Impact of Valve Academic Research Consortium 3 (VARC-3) minor access site vascular complications in patients undergoing percutaneous transfemoral transcatheter aortic valve implantation

Abstract

Open in new tabDownload slide

OBJECTIVES

The aim of this study was to investigate the impact of Valve Academic Research Consortium 3 minor access site vascular complications (VCs) in patients who underwent percutaneous transfemoral (TF) transcatheter aortic valve implantation (TAVI).

METHODS

This single-centre retrospective study included consecutive patients who underwent percutaneous TF-TAVI from 2009 to 2021. A propensity score-matched analysis was performed to compare early and long-term clinical results between patients with VC and without VC (nVC).

RESULTS

A total of 2161 patients were included, of whom 284 (13.1%) experienced access site VC. Propensity score analysis allowed to match 270 patients from the VC group with 727 patients from the nVC group. In the matched cohorts, the VC group showed longer operative times (63.5 vs 50.0 min, P < 0.001), higher operative and in-hospital mortality (2.6% vs 0.7%, P = 0.022; and 6.3% vs 3.2%, P = 0.040, respectively), longer hospital length of stay (8 vs 7 days, P = 0.001) and higher rates of blood transfusion (20.4% vs 4.3%, P < 0.001) and infectious complications (8.9% vs 3.8%, P = 0.003). Overall survival during follow-up was significantly lower in the VC group (hazard ratio 1.37, 95% CI 1.03–1.82, P = 0.031) with 5-year survival rates being 58.0% (95% CI 49.5–68.0%) and 70.7% (95% CI 66.2–75.5%) for the VC and nVC groups, respectively.

CONCLUSIONS

This retrospective study observed that minor access site VCs during percutaneous TF-TAVI can be serious events affecting early and long-term outcomes.

INTRODUCTION

As the indication to transcatheter aortic valve implantation (TAVI) is expanding to patients at progressively lower risk [1, 2], the reduction of procedure-related vascular complications (VCs) represents a main objective [3].

Current available clinical trials reported an incidence of VCs ranging from 5% to 18% during TAVI procedures [4–9]. Even if the rate of VCs progressively declined over time, these events are still common and are associated with adverse early clinical outcomes [10, 11] and long-term mortality [12–15]. The Valvular Academic Research Consortium (VARC) is an established independent collaboration between academic research organizations and specialty societies in the USA and Europe with the main objective to create consistent end point definitions and consensus recommendations for implementation in TAVI clinical research programs. The new VARC-3 criteria were recently introduced to better classify the postoperative outcomes of patients who underwent aortic valve replacement procedures [16]. Based on this updated classification, major VCs are associated with increased mortality by definition over minors, and, up to now, no data still exist about the impact of minor VCs on patients treated with transfemoral (TF) TAVI. Supplementary Material, Fig. S1 reports the VARC-3 classification of VCs.

Thus, the aim of the present study was to investigate the early and long-term impact of minor VARC-3 access site VCs in a well-defined population of patients who exclusively underwent percutaneous (TF-TAVI) at our institution.

METHODS

Ethics statement

The local ethics committee (University hospital of Bordeaux, France) approved the study design, consent process and review and analysis of the data (IRB n. CER-BDX 2023-62).

Study oversight

The use of data for scientific and research purposes is included in the formal written informed consent agreements used at our institution. We also guarantee the respect of anonymity, professional secrecy and the use of the collected data and the statistical analysis solely for the scientific purposes granted in accordance with the law in force (GDPR).

Study population study design and rationale

From March 2009 to December 2021, a total of 2896 consecutive patients underwent TAVI at the University Hospital of Bordeaux (France). The primary indication for the procedure included severe aortic valve stenosis in all patients. The strategy for aortic valve replacement, the anatomical access and the type and size of the prosthesis were formally discussed for each patient by the heart team, with the support of computed tomography exams.

The study was designed to analyse patients who underwent TF-TAVI, specifically focused on those deemed ‘suitable’ for percutaneous treatment, and for this reason, patients with TF-TAVI with surgical exposure of femoral vessels were explicitly excluded from the study (Fig. 1). Transcarotid and transapical approaches, as well as patients who experienced major VARC-3 VCs, were excluded from the study. Thus, a total of 2161 (85.5%) patients were considered in this study. Figure 1 shows the study design.

The entire cohort was subsequently divided into 2 groups based on whether they did (group 1: VC) or did not experience minor access site VC (group 2: nVC). Minor VCs were defined accordingly with VARC-3 criteria as percutaneous closure device failure, distal embolization, arterial wall dissection, arterial perforation, rupture, stenosis, ischaemia, arterial or venous thrombosis, artero-venous fistula, pseudoaneurysm and haematoma. The analysis of surgical outcomes was performed by reviewing electronic health records, to investigate early postoperative clinical outcomes and long-term mortality between 2 cohorts.

Transcatheter aortic valve implantation protocol and devices used

All patients were previously discussed by the heart team to evaluate the indications, risks and benefits of the transcatheter procedure. Preoperative total body computed tomography scan was mandatory for all patients to evaluate the eligible criteria for TAVI and to choose the best approach (i.e. TF, transcarotid, transapical). All procedures were performed by 1 interventional cardiologist and 1 cardiac surgeon. The prostheses used during the study period were the following: Edwards SAPIEN valve, SAPIEN-XT, SAPIEN 3 and 3-Ultra System (Edwards Lifesciences, Irvine, CA, USA), Medtronic CoreValve, EVolut R or PRO (Medtronic, Minneapolis, MN, USA), Portico TAVI system (Abbott Vascular, Santa Clara, CA, USA), Accurate Neo Aortic Valve system (Boston Scientific, Marlborough, MA, USA) and Lotus Valve system, Edge (Boston Scientific). Doppler ultrasound was systematically used for femoral puncture and different vascular closure devices (VCD) were used based on the preference, practice and experience of the operators. The VCD used during the study period were the following: Prostar and Proglide (Abbott Vascular, Santa Clara, CA, USA), MANTA (Essential Medical, Malvern, PA, USA), PerQseal^® (Vivasure Medical) and Angio-Seal (St. Jude Medical, St. Paul, MN, USA). In all cases, the use of Angio-Seal was added to the use of Proglide to obtain more effective hemostasis and arterial closure.

At the end of the procedure, an angiographic control of the vascular access was systematically performed, and a compression bandage of the femoral access or eventual additional upstream transient balloon inflation was placed if needed.

Diagnosis and management of vascular complications

The time to onset of VCs was variably reported from the time of the procedure to the end of hospitalization. The most difficult and demanding procedures showing immediate bleeding, femoral artery rupture or arterial dissections were clearly visible in the operating room. In this case, a vascular treatment was immediately performed.

Femoral angiography at the end of the TAVI procedure allowed us to identify arterial ruptures, dissections or other vascular injuries. However, all patients underwent Doppler ultrasound control of the femoral vessels before discharge. The indication to perform a vascular procedure, both surgical and endovascular, has always been taken collectively by the operating team (surgeon, anaesthesiologist, interventional cardiologist). If VC occurred or it was detected during the hospitalization, the heart team discussed which treatment to propose (percutaneous or open surgery).

Statistical analysis

All continuous variables were checked for the normality of their distribution using the Shapiro–Wilk test. For the entire cohort (‘before matching’), continuous variables were reported as mean ± standard deviation or median (interquartile range, IQR) for normally and not-normally distributed variables, respectively, while categorical variables were reported as frequencies and percentages. Patients were stratified based on whether they developed VCs or not and were compared using parametric (Student’s paired t, for normally distributed variables) or nonparametric (Mann–Whitney U; for not-normally distributed variables) tests, as appropriate, for continuous variables and the Chi-squared test for categorical variables. Before performing all analyses, missing data in the baseline dataset were imputed using a random forest model via the R package ‘missForest’ (version 1.4). This algorithm initially imputes all missing data using the mean/mode; then for each variable with missing values, it fits a random forest on the observed part and then predicts the missing part.

Propensity score matching (Fig. 2A) was used to adjust for prespecified baseline characteristics that were potentially confounding variables. We calculated propensity scores using logistic regression models with all baseline variables listed in Fig. 2B. The C-statistic for the model was 0.75. Cases from the VC group were matched 1:3 with cases from the nVC group, using the propensity score with a calliper of 0.1 of the standard deviation of the logit of the propensity score, without replacement [17, 18]. Standardized mean differences (SMD) were determined to compare the baseline characteristics of all patients; an SMD of <0.2 was considered an indicator of a good balance between groups. For the matched cohort (‘After Matching’), data were again presented as described above. Both approaches were compared by paired univariate analysis. Categorical variables were compared using McNemar’s test and continuous variables were compared by Wilcoxon signed-rank test.

All patients were followed up until death or the latest follow-up time. For time-to-event outcomes (overall survival), Kaplan–Meier curves were constructed. Cox proportional hazards models were used to calculate hazard ratios and 95% confidence intervals (CIs) for each of these outcomes. All analyses were completed with R Statistical Software (version 4.0.5, Foundation for Statistical Computing, Vienna, Austria).

RESULTS

Study population and vascular complication characteristics

A total of 2161 consecutive patients (52.9% male, median age 84.0 years) underwent percutaneous TF-TAVI between March 2009 and December 2021. The overall median European System for Cardiac Operative Risk Evaluation II was 4.05% (IQR 2.5–6.69%) and was classified as being low in 8 (0.37%) patients, intermediate in 85 (3.93%) patients and high in 2068 (95.70%) patients. Other demographics and preoperative characteristics are summarized in Table 1.

A total of 284 patients (13.1%) experienced minor access site VCs associated with their percutaneous TF-TAVI procedure. VCs included percutaneous closure device failure in 271 patients, distal embolization in 2 patients, arterial wall dissection in 7 patients, arterial perforation or rupture in 3 patients, and artero-venous fistula in 1 patient. VCs were treated through an endovascular approach in 230 patients (81.9%), including stent implantation in 227 patients, thrombectomy with Fogarty catheter in 2 patients, and balloon dilatation in 1 case. Thirty patients (10.6%) required surgical repair with patch reconstruction or ilio/femoral bypass. The remaining 24 patients (8.5%) were treated with local manual compression.

Before propensity score matching, patients in the VC group were less likely to be male (SMD = −0.331) and to have received Angio-Seal (SMD = −0.567) or Edwards valve implantation (SMD = 0.218) (Table 1). On the other hand, they more frequently underwent procedures with a left-sided approach (SMD = 0.273) and Corevalve valve implantation (SMD = 0.234). Following propensity score matching, 270 patients in the VC group were matched with 727 patients in the nVC group, showing adequate balancing of baseline characteristics (Fig. 2). The characteristics of the matched populations are summarized in Table 1.

In-hospital outcomes

Intraoperative characteristics and postoperative outcomes before and after matching are summarized in Table 2. Before matching, we reported 17 cases of intraprocedural deaths, 8 patients in the VC group and 9 patients in the nVC group. The cause of death was cardiac arrest/not responsive arrhythmia in 14 cases (8 in the VC group and 6 in the nVC group) and valve-related in 3 patients. A total of 57 patients died during the hospitalization period. Ten deaths were due to arrhythmias/conduction disorders, 18 to multi-organ failure, 21 to cardiogenic shock and 8 to septic shock. In the matched cohorts, patients in the VC group showed higher intraoperative (2.59% vs 0.69%, P = 0.022) and in-hospital (6.3% vs 3.16%, P = 0.04) mortality. VC was associated with longer operative times (63.5 vs 50.0 min, P < 0.001), longer hospital length of stay [8 vs 7 days, P = 0.001], higher rates of blood transfusion (20.4% vs 4.26%, P < 0.001) and infectious complications (8.89% vs 3.85%, P = 0.003) than the nVC group. No differences were noted between the 2 groups regarding other in-hospital complications (Table 3).

Long-term outcomes

Figure 3 shows the overall mortality before and after matching in 2 cohorts. Follow-up data were available for a median of 2.7 (IQR 1.1–4.2) years after TF-TAVI. In the matched analyses, overall survival was significantly lower in the VC group (hazard ratio 1.37, 95% CI 1.03–1.82, P = 0.031), with 5-year survival rates being 58.0% (95% CI 49.5–68.0%) and 70.7% (95% CI 66.2–75.5%) for the VC and no VC groups, respectively.

Supplementary Material, Fig. S2 shows the timeline distribution of VCs over the study period. The rate of VCs reported was stable over time ranging from 11.5% to 16.5%.

DISCUSSION

The introduction in clinical practice of VCD has allowed for totally percutaneous TAVI procedures reducing procedural times and complications. Nevertheless, the incidence of VCs following TAVI remains an open issue [19–22]. Although has been well demonstrated that major VCs are associated with adverse outcomes [15], and the difference between major and minor VCs is based on death by definition, little data still exist about early outcomes and the long-term impact of minor VCs in patients treated with percutaneous TF-TAVI.

The main findings of this study are as follows:

1. The rate of minor access site VCs during percutaneous TF-TAVI is 13.1%.

2. Access site VC is a potentially serious event during TF-TAVI procedures, increasing hospital length of stay, the rates of blood transfusion and infectious complications, as well as intraoperative and in-hospital mortality.

3. The long-term overall survival is significantly lower in patients who experienced minor access site VC during TF-TAVI.

Our results showed that minor access site VCs are common events in patients deemed suitable for a percutaneous approach. In our experience, most of VCs reported (95%) are attributable to a failure of the VCD used. This aspect could be related to the vessel’s characteristics (tortuosity and calcifications), device sheath size and type of VCD used. It is well known that pre-existing peripheral arterial disease (PAD) is associated with an increased risk of all-cause mortality and VCs after TAVI procedures [23, 24]. Nevertheless, there is still no consensus and established criteria for PAD definition and its involvement in access site VCs. However, although most of our study population was at high risk (96%), the incidence of PAD was not significantly different between the 2 groups. This could be explained because patients with PAD were submitted to different approaches than percutaneous TF-TAVI (i.e. transcarotid, transapical). A recent meta-analysis by Montalto et al. [25] investigated the outcomes according to the type of VCD used during TAVI procedures. The authors found that ProGlide sutured closure system (Abbott) and MANTA (Teleflex, Morrisville, NC, USA) devices appear to be valid options to minimize VCs in patients undergoing TAVI. In agreement with these findings, our study showed no significant differences between the use of all VCD considered, highlighting that all devices could represent an acceptable alternative for vascular closure during TF approaches. Moreover, we found that the addition of Angio-Seal^® (Terumo Interventional Systems) to ProGlide to complete the haemostasis was effective in reducing the rate of VCs.

The present study permits some more considerations.

First, the higher operative and in-hospital mortality in the VCs group lead us to speculate that if a minor access site VC has already occurred during the procedure, and any type of other complication is superimposed, the operative management becomes more complicated reducing the chances of survival following percutaneous TF-TAVI procedure. In addition, intraoperative blood transfusions and potential transient haemodynamic instability episodes during the acute period of VC may further increase the mortality risk.

Second, previous studies [12, 13, 14, 15, 26] have demonstrated that VCs are associated with worse short- and 1-year clinical outcomes including all-cause mortality. However, these trials have reported short follow-up times and have not offered separate analyses on the impact of minor VCs.

The main objective of the present study was to investigate the long-term impact of access site VCs that were not directly responsible for the death of the patients. In this regard, although the disadvantages of VCs in the early outcomes are clear, the reasons behind the late outcomes are still unexplored. The reasons that led to this high incidence of mortality in the VC group during the FU period are certainly multifactorial and probably occur in more frail patients, with characteristics that we are not yet fully able to screen for during procedural planning.

Third, our findings suggest that not only major VCs pose a serious risk but also minor access site VCs can lead to serious consequences that can impact patient survival. In this context, timely and effective management must be ensured by operators who can offer both surgical and endovascular solutions.

Last, in contrast with previous experiences that report a progressive reduction over time, of VCs, our findings show a stable rate of VCs during the study period. Although many factors should be considered, such as the improvement of devices with smaller delivery systems, better VCDs or the systematic use of Doppler ultrasound for femoral vessel puncture, the incidence of VCs remains somewhat constant, and the learning curve was not able to reduce the incidence. This evidence reveals that minor access site VCs are a surely underestimated and frequent problem during our daily transcatheter procedures that should be considered with attention during procedural planning.

Limitations

The main limitation of the study is its retrospective rather than randomized nature. Moreover, this being a single-centre study can represent an important limitation that makes the generalization of the results more difficult. In this regard, there could be some selection bias during the retrospective assessment of VARC-3 criteria and some data could be missed especially in the initial period of our experience with TF-TAVI. Two other important limitations concern the lack of standardized criteria to evaluate the preoperative access site vascular risk and the lack of data regarding the causes of mortality during the follow-up period. In addition, the extended study period ranging from 2009 to 2021 may introduce relevant bias considering the device and technique modifications over time.

CONCLUSION

In conclusion, our findings suggest that minor VCs should be evaluated with care due to their strong impact on short- and long-term outcomes.

More specific criteria are needed to standardize VC risk for each patient, based not only on vessel characteristics but also on patients’ clinical and functional status. In addition, cardiac surgeons together with interventional cardiologists represent an essential combination during transcatheter procedures, combining endovascular and surgical skills to ensure timely management and treatment of VCs.

SUPPLEMENTARY MATERIAL

Supplementary material is available at EJCTS online.

Funding

No funding source was provided for this study.

Conflict of interest: Thomas Modine has served as a consultant for Medtronic, Abbott, Microport and Edwards Lifesciences. All other authors have no conflicts of interest or financial conflicts to disclose.

DATA AVAILABILITY

The data can be shared on reasonable request to the corresponding author.

Author contributions

Antonio Piperata: Conceptualization; Data curation; Writing—original draft. Jef Van den Eynde: Formal analysis. Mathieu Pernot: Data curation; Project administration.Martina Avesani: Conceptualization; Supervision; Writing—review & editing. Benjamin Seguy: Data curation; Project administration. Guillaume Bonnet: Data curation; Project administration. Walid Ben Ali: Supervision. Lionel Leroux: Data curation; Project administration; Supervision. Louis Labrousse: Project administration; Supervision. Thomas Modine: Methodology; Project administration; Supervision.

Reviewer information

European Journal of Cardio-Thoracic Surgery thanks Roman Gottardi, Ferdinand Aurel Vogt and Mario Lescan for their contribution to the peer review process of this article.

REFERENCES

ABBREVIATIONS

ABBREVIATIONS
 
• CIs

  Confidence intervals

 
• IQR

  Interquartile range

 
• PAD

  Peripheral arterial disease

 
• SMD

  Standardized mean differences

 
• TAVI

  Transcatheter aortic valve implantation

 
• TF

  Transfemoral

 
• VARC-3

  Valvular Academic Research Consortium 3

 
• VCs

  Vascular complications

 
• VCD

  Vascular closure device

Author notes