Comparison of Transfemoral versus Transsubclavian/Transaxillary access for transcatheter aortic valve replacement: A systematic review and meta-analysis

Femoral access is the gold standard for transcatheter aortic valve replacement (TAVR). Safe alternative access, that represents about 15 % of TAVR cases, remains important for patients without adequate transfemoral access. We aimed to perform a systematic review and meta-analysis of studies comparing transfemoral (TF) access versus transsubclavian or transaxillary (TSc/TAx) access in patients undergoing TAVR. We searched PubMed, Cochrane CENTRAL Register, EMBASE, Web of Science, Google Scholar and ClinicalTrials.gov (inception through May 24, 2022) for studies comparing (TF) to (TSc/TAx) access for TAVR. A total of 21 studies with 75,995 unique patients who underwent TAVR (73,203 transfemoral and 2,792 TSc/TAx) were included in the analysis. There was no difference in the risk of in-hospital and 30-day all-cause mortality between the two groups (RR 0.64, 95 % CI 0.36–1.13, P = 0.12) and (RR 0.95, 95 % CI 0.64–1.41, P = 0.81), while 1-year mortality was significantly lower in the TF TAVR group (RR 0.79, 95 % CI 0.67–0.93, P = 0.005). No significant differences in major bleeding (RR 0.82, 95 % CI 0.65–1.03, P = 0.09), major vascular complications (RR 1.14, 95 % CI 0.75–1.72, P = 0.53), and stroke (RR 0.66, 95 % CI 0.42–1.02, P = 0.06) were observed. In patients undergoing TAVR, TF access is associated with significantly lower 1-year mortality compared to TSc/TAx access without differences in major bleeding, major vascular complications and stroke. While TF is the preferred approach for TAVR, TSc/TAx is a safe alternative approach. Future studies should confirm these findings, preferably in a randomized setting.


Introduction
Aortic valve replacement for symptomatic severe aortic stenosis (AS) has class I indication in both the current guidelines. [1,2] Transcatheter aortic valve replacement (TAVR) has been approved for aortic valve replacement in high-, intermediate-or low-risk patients with symptomatic severe AS [3][4][5][6] becoming a predominant therapy for the treatment of severe AS, exceeding surgical aortic valve replacement in the US since 2019. [7] As delivery systems have evolved, corresponding sheath sizes have also become smaller to facilitate greater rates of transfemoral (TF) TAVR. While TF access remains the preferred access route for TAVR, [8] 10-15 % of cases are unsuitable for TF access. [9] TF route allows a fully-percutaneous TAVR under conscious sedation/local anesthesia. Careful procedural planning by CT scan and accurate choice of the proper site for vascular puncture are keys for procedural success. Analysis of CT scan images will help to identify potential challenges such as tortuosity, presence of aneurysms, thrombotic appositions, or aortic arch calcifications. All these anatomic features are potential sources of complications when large catheters are inserted and, therefore, can be considered as relative contraindications for a transfemoral approach. When TF access is contraindicated, an alternate access like trans subclavian or transaxillary or trans carotid or transaortic can be considered. Due to unfavorable outcomes associated with transapical and transaortic access, [10,11] other alternative access routes have been developed including transsubclavian (TSc) and transaxillary (TAx). [12][13][14] While preferences for alternative access TAVR approaches vary and depend on operator preference, institutional experience and patient anatomy, alternative access site choice is critical.
Important TAVR outcomes include access-related complications like pseudoaneurysm or bleeding. [15,16] Comparing outcomes between non-TF access versus TF access is clinically important in defining outcomes associated with alternative access. While evidence shows advantages and disadvantages of each access routes, due to lack of head-tohead randomized comparator trials, appropriate access choice remains a debatable/controversial issue. We, therefore, aimed to review all studies comparing TF to TSc/TAx accesses in regards to their safety and efficacy endpoints through a systematic review and meta-analysis.

Data sources and search strategy
A meta-analysis was performed according to the Preferred Reporting Items for Systematic Reviews on meta-Analyses (PRISMA) 2015 guidelines. [17] Two reviewers (WA, MI) independently identified the relevant studies by an electronic search of the PubMed, EMBASE, Cochrane Central Register of Controlled Trials, and ClinicalTrials.gov databases (from inception to May 2022). Reference lists of the retrieved studies were also screened further for relevant studies. The following search terms and key words were used: "aortic stenosis" and "transcatheter aortic valve replacement" or "TAVR" or "TAVI" and "subclavian artery access" or "axillary artery access" or "femoral artery access". The metaanalysis was registered in the PROSPERO (International Prospective Register of Systematic Reviews) Registry, under PROSPERO CRD42022340351.

Study selection
Two reviewers (WA, MI) independently assessed studies' eligibility based on titles, abstracts, and full-text reports. Discrepancies in study selection were discussed and resolved with a third investigator (KD). Eligible studies had to compare between transsubclavian and/or transaxillary vs transfemoral access for TAVR, and present clinical outcomes data of interest. Exclusion criteria were: (a) lack of any clinical outcome data, (b) single arm studies, (c) duplicate publications, (d) reviews, editorials, letters, and non-human studies. Only studies published in the English language were included in this meta-analysis.

Data extraction and quality assessment
Two investigators (WA, AMB) independently extracted data (baseline characteristics, outcomes and number of events) using a standardized data abstraction form. Funnel plots for the outcomes were used to assess for publication bias when data were available for at least three studies (Supplemental Figs. 1, 2). The studies' methodological quality was assessed systematically using the Newcastle-Ottawa Scale for observational studies (Supplemental Table 2), and disagreements were resolved by a third author (KD).

Outcome measures
The co-primary outcomes for study selection were in-hospital, 30day, and 1-year all-cause mortality, major vascular complication, major bleeding and stroke. Peri-procedural myocardial infarction (MI), cardiac tamponade, pacemaker placement, conversion to open surgery, acute kidney injury (AKI), procedure success, procedure time and fluoroscopy time were secondary outcomes. Outcome definitions were as determined in each individual study. Most outcomes were assessed according to Valve Academic Research Consortium (VARC) definitions. [18].

Statistical analysis
For dichotomous outcomes, risk ratios (RRs) with 95 % confidence intervals (CIs) were calculated from the available data, and trial-specific RRs were combined using the DerSimonian and Laird random effects model with the estimate of heterogeneity taken from the Mantel-Haenszel model. We used I 2 statistics to measure heterogeneity among the included trials. A value of 0 % indicated no observed heterogeneity, and I 2 values of 25 %, 50 %, and 75 % were considered to represent low, moderate, and high heterogeneity, respectively. Analyses were performed using Review Manager (RevMan) Version 5.3 (The Nordic Cochrane Centre, The Cochrane Collaboration, Copenhagen,

Discussion
Our analysis of 21 studies including more than 75,000 patients showed: (1) There was no significant difference in in-hospital, and 30day mortality between patients undergoing TF vs TSc/TAx TAVR, while 1-year mortality was lower in the TF group. (2) There were no significant differences between the two groups in the risks of major bleeding and major vascular complications. (3) Rate of pacemaker placement were significantly less in the TF access group. (4) Stroke did not differ significantly between the groups. (5) Cardiac tamponade, periprocedural MI, and AKI, did not differ significantly between the groups. (6) The procedure time was noted to be lower in the TF group, with no significant difference in fluoroscopy time.

Mortality
Ruge and colleagues [37] were the first to describe successful TSc/ TAx-TAVR in a patient with aortoiliac occlusive disease and concomitant left subclavian arterial stenosis (the right SCA was accessed). Use of the left SCA was later described by Asgar and colleagues [38] in 2009 after they treated a woman with severe aortic stenosis and very small iliofemoral arteries. Data from previous studies support the TSc/TAx as the preferred non-TF route due to several advantages. [34,39,40] Gleason et al [27] compared a cohort of TSc patients to TF patients within the CoreValve US Pivotal Trial and Continued Access Study, and reported TSc patients, demonstrated no significant differences in outcomes, with 30-day and 1-year mortality rates equivalent to TF procedures. That aligns with the results of our pooled analysis with respect to 30-day mortality, while contrasting with our study with respect to 1 year mortality, which was noted to be higher in patients with TSc/TAx approach. A possible explanation may relate to the higher rate of comorbidities seen in TSc and TAx groups than the TAVI procedure itself. TSc/TAx patients have increased risk, reflected by the higher Logistic EuroSCORE as seen in table 2, which may explain the worse late survival. Iliofemoral disease is the most common reason that makes iliofemoral access undesirable. Peripheral vascular disease (PVD)   especially iliofemoral disease is frequently seen in patient being referred for TSc/TAx TAVR with prevalence ranging from 43 % to 60 % and. [41] It is a well-known fact that PAD is an independent predictor of long term mortality and stroke in these patients. [41,42] Table 2 shows higher prevalence of baseline PVD in patients undergoing TSc/TAx -TAVR than TF approach which is in accordance with prior studies.

Bleeding and vascular complications
Percutaneous or surgical cut down access for TSc/TAx TAVR were used in the included studies. However, major vascular complication rates appear not to be significantly different between the two groups. This supports the notion that the TSc/TAx approach may be the preferred alternative-access option in the current era of newergeneration devices. Pooled results deriving from unadjusted data in our meta-analysis found no difference in the risk of major bleeding in both groups although was a non-significant trend towards decrease in the risk of major bleeding the TF TAVR group. This is in line with a previous analysis [43] that used adjusted data and showed no statistically difference in the risk of bleeding during transcarotid/transsubclavian TAVR in comparison with transfemoral TAVR. Another report from the FRANCE registry who grouped TC and subclavian/ axillary TAVR (1,616 patients) reported similar outcomes compared to TF in the term of major bleeding. [44] In our meta-analysis we compared transsubclavian/transaxillary TAVR without transcarotid (TC) TAVR group to the femoral TAVR.

Pacemaker rates
Conduction dysfunction originating from the mechanical injury due to the anatomical interaction between the valve prosthesis and the atrioventricular node and bundle of His are the implicated causes requiring pacemaker implantation. [45] In our metanalysis, we found that the rate of pacemaker placement was surprisingly lower in TF approach compared to TSc/TAx approach. Moreover, there was no differences in procedural complications such as MI, AKI, and cardiac tamponade. These findings are consistent with Italian CoreValve Registry data that showed comparable procedural and 2-year results after TSc and TF approaches. [40] Our meta-analysis showed that the TSc/TAx approach had a longer procedure time when compared to TF approach, which exposes the AV node area to longer manipulation time that could lead to AV nodal dysfunction.

Stroke
Previous studies have reported conflicting data in regards to periprocedural stroke events. Dahle et al reported higher stroke rates in TAx approach, which may be partially related to an increased risk of access site trauma with TSc/TAx approach. [38,46] providing nidus for thrombus formation with subsequent embolization or embolization of an atheromatous plaque located in the subclavian artery. This makes potentially relevant the use of embolic protection during TSc/TAx approach. [47]. However, other included studies as well as our aggregate analysis reported no significant difference in post-procedural stroke with either access.

Procedure and fluoroscopy time
The procedure time was reported in 6 studies and fluoroscopy time in 7 studies. Petronio et al. [40] found that the overall procedural time was longer in the TSc group compared to the TF (120 vs 75 min, p < 0.0001), however the fluoroscopy time was similar (18 vs 21 min, p = 0.15). Muensterer et al. [13] also failed to demonstrate a significant difference in fluoroscopy time between the TSc and TF groups (22.24 vs 25.48 min, p = 0.053) which is aligned with our pooled analysis that showed comparable fluoroscopy time in both groups, however the procedural time was significantly longer in the TSc group (105 vs 82 min, p = 0.001). Dahle et al. [28] found that the mean total fluoroscopy time and procedure time were slightly longer in the TSc group compared to TF groups (21.7 vs 17.7 min and 137.6 vs 97.7 min, respectively). Our meta- analysis showed no difference in the fluoroscopy time while procedure time was shorter in the TF TAVR group. Since the fluoroscopy time is similar in both groups our findings could be explained by the longer surgical vascular access and wound closure required in the TSc/TAx group, in addition to that fact that most operators will have potentially more expertise with the TF approach.
A previous meta-analysis by Zhan et al. [48] of 5 studies comparing the TF and TSc/TAx and another network meta-analysis [49] comparing several access sites demonstrated a lower but statistically nonsignificant 1-year mortality with the TF group, while our study showed statistically significant lower 1 year mortality in the TF group. Our present meta-analysis included 21 studies with subclavian/transaxillary and 75,995 patients. We also found a lower rate of new pacemaker placement in TF group compared TSc/TAx, which is different compared to the previous meta-analyses who reported comparable risk of new pacemaker between the two groups. Furthermore, we evaluated other outcomes such as procedure time and fluoroscopy time, which were not part of the outcomes of interest in the prior meta-analysis, thus, our meta-analysis adds methodological rigor and novel findings to the literature.

Limitations
Our study has several important limitations. In this study, the data analyzed were from observational studies and not randomized trials comparing TF and TSc/TAx access. There is intrinsic heterogeneity between different studies in terms of representation of baseline data, study design, and outcome measures. Only one study included in this metaanalysis was propensity-matched with similar patient demographics, other studies exhibited major differences in baseline characteristics between the TF and TSc/TAx cohorts. There is a possibility of publication bias among the outcomes where significant asymmetry was observed. Moreover, data included in our analysis represents a conglomerate of both self-expandable and balloon-expandable prostheses making it unattainable to carry out a head to head comparison between such devices. The data of pre-dilation or direct implantation were not available in the studies as well. Lastly, the quality of this metaanalysis reflects the quality of individual studies. Nevertheless, our metaanalysis is strengthened by inclusion of a large number of real world studies (total 21) and therefore, is the most current and comprehensive meta-analysis on this important clinical issue.

Conclusion
In patients undergoing TAVR, TF access is associated with significantly lower 1-year mortality compared to TSc/TAx access, while there were no differences in major vascular complications, major bleeding or stroke. While TF is the preferred approach for TAVR, TSc/TAx appears to be a safe alternative approach. Future studies should confirm these findings, preferably in a randomized setting.

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.