Comparisons of transcatheter aortic valve implantation (TAVI) to surgical aortic valve replacement (SAVR) in patients with severe aortic stenosis remain sparse or limited by a short follow-up. We sought to evaluate early and midterm outcomes of consecutive patients (n = 618) undergoing successful TAVI (n = 218) or isolated SAVR (n = 400) at 2 centers. The primary end point was incidence of Valvular Academic Research Consortium–defined major adverse cerebrovascular and cardiac events (MACCEs) up to 1 year. Control of potential confounders was attempted with extensive statistical adjustment by covariates and/or propensity score. In-hospital MACCEs occurred in 73 patients (11.8%) and was more frequent in patients treated with SAVR compared to those treated with TAVI (7.8% vs 14.0%, p = 0.022). After addressing potential confounders using 3 methods of statistical adjustment, SAVR was consistently associated with a higher risk of MACCEs than TAVI, with estimates of relative risk ranging from 2.2 to 2.6 at 30 days, 2.3 to 2.5 at 6 months, and 2.0 to 2.2 at 12 months. This difference was driven by an adjusted increased risk of life-threatening bleeding at 6 and 12 months and stroke at 12 months with SAVR. Conversely, no differences in adjusted risk of death, stroke and myocardial infarction were noted between TAVI and SAVR at each time point. In conclusion, in a large observational registry with admitted potential for selection bias and residual confounding, TAVI was not associated with a higher risk of 1-year MACCEs compared to SAVR.
Although some large single-arm registries have demonstrated that short- and long-term outcomes after transcatheter aortic valve implantation (TAVI) have the potential to compare favorably to surgical aortic valve replacement (SAVR) in high-risk patients with aortic stenosis, direct comparisons of TAVI to SAVR remain limited to a single randomized clinical trial and 2 observational studies reporting 30-day adjusted outcomes. These studies did not use the Valvular Academic Research Consortium (VARC) definitions that are currently available and should be used to enable meaningful comparisons across different TAVI studies. The aim of this study was to investigate early and 1-year outcomes of TAVI compared to SAVR.
Methods
Consecutive patients who underwent successful TAVI or SAVR for severe aortic stenosis from January 2005 through March 2011 in 2 centers were included in a central database. In total 206 patients with SAVR were excluded because of concomitant coronary artery bypass grafting (n = 150), mitral valve procedures (n = 55), ventricular repair/aortic homograft (n = 92), and/or septal myectomies (n = 49). Follow-up data were available in all cases. The local ethics committee at each center approved the use of clinical data for this study, and all patients provided written informed consent. For TAVI and SAVR screening included a comprehensive clinical evaluation, transthoracic echocardiography, and, if necessary, transesophageal echocardiography, coronary angiography, aortic and iliofemoral angiography, and 64-multislice computed tomography. A multidisciplinary team including cardiologists, cardiothoracic surgeons, anesthesiologists, geriatricians, and interventional cardiologists evaluated all available clinical and imaging data and a consensus decision was obtained to determine individual eligibility for TAVI. The decision to perform TAVI was conditional on the presence of severe symptomatic aortic stenosis with contraindications to or high-risk for SAVR, life expectancy ≥1 year, anatomy suitable for intervention, and no need for coronary artery bypass surgery.
All TAVI procedures were performed under local anesthesia and analgesia under fluoroscopic guidance in a standard cardiac catheterization laboratory with surgical back-up. General anesthesia was employed when the trans-subclavian approach was used. Technical details of the transfemoral and trans-subclavian procedures have been detailed previously. Briefly, the stenotic valve was crossed in retrograde fashion and predilated with an appropriately sized balloon. Next, the self-expanding CoreValve ReValving System (Medtronic, Inc., Minneapolis, Minnesota) or the balloon-expandable Edwards SAPIEN XT (Edwards Life Sciences, Irvine, California) devices were positioned across the aortic annulus and deployed with aortic root angiographic guidance at each stage to ensure optimal deployment. Upon completion of the procedure, all patients were transferred to the cardiac intensive care unit for postprocedure monitoring and continued medical management. Vital parameters, serum electrolytes, cardiac biomarkers (creatine kinase-MB and troponin I), and hematologic indexes were monitored as described elsewhere. The surgical technique was the same for all patients undergoing SAVR. After general anesthesia, a median sternotomy was performed together with standard cannulation from the right atrium and ascending aorta. Extracorporeal circulation was set up under moderate hypothermia with nonpulsating flow. Myocardial protection was ensured by intermittent cold blood cardioplegia. The choice of implanting a biological or mechanical prosthesis and the technique for SAVR were left to the cardiac surgeon’s discretion and standard practice. After the intervention all patients were transferred to the intensive care unit for postintervention monitoring.
All patients received clinical and echocardiographic follow-up at 1 month, 6 and 12 months and every 12 months thereafter. All relevant clinical, procedural, and adverse event data were recorded in a prospective database. Referring cardiologists, general practitioners, and patients were contacted whenever necessary for further information. All outcomes of interest were confirmed by source documentation collected at each center and were centrally adjudicated by an independent blinded end-points committee. Major neurologic events were adjudicated by multidisciplinary consensus. To maximize quality ascertainment of the primary end point, administrative data on survival status were obtained. Clinical end points were defined according to the VARC. Major adverse cerebrovascular and cardiac events (MACCEs) were defined as the composite of death from any cause, spontaneous myocardial infarction, and stroke, urgent or emergency conversion to surgery, or life-threatening/disabling bleeding. Cardiovascular mortality, spontaneous myocardial infarction, stroke, stage 3 acute kidney injury (AKI), and life-threatening/disabling bleeding were also defined according to VARC definitions.
Continuous variables are presented as mean ± SD or median and interquartile range and were compared using Student’s unpaired t test or Mann–Whitney rank-sum test based on a test for normal distribution. Categorical variables are presented as count and percentage and were compared using chi-square test when appropriate (expected frequency >5). Otherwise, Fisher’s exact test was used. Kaplan–Meier estimates were used to plot rates of clinical end points in patients with TAVI and SAVR and differences between groups were analyzed with log-rank test. To decrease the effect of selection bias and potential confounding in this observational study, outcome parameters were adjusted by a Cox multivariable proportional hazard regression model for observed differences with respect to variables with a p value <0.20 in univariate analysis, which were age, hypertension, dyslipidemia, previous congestive heart failure, previous myocardial infarction, previous stroke, previous percutaneous coronary intervention, chronic renal failure, chronic obstructive pulmonary disease, porcelain aorta, previous pacemaker implantation, New York Heart Association class III to IV, logistic European System for Cardiac Operative Risk Evaluation score, left ventricular ejection fraction, and mean transaortic gradient. The assumption of the proportional hazard was verified by a visual examination of the log(−log) curves and the linearity assumption was assessed by plotting the martingale residuals against continuous covariates. Control of potential confounders was also attempted by developing a propensity score using logistic regression. The propensity score is the conditional probability of being part of a group according to a set of measured covariates. In our context, it was computed for each patient using a logistic regression model including logistic European System for Cardiac Operative Risk Evaluation score, age, hypertension, chronic renal failure, chronic obstructive pulmonary disease, previous coronary artery bypass grafting, previous percutaneous coronary intervention, previous myocardial infarction, previous transient ischemic attack/stroke, previous acute heart failure, dyslipidemia, New York Heart Association class III to IV, mean transaortic gradient, and left ventricular ejection fraction. Selection of variables, which formed a “minimum relevant” information set, was based on a close relation with the treatment effect and/or choice of treatment as assessed by univariate analysis. The population was then divided into quintiles according to propensity score. Within each quintile, mean propensity scores of the 2 groups were compared, as were their clinical and procedural characteristics. Covariate interactions and higher-order terms for continuous variables proved unnecessary for the balance of baseline characteristics across quintiles. The model was well calibrated (Hosmer–Lemeshow test 0.23) with a good discrimination (c-statistic 0.79). Clinical end-points adjustment was performed for the listed covariates, for propensity score alone, and for covariates with propensity score added as an additional covariate. For all analyses a 2-sided p value <0.05 was considered statistically significant. All data were processed using SPSS 17 (SPSS, Inc., Chicago, Illinois).
Results
Baseline characteristics of 618 patients undergoing successful TAVI (n = 218) or isolated SAVR (n = 400) procedures in the study period are listed in Table 1 .
| Variable | Overall Population (n = 618) | TAVI Group (n = 218) | SAVR Group (n = 400) | p Value ⁎ |
|---|---|---|---|---|
| Age (years), mean ± SD | 74.0 ± 9.9 | 80.9 ± 5.2 | 70.3 ± 9.9 | <0.001 |
| Body mass index (kg/m 2 ), mean ± SD | 26.6 ± 4.8 | 26.6 ± 4.8 | 26.6 ± 4.9 | 0.668 |
| Women | 322 (52.1%) | 117 (53.7%) | 205 (51.2%) | 0.565 |
| Hypertension | 438 (70.9%) | 186 (85.3%) | 252 (63.0%) | <0.001 |
| Diabetes mellitus | 142 (22.9%) | 53 (24.3%) | 143 (89%) | 0.560 |
| Bicuspid aortic valve | 27 (4.4%) | 0 (0%) | 27 (6.8%) | <0.001 |
| Dyslipidemia † | 293 (47.4%) | 116 (53.2%) | 177 (44.3%) | 0.033 |
| Acute heart failure ‡ | 152 (24.6%) | 86 (39.4%) | 66 (16.5%) | <0.001 |
| Previous myocardial infarction | 74 (11.9%) | 42 (19.3%) | 32 (8.0%) | <0.001 |
| Previous stroke | 32 (5.2%) | 16 (7.4%) | 16 (4.0%) | 0.073 |
| Previous transient ischemic attack | 28 (4.5%) | 14 (6.4%) | 14 (3.5%) | 0.095 |
| Previous coronary bypass graft surgery | 31 (5.0%) | 25 (11.5%) | 6 (1.5%) | <0.001 |
| Previous cardiac surgery § | 12 (1.9%) | 5 (2.3%) | 7 (1.8%) | 0.424 |
| Previous percutaneous coronary intervention | 97 (15.7%) | 68 (31.2%) | 29 (7.3%) | <0.001 |
| Peripheral vascular disease | 43 (6.9%) | 13 (5.9%) | 30 (7.5%) | 0.473 |
| Chronic obstructive pulmonary disease | 152 (24.6%) | 74 (33.9%) | 78 (19.5%) | <0.001 |
| Cirrhosis Child class A or B | 8 (1.3%) | 4 (1.8%) | 4 (1.0%) | 0.299 |
| Renal insufficiency ‖ | 59 (9.5%) | 51 (23.4%) | 8 (2.0%) | <0.001 |
| Atrial fibrillation | 57 (9.2%) | 28 (12.8%) | 29 (7.3%) | 0.017 |
| Previous pacemaker | 25 (3.2%) | 21 (9.6%) | 4 (1.0%) | <0.001 |
| Porcelain aorta ¶ | 30 (4.8%) | 28 (12.8%) | 2 (0.5%) | <0.001 |
| New York Heart Association class III and IV | 231 (37.4%) | 136 (62.4%) | 166 (41.5%) | 0.003 |
| Logistic European System for Cardiac Operative Risk Evaluation score (%), mean ± SD | 11.8 ± 11.8 | 21.1 ± 14.2 | 6.8 ± 5.9 | <0.001 |
| Society of Thoracic Surgeons score, mortality (%), mean ± SD | 4.6 ± 4.0 | 8.5 ± 4.3 | 2.5 ± 1.9 | <0.001 |
| Baseline echocardiographic parameters | ||||
| Left ventricular ejection fraction (%), mean ± SD | 53.9 ± 9.7 | 51.1 ± 10.6 | 55.4 ± 8.9 | <0.001 |
| Peak pressure gradient (mm Hg), mean ± SD # | 94.4 ± 24.8 | 89.8 ± 27.6 | 96.1 ± 24.2 | 0.015 |
| Mean systolic pressure gradient (mm Hg), mean ± SD # | 58.2 ± 16.8 | 54.2 ± 16.4 | 61.7 ± 16.7 | <0.001 |
| Aortic valve area (cm 2 ), mean ± SD | 0.8 ± 0.2 | 0.6 ± 0.2 | 0.9 ± 0.3 | 0.004 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
