Vascular complications (VC) after transcatheter aortic valve implantation (TAVI) are reported using various criteria and several access site approaches. We aimed to describe them in a solely percutaneous transfemoral TAVI approach and their association with survival using both the updated Valve Academic Research Consortium (VARC)-2 criteria and the former VARC-1 criteria. From March 2009 to September 2013, 403 consecutive patients at a mean age (±SD) of 83 ± 6 years underwent percutaneous transfemoral TAVI. VC were defined by both VARC-1 and VARC-2 criteria and analyzed separately. Cox proportional hazard ratio models for all-cause mortality were adjusted separately as defined by each criteria. VARC-1–defined and VARC-2–defined VC occurred in 71 (18%) and 78 (19%) patients, respectively, with 15 (4%) and 33 (8%) defined as major VC. The difference in frequency of major and minor VC was mainly driven by VARC-2 implementation of major bleeding events. With either VARC definition, patients with minor VC had similar mortality and complications rates as those patients without VC. In multivariate analyses, referenced to patients with minor or no VC, only VARC-1–defined major VC were significantly associated with increased mortality (hazard ratio 3.52; confidence interval 1.5 to 8.4; p = 0.005), whereas VARC-2–defined major VC were found to be only marginally significant (hazard ratio 1.9; confidence interval 0.9 to 3.9; p = 0.08). In conclusion, the implementation of the VARC-2 criteria resulted in a higher rate of reported major VC after TAVI compared with VARC-1 criteria, mainly by the inclusion of major bleeding events and a reduced association with patient mortality.
Highlights
- •
Vascular complications after transcatheter aortic valve implantation (TAVI) are reported using various criteria.
- •
We compared them using 2 Valve Academic Research Consortium criteria (VARC).
- •
The work analyzed data of 403 patients who underwent percutaneous transfemoral TAVI.
- •
VARC-2 criteria resulted in a higher rate of reported major vascular complications after TAVI.
- •
This difference was mainly driven by VARC-2 implementation of major bleeding events.
Vascular complications (VC) after transcatheter aortic valve implantation (TAVI) have been associated with a significant increased morbidity and mortality. Expert-led attempts were made to standardize the classification of VC associated with TAVI using the end point definitions of the “Valve Academic Research Consortium” (VARC) to select patient outcomes that would best reflect the safety and efficacy of TAVI. Despite this, large variability still exists regarding the reported “real-life” occurrence of major and minor VC ( Supplementary Table 1 ). Furthermore, many of the previous reports included a mixture of different anatomical access sites and transcatheter approaches and various definitions for VC. In the present analysis, we sought to compare the association of VC defined by VARC-1 and VARC-2 with patient survival. We report here our single-center experience with respect to the prevalence and clinical risk predictors of VC during a solely full percutaneous transfemoral TAVI approach and compare the association of the 2 VARC definitions with patient mortality.
Methods
From March 2009 to September 2013, 410 consecutive patients underwent percutaneous aortic valve implantation at our medical center. Informed consent was obtained from each patient as approved by the Institutional Ethics Committee. Patients were recruited as part of their participation in the Tel-Aviv Prospective Angiography Study. The diagnosis of severe AS was based on clinical, echocardiographic, and hemodynamic criteria. Suitability and eligibility for TAVI were determined by a joint team including an interventional cardiologist, a cardiac surgeon, and a senior echocardiographist. According to recent guidelines, transfemoral TAVI is contraindicated in patients with inadequate femoral vascular access because of significantly reduced vessel size, severe calcification, or severe tortuosity. Therefore, during the preprocedural patient evaluation, we referred 7 patients to either a transaxillary or a transapical approach, all of whom had severe femoral artery calcifications, extremely tortuous arteries, or femoral artery diameters smaller than 6 mm. Mild femoral artery tortuosity, however, was not a contraindication for a transfemoral approach.
We have previously reported the methodologic aspects of the TAVI procedure. Elective preprocedural angiography was performed in all patients. After contralateral arterial puncture using the Seldinger technique, either a pigtail or IMA catheter were introduced as a landmark for femoral artery puncture and sheath delivery. All procedures were performed using either 18F sheaths (Corevalve system, Medtronic, MN; Cook medical, IN; St. Jude, MN) or 18F/19F sheaths (Edwards Sapien or Edwards Sapien XT System, Edwards Lifesciences, CA). At the end of the procedure, femoral artery access was closed using a Prostar XL closure device (Abbott Vascular, Abbott Park, Illinois) followed by angiography of the femoral artery through the contralateral femoral artery to detect VC.
The original consensus report of the VARC-1 standardized the end point definitions of TAVI procedures including the occurrence of VC. During 2012, these end points were updated by the VARC—known as VARC-2. Although the general definitions of VC were unchanged after the update, VARC-2 incorporated a more rigorous approach to bleeding and a decrease in hemoglobin (Hb) after TAVI, stating that major bleeding definitions (an Hb decrease >3 g/dL or transfusion of 2 or more packed red blood cells [RBC] units) should be considered also as a major VC, whereas in the former VARC-1 end points, only transfusion of ≥4-pack RBC units was considered a major VC. In addition, VARC-2 recommended that all VC be recorded as either access or nonaccess site related. Importantly, in both consensus documents, interventional or surgical repair for failed percutaneous closure during the initial procedure without other clinical consequence was considered a minor VC. Accordingly, we classified VC using both definitions ( Supplementary Table 2 ).
All data are displayed as mean (±standard deviation) for continuous variables and as the number (percentage) of patients in each group for categorical variables. Significance for continuous variables was determined with the Student’s t test. Significance for categorical variables was determined with the chi-square and Fisher’s exact tests. Logistic regression models were VC where the dependent variables were adjusted to age, gender, height, diabetes mellitus, hyperlipidemia, hypertension, peripheral vascular disease, society of thoracic surgeons (STS) score, aortic valve area, preprocedural mean pressure gradient, and vessel diameter. The log-rank (Mantel-Cox) test was used to estimate the significance in Kaplan–Meier survival plots. Cox proportional hazard models for all-cause mortality were adjusted separately for major VC defined by VARC-1 and VARC-2 and to other predictors of mortality after TAVI : gender, age, body mass index, systolic heart failure, previous coronary artery bypass grafting and percutaneous coronary intervention, atrial fibrillation, pulmonary disease, STS score, preoperative mean gradient, and creatinine clearance test. The proportional hazards assumption was evaluated with the “log minus log” plots. All the analyses were considered significant at a 2-tailed p value <0.05. The SPSS statistical package was used to perform all statistical evaluation (SPSS, Chicago, IL).
Results
The final study population included 403 patients (166 men, 41%) at a mean age and range of 83 ± 6 (±SD) and 61 to 98 years, respectively ( Table 1 ). The Edwards Sapien XT and Medtronic CoreValve prostheses were implanted in 96 (24%) and 306 (76%) patients, respectively. One prosthetic valve was not implanted because of aortic trauma (see later). Valve sizes for the Edwards Sapien XT and Medtronic CoreValve prostheses were 23, 26, and 29 mm in 42, 51, and 3 patients and 23, 26, 29, and 31 mm in 9, 172, 123, and 2 patients, respectively. Most patients underwent the procedure through the right femoral artery (n = 277; 68.7%). Mean access site artery diameter was 7.7 ± 1.3 mm.
| Variable | Total Population (n=403) |
|---|---|
| Age, years (mean±SD, range) (years) | 83±6, 61-98 |
| Men | 166 (41%) |
| Diabetes Mellitus | 138 (34%) |
| Dyslipidemia ∗ | 313 (78%) |
| Hypertension ∗ | 355 (88%) |
| Smoker or past smoker | 98 (24%) |
| Height (mean±SD) (cm) | 163±9 |
| Weight (mean±SD) (Kg) | 73±15 |
| Body mass index (mean±SD) (Kg/m 2 ) | 27±5 |
| Creatinine clearance test (mean±SD) (ml/min) | 50±19 |
| Peripheral vascular disease | 28 (7%) |
| Stroke | 41 (10%) |
| Systolic heart failure | 68 (17%) |
| Coronary artery disease ∗ | 237 (59%) |
| Prior myocardial infarction | 70 (17%) |
| Prior coronary artery bypass grafting | 70 (17%) |
| Left main disease | 42 (10%) |
| Prior cardiac valve surgery | 8 (2%) |
| Permanent pacemaker | 43 (11%) |
| Atrial fibrillation (any type) | 122 (30%) |
| Asthma or chronic obstructive pulmonary disease | 77 (19%) |
| Ejection fraction (mean±SD, range) (%) | 56±8, 25-69 |
| STS score (mean±SD, range) | 4.3±3, 1-25 |
| Euroscore (mean±SD, range) | 24±14, 2-73 |
| Aortic valve area (mean±SD) (cm 2 ) | 0.7±0.19 |
| Peak pressure gradient (mean±SD) (mmHg) | 78±23 |
| Mean pressure gradient (mean±SD) (mmHg) | 47±15 |
∗ Dyslipidemia and hypertension were defined by patient’s history or current use of relevant medications. Coronary artery disease was defined by the presence of >50% stenosis in large epicardial vessel as demonstrated in preprocedural coronary angiography.
VC according to VARC-1 were encountered in 71 patients (18%), with only 15 events (4%) classified as major VC. When applying the VARC-2 definitions, VC occurred in 78 (19%) patients, and the rate of major and minor VC was 8% (33 patients) and 11% (45 patients), respectively ( Supplementary Table 2 ). Femoral artery perforations and local hematomas were the most prevalent VC observed and were commonly treated by immediate prolonged balloon inflation during the index procedure ( Supplementary Table 2 ). Stenosis of the femoral artery occurred in 11 patients and was treated successfully in all during the index procedure by balloon inflation. Dissection, in contrast, occurred in 5 patients and was treated with balloon angioplasty. A single patient with a femoral artery pseudoaneurysm required surgical repair, whereas all others were treated successfully by ultrasound guided thrombin injections during their hospital stay.
A nonsignificant trend was noticed toward increased VC in the early experience groups defined as either the first 50 patients versus the rest (VARC-1 VC: 24% vs 17%, VARC-2 VC: 24% vs 19%, respectively) or the first 100 patients versus the rest (VARC-1 VC: 23% vs 16%, VARC-2 VC: 23% vs 18%, respectively). A nonsignificant trend was also noticed for increased VC in patients implanted with the Edwards Sapien XT compared with the Medtronic CoreValve prosthesis (VARC-1 VC: 23% vs 16%, VARC-2 VC: 24% vs 18%, respectively). In univariate analysis models, dyslipidemia and shorter stature were significantly associated with increased risk for VARC-2–defined VC, with female gender reaching marginal significance (p = 0.051). In the multivariable logistic regression model, the only patient variable found to predict the occurrence of VC was dyslipidemia (p = 0.03), however not shorter stature (p = 0.081). Interestingly, diabetes was found to have a protective effect against VC (p = 0.003).
The association of VC with all-cause mortality and procedural outcomes is presented in Figures 1 and 2 and Tables 2 and 3 . With either VARC definition, patients with minor VC had similar mortality rates during the entire follow-up as those patients without VC ( Table 2 ). Similar complications rates were also noted between these 2 groups. Compared with patients with no VC, patients with VARC-2–defined major VC had increased 30-day mortality and a trend toward higher in-hospital and 6-month mortality and increased rates of respiratory failure and cardiogenic shock ( Table 2 ). A stronger association with mortality of up to 1-year follow-up was noted for VARC-1–defined VC. We further analyzed the survival of patients with major VC compared with those with no or minor VC. When using VARC-2 definitions, a small trend toward an elevated mortality was observed in the early postprocedural period ( Figure 1 ) in patients with major VC, only to equalize later during follow-up (p log-rank test = 0.443). When using VARC-1 definitions, the enhanced mortality observed early after the procedure remained significant throughout the follow-up (p log-rank test = 0.046). In Cox proportional hazard models ( Table 3 and Figure 2 ), only VARC-1–defined VC were significantly associated with increased mortality (hazard ratio 3.52; confidence interval 1.5 to 8.4; p = 0.005), whereas VARC-2–defined VC were found to be only marginally significant (hazard ratio 1.9; confidence interval 0.9 to 3.9; p = 0.08).
| A. | ||||||
|---|---|---|---|---|---|---|
| Valve Academic Research Consortium 1 defined vascular complications | Vascular complications | p value Minor vs. No VC | p value Major vs. No VC | p value Major or Minor VC vs. No VC | ||
| None (n=332) | Minor (n=56) | Major (n=15) | ||||
| Days in hospital (mean±SD) | 8.2±5.9 | 8.6±6.6 | 13.8±15 | 0.545 | 0.01 | 0.160 |
| Mortality (months) | ||||||
| In hospital | 8 (2%) | 1 (2%) | 3 (20%) | >0.999 | 0.009 | 0.237 |
| 1 | 8 (2%) | 1 (2%) | 4 (27%) | >0.999 | 0.001 | 0.060 |
| 6 | 25 (8%) | 4 (9%) | 5 (33%) | >0.999 | 0.009 | 0.153 |
| 12 | 40 (16%) | 5 (14%) | 6 (46%) | 0.813 | 0.014 | 0.306 |
| All mortality | 64 (19%) | 10 (18%) | 6 (40%) | >0.999 | 0.091 | 0.516 |
| Post procedural complications : | ||||||
| Myocardial infarction | 0 (0%) | 0 (0%) | 0 (0%) | >0.999 | >0.999 | >0.999 |
| Cardiogenic shock | 6 (2%) | 0 (0%) | 3 (20%) | 0.599 | 0.005 | 0.199 |
| Respiratory failure | 18 (5%) | 1 (2%) | 7 (47%) | 0.333 | 0.001 | 0.104 |
| Stroke | 5 (2%) | 1 (2%) | 0 (0%) | >0.999 | >0.999 | >0.999 |
| Pacemaker implantation | 78 (24%) | 7 (13%) | 2 (14%) | 0.080 | 0.535 | 0.056 |
| Acute kidney injury | 39 (12%) | 7 (13%) | 4 (27%) | 0.826 | 0.102 | 0.428 |
| In hospital sepsis | 6 (2%) | 2 (4%) | 1 (7%) | 0.325 | 0.268 | 0.199 |
| B. | ||||||
|---|---|---|---|---|---|---|
| Valve Academic Research Consortium 2 defined vascular complications | Vascular complications | p value Minor vs. No VC | p value Major vs. No VC | p value Major or Minor VC vs. No VC | ||
| None (n=325) | Minor (n=45) | Major (n=33) | ||||
| Days in hospital (mean±SD) | 8.2±5.9 | 7.4±2.6 | 12.2±12 | 0.388 | 0.08 | 0.123 |
| Mortality (months) | ||||||
| In hospital | 8 (2%) | 1 (2%) | 3 (9%) | >0.999 | 0.071 | 0.258 |
| 1 | 8 (2%) | 1 (2%) | 4 (12%) | >0.999 | 0.018 | 0.143 |
| 6 | 25 (9%) | 3 (8%) | 6 (19%) | >0.999 | 0.099 | 0.253 |
| 12 | 40 (16%) | 4 (14%) | 7 (26%) | >0.999 | 0.279 | 0.552 |
| All mortality | 64 (20%) | 7 (16%) | 9 (27%) | 0.686 | 0.363 | 0.875 |
| Post procedural complications : | ||||||
| Myocardial infarction | 0 (0%) | 0 (0%) | 0 (0%) | >0.999 | >0.999 | >0.999 |
| Cardiogenic shock | 6 (2%) | 0 (0%) | 3 (9%) | >0.999 | 0.041 | 0.385 |
| Respiratory failure | 18 (6%) | 0 (0%) | 8 (24%) | 0.145 | 0.001 | 0.130 |
| Stroke | 5 (2%) | 1 (2%) | 0 (0%) | 0.543 | >0.999 | >0.999 |
| Pacemaker implantation | 77 (24%) | 5 (11%) | 5 (15%) | 0.058 | 0.384 | 0.045 |
| Acute kidney injury | 39 (12%) | 6 (13%) | 5 (15%) | 0.808 | 0.581 | 0.702 |
| In hospital sepsis | 6 (2%) | 1 (2%) | 2 (6%) | 0.600 | 0.162 | 0.385 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
