) | 0.01 |
| Procedural characteristics | |||||
| Local anaesthesia | 5 (3.0%) | 15 (2.0%) | 2 (1.0%) | 501 (18%) | <0.0001 |
| Procedural characteristics | <0.0001 | ||||
| CoreValve | 186 (99%) | 3 (0.4%) | 67 (36%) | 1626 (58%) | |
| Edwards | 2 (1.0%) | 753 (99) | 118 (64) | 1161 (41%) | |
| Valve-in-valve procedure | 3 (2.0%) | 44 (6.0%) | 5 (3.0%) | 97 (3.4%) | 0.1 |
| Valve size (mm) | 29 (26-29) | 26 (23-26) | 26 (23-29) | 26 (26-29) | <0.0001 |
| Aortic regurgitation post implant | |||||
| moderate to severe | 15 (8.0%) | 21 (4.0%) | 8 (5.0%) | 214 (8.0%) | <0.0001 |
∗ Non-significant among non-femoral access routes.
Inhospital mortality was lowest in the TF group (3.7%, n = 105, p <0.0001 vs pooled non-TF). Among the nonfemoral access groups, only the SC route (4.3%, n = 8, p = 0.69) was not significantly different from TF, whereas TA (9.5%, n = 72, p <0.0001) and DA (7.6%, n = 14, p <0.02) were associated with higher mortality. Inhospital morbidity is summarized in Table 2 .
| Variable | Subclavian (n=188) | Transapical (n=761) | Direct aortic (n=185) | Femoral (n=2828) | p-value |
|---|---|---|---|---|---|
| In-hospital death | 8 (4.3%) | 72 (9.5%) | 14 (7.6%) | 105 (3.7%) | <0.0001 |
| 30-day mortality | 5 (2.9%) | 80 (11%) | 15 (8.4%) | 121(4.7%) | <0.0001 |
| 12-months mortality | 33 (20%) | 187 (27%) | 42 (29%) | 388 (18%) | <0.0001 |
| Stroke in hospital | 6 (3.0%) | 23 (3.0%) | 1 (1.0%) | 58 (2.1%) | 0.12 |
| TIA in hospital | 3 (2.0%) | 4 (1.0%) | 0 | 16 (0.6%) | 0.22 |
| Tamponade | 4 (2.0%) | 4 (1.0%) | 1 (1.0%) | 22 (0.8%) | 0.07 |
| Major vascular complication | 4 (2.0%) | 3 (0.4%) | 6 (3.0%) | 98 (3.5%) | <0.0001 |
| Need for vascular surgery | 3 (2.0%) | 7 (1.0%) | 3 (2.0%) | 63 (2.3%) | 0.59 |
| Emergency valve in valve procedure | 7 (4.0%) | 7 (1.0%) | 2 (1.0%) | 77 (2.7%) | 0.01 |
| Need for haemofiltration | 7 (4.0%) | 54 (7.0%) | 19 (10%) | 71 (2.5%) | <0.0001 |
| GI bleeding | 2 (1.0%) | 15 (2.0%) | 0 | 21 (0.8%) | 0.01 |
| Pacemaker implantation post TAVI | 43 (23%) | 37 (5.0%) | 13 (7.0%) | 363 (13%) | <0.0001 |
| Hospital stay (days) | 7.0 (5.0-10.0) | 8.0 (5.0-15.0) | 8.0 (5.0-16.0) | 5.5 (4.0-8.0) | <0.0001 |
| Follow-up (days) | 609 (312-994) | 567 (225-1056) | 421(202-680) | 544 (283-929) | <0.0001 |
An unadjusted Kaplan-Meier survival chart is shown in Figure 1 . There was no difference in survival rate between TA (1-year estimator 74.7 ± 1.6%, p <0.0001) and DA (1-year estimator 75.2 ± 3.3%) approaches, both of which were associated with significantly lower long-term survival rate than TF (1-year estimator 84.6 ± 0.7%, p <0.0001). In contrast, unadjusted survival rate of the SC cohort was not significantly different from TF (1-year estimator 80.5 ± 3%, p = 0.27).
