Transcatheter Aortic Valve Implantation With or Without Predilation: A Meta-Analysis

Aims. To evaluate the impact of systematic predilation with balloon aortic valvuloplasty (BAV) on transcatheter aortic valve implantation (TAVI). Methods and Results. We performed a systematic meta-analysis investigating patients undergoing TAVI with systematic BAV vs no BAV in RCT or in adjusted studies. Device success was the primary endpoint, while all-cause mortality, 30-day moderate/severe aortic regurgitation (AR), stroke, permanent pacemaker implantation (PPI) and acute kidney injury (AKI) were the secondary endpoints. Subanalysis according to design of the study (RCT and adjusted analysis) and to the type of valve (balloon-expandable [BE] vs self-expanding [SE]) were conducted. We obtained data from 15 studies, comprising 16,408 patients: 10,364 undergoing BAV prior to TAVI and 6,044 in which direct TAVI has been performed. At 30-day follow-up, BAV did not improve the rate of device success in the overall population (OR, 1.09; 95% CI, 0.90-1.31), both in SE (OR, 0.93; 95% CI, 0.60-1.45) and in BE (OR, 1.16; 95% CI, 0.88-1.52) valves. Between BAV and direct TAVI, no differences in secondary outcomes were observed neither in overall population nor according to valve type between BAV and direct TAVI strategies. All endpoints results were consistent between RCTs and adjusted studies except for postdilation rate that did not differ in observational studies (OR, 0.70; 95% CI, 0.47-1.04), while it was lower in BAV when only RCTs were included in the analysis (OR, 0.48; 95% CI, 0.24-0.97). Conclusions. Direct TAVI is feasible and safe compared to predilation approach with similar device success rates and clinical outcomes. Direct TAVI could represent a first-choice approach in contemporary TAVI procedures.

J INVASIVE CARDIOL 2022;34(2):E104-E113. Epub 2022 January 6.

Key words: aortic stenosis, balloon aortic valvuloplasty, transcatheter aortic valve implantation

Transcatheter aortic valve implantation (TAVI) is rapidly emerging as the default strategy for patients with severe aortic stenosis (AS) independently from their surgical risk due to promising results in recent RCTs (randomized controlled trials)^1,2 and due to reassuring long-term outcomes in high-risk patients.^3-5

Consequently, attention is increasingly focused towards technological and clinical strategies to simplify the procedure itself, while maintaining clinical safety and efficacy. Contrasting data have been published about the efficacy and effective need of BAV (balloon aortic valvuloplasty) before TAVI. Historically, BAV was considered a mandatory step to facilitate implantation of the device, to reduce the radial counterforce for optimal device expansion and predict possible coronary obstruction by native valve leaflets.⁶ However, BAV has been shown to be associated with specific complications, including annular rupture, massive aortic regurgitation, destabilization of hemodynamic status related to rapid pacing, and possible cerebral embolization.⁷

Recently, two randomized controlled trials (RCTs)^8,9 investigating BE (balloon expandable) valves and the other SE (self expanding) valves, have reported a neutral effect of BAV on the success of the procedure, despite being limited by reduced sample size. Consequently, we performed a meta-analysis of RCTs and studies with multivariate adjustment to evaluate the impact of BAV on TAVI procedures.

The present work was conducted in accordance with current guidelines, including the recent Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) amendment to the Quality of Reporting of Meta-analyses (QUOROM) statement, as well as recommendations from The Cochrane Collaboration and Meta-analysis Of Observational Studies in Epidemiology (MOOSE).^10-13 The review was also registered online at its inception on metcardio.org.¹⁰

Search strategy. MEDLINE/PubMed was searched for pertinent articles published in English according to the following key words, in keeping with established methods¹⁴ and incorporating wild cards (identified by *): TAVI* or TAVR* and BAV* and Balloon Or Self Expanding NOT (review[pt] OR editorial[pt] OR letter[pt]). We did not search The Cochrane Collaboration CENTRAL database, as it only includes controlled clinical trials. Nonetheless, all corresponding authors of shortlisted studies were systematically and repeatedly queried for additional quantitative details. Concomitantly, they were asked for additional pertinent studies on the topic and offered coauthorship in the present work.

Study selection. Retrieved citations were first screened independently by two unblinded reviewers (FC, FDA) at the title and/or abstract level, with divergences resolved after consensus. If potentially pertinent, they were then appraised as complete reports according to the following explicit selection criteria, which were piloted over the first 5 cases.

Inclusion criteria were (all had to be met for inclusion): (1) human studies; (2) investigating patients undergoing TAVI comparing BAV vs not BAV; and (3) evaluating impact of BAV through randomization (RCT) of after comparison with multivariable analysis in observational studies.

Exclusion criteria were (any one alone was enough for exclusion): (1) non-human setting; and (2) duplicate reporting (in which case the manuscript reporting the largest sample of patients was selected).

Data extraction and endpoint. The following data were independently abstracted by two unblinded reviewers (FC, FDA) on pre-specified electronic forms, which were piloted over the first 5 cases, with divergences resolved after consensus. In particular, authors, journal, year of publication, location of the study group, baseline and procedural features were evaluated. Nonetheless, all corresponding authors of included studies were systematically queried to confirm or disprove abstracted data.

Device success was the primary endpoint (definition for each study is reported in Supplemental Table S1). All-cause mortality, 30-day moderate/severe AR (aortic regurgitation), stroke, permanent pacemaker implantation (PPI) and AKI (acute kidney injury) were the secondary ones. Analyses were performed according to design of the study (RCT and adjusted analysis) and type of valve (BE valves and SE valves).

Internal validity and quality appraisal. The quality of included studies was independently appraised by two unblinded reviewers (FC, FDA), on pre-specified electronic forms, which were piloted over the first 5 cases, with divergences resolved after consensus. Modifying the MOOSE item list in order to take into account the specific features of included studies,¹² we separately abstracted and appraised study design, setting, data source, and statistical methods for multivariable analysis, as well as, in keeping with The Cochrane Collaboration approach, the risk of analytical, selection, adjudication, detection, and attrition bias (expressed as low, moderate, or high risk of bias, as well as incomplete reporting leading to inability to ascertain the underlying risk of bias).

Statistical analysis. Continuous variables are reported as mean (standard deviation) or median (range). Categorical variables are expressed as n/N (%). Statistical pooling was performed according to a random-effects model with generic inverse-variance weighting, computing risk estimates with 95% confidence intervals, using RevMan 5.2 (The Cochrane Collaboration, The Nordic Cochrane Centre, Copenhagen, Denmark). A pooled analysis of odds ratios, relative risks or hazard ratios derived from each single study (from RCT or multivariate analysis) was performed after logarithmic transformation. We assumed similarity between the odds ratio and other relative measures such as relative risks and hazard ratios because endpoints were rare events. Graphical inspection of funnel plots was used to assess for study bias. Standard hypothesis testing was set at the two-tailed .05 level.

Study selection. A total of 467 reports were initially screened at the title and abstract level, of which 31 were retrieved in full text and evaluated for potential inclusion (Figure 1). We identified studies that compared TAVI with and without predilation with BAV and finally including 15 studies^15-29 that fulfilled the pre-specified inclusion criteria to be considered in the analysis (Figure 1 and Table 1). Twelve studies were observational and three were randomized controlled trials (RCTs) (Table 1).

Studies and patients’ characteristics. Studies were published between 2015 and 2020. Overall, 16408 patients were evaluated in 15 studies, among those 10364 (63.2%) underwent BAV prior to TAVI while 6044 (36.8%) were treated with direct implantation. SE valves were used exclusively in 2 studies (n=1434; 1137 BAV pre TAVI, 296 direct TAVI) and BE exclusively in 8 (n=2444; 1126 BAV pre TAVI, 1318 direct TAVI) (Table 1).

Baseline characteristics of overall population and of the two groups are shown in Supplemental Table S2. The mean age was 81.48 ± 1.44 years, with 49% of women. Diabetes was present in 30% (IQR 23-38) of patients, arterial hypertension in 86% (IQR 72-69), coronary artery disease in 52% (IQR 29-85). Mean ejection fraction (EF) was 55 ± 3.3%. Preprocedural risk was assessed by the Society of Thoracic Surgeons Predicted Risk of Mortality (STS-PROM) score in the majority of studies, being 6.6 ± 1.8.

Primary endpoint. At 30-days, predilation with BAV did not improve the rate of device success neither in the overall population (OR, 1.09; 95% CI, 0.90-1.31) (Figure 2) nor according to valve type: SE valves (OR, 0.93; 95% CI, 0.60-1.45), BE valves (OR, 1.16; 95% CI, 0.88-1.52) (Figure 3).

Secondary endpoint. Between BAV and direct TAVI approach, no differences in 30-day mortality (OR, 1.07; 95% CI, 0.87-1.31), 30-day stroke (OR, 0.83; 95% CI, 0.62-1.10), 30-day permanent pacemaker implantation (PPI; OR, 0.96; 95% CI, 0.80-1.16), AKI (OR, 1.13; 95% CI, 0.73-1.73) and 30-day moderate-to-severe AR (OR, 1.17; 95% CI, 0.91-1.49) were observed (Figures 4,5,6, and 7 and Supplemental Figure S1).

Postdilation rate was lower in BAV when accounting all studies (OR, 0.66; 95% CI, 0.47-0.94), and also when only RCTs were included in the analysis (OR, 0.48; 95% CI, 0.24-0.97) (Supplemental Figure S2).

According to the type of valve implanted, no differences were observed in 30-day stroke (OR, 0.78; 95% CI, 0.54-1.11 for BE valves and OR, 1.06; 95% CI, 0.62-1.80 for SE valves), 30-day PPI (OR, 0.90; 95% CI, 0.74-1.10 for BE and OR, 0.99; 95% CI, 0.63-1.55 for SE), and AR (OR, 1.23; 95% CI, 0.86-1.77 for BE and OR, 1.14; 95% CI, 0.81-1.61 for SE) (Supplemental Figure S3, Supplemental Figure S4, and Supplemental Figure S5).

The main findings of this meta-analysis can be summarized as following: (1) no differences were found in device success comparing direct TAVI with systematic predilation with BAV; (2)the use of postdilation was reported to be lower in the predilation group compared to direct TAVI, specifically in RCT;(3) 30-day moderate-to-severe aortic regurgitation rates were similar between the two approaches with no significant differences between SE and BE valves; (4) the safety of direct TAVI was comparable to systematic predilation, without significant differences in terms of 30-day all-cause mortality, 30-day stroke, 30-day PPI and AKI, with no difference according to valve type.

Device success, postdilation, and aortic regurgitation. Originally, when transcatheter aortic valve implantation (TAVI) was introduced, predilation with BAV was a mandatory step of the procedure.^30,31 Main assumed advantages are preparation of the native valve landing zone to facilitate passage of large profile valves and delivery systems, easier and more precise valve implantation, estimation of annular size and predicting risk of coronary occlusion. Nevertheless, with the use of multi-slice computed tomographic angiography for optimal pre-operative valve selection and with further device developments, its use has become more selective. Improved prosthesis expansion with new generation valves provides sufficient radial-force in most cases as reported in previous studies and meta-analysis.^32-34 Based on these studies, direct implantation has been more widely adopted as a means to simplify the procedure and its use has grown over recent years as reported in a recent published French registry.²⁷

Most of the comparisons between these two techniques were based on observational and adjusted studies.^17-22,24-29 Recently, two important RCTs, the DIRECT and the DIRECT-TAVI trial,^8,9 compared these two strategies both in SE and BE valves. Both trials reported non-inferiority of direct TAVI compared to a predilation strategy. Only non-significant differences were reported in specific subsets in which predilation might be considered as a first option. For instance, in the DIRECT trial,⁸ a trend towards better success rates for predilation in severe stenosis (pre-TAVI AVA <0.6 cm²) was observed while in the DIRECT-TAVI,⁹ BAV was still necessary in 7 patients (5.8%) initially allocated to direct implantation because of the failure to cross the valve or due to a clinical decision.

It is well recognized that post-TAVI AR incidence is associated with worse short- and long-term outcomes.^35-37 According to our results, 30-day moderate-to-severe AR was comparable between the two strategies. However, a positive trend was seen for the use of the direct approach in the BEV subgroup. This could be explained by the fact that BAV is frequently used in more complex anatomies, especially in retrospective studies where similar results between the two approaches were found when adjusting for the amount of aortic valve calcifications.²⁹

Interestingly, rates of postdilation were low (2.8%) and comparable to those between predilation and direct TAVI in the DIRECT-TAVI trial with BE valves, while they were significantly higher for the direct approach in the DIRECT trial using SE valves. Nevertheless, the authors of the DIRECT trial reported similar rates of postdilation (29.4%) compared to the SURTAVI trial (29%) with a total rate of postdilation reaching 22.2%.⁸ Whereas postdilation could be effective in reducing the severity of paravalvular leak, concerns have been raised about the possible complications of this procedure such as prosthesis damage, annular rupture and PPI. Nonetheless, some authors reported the safety and efficacy of performing postdilation in patients with moderate-to-severe AR following SE valve implantation without reporting any case of annular rupture or valve malfunction.³⁸ On the other hand, a higher rate of PPI was observed in these series (33.3% vs 20.7%),³⁸ and some concerns about higher rates of cerebrovascular events were raised regarding its use with BE valves.³⁹ Aortic leaflet calcium on CT scan, higher aortic annulus diameters and higher annulus-to-prosthesis area ratio were reported to be possible predictors of postdilation rate but these parameters were not specifically evaluated in RCTs.⁴⁰

Safety and cerebrovascular events. As TAVI indication has been recently extended to low-risk patients,^41,42 procedure related complications and possible adverse outcomes are becoming far more important. Stroke is a possible drawback of using a balloon to predilate a calcific and stenotic valve as embolic migration of calcified particles during BAV might be correlated with clinical and subclinical cerebrovascular events. Nonetheless, in our meta-analysis, no significant difference was found on 30-day stroke rate between groups. Yet, a positive trend for the predilation strategy was seen, especially in the BE subgroup. These results are in line with previous studies in which a higher risk of subclinical strokes was reported after direct TAVI with both valve types.^20,43

Post-TAVI AKI has a multifactorial etiology including procedural related factors such as contrast medium, hypotension during rapid pacing and embolization.⁴⁴ Avoiding BAV before TAVI may shorten procedural time, with consecutive reduction in total radiation exposure and amount of contrast agent. This may be relevant, particularly in those patients with compromised renal function where a lower contrast volume might be associated with lower rates of AKI. Nevertheless, none of the two approaches to TAVI were found to be effective in reducing AKI stage 2 and 3 at 30-day but a favorable trend was seen for direct TAVI. This might be explained by a relatively low difference in contrast quantity between the two techniques or by the fact that the contrast spared from an easy BAV could be used in a difficult direct valve release, especially in challenging anatomies.

Permanent pacemaker implantation. Post-TAVI conduction abnormalities and PPI are common with both early and newer generation prosthesis.⁴⁵ They are associated with the radial force exerted by the balloon or the valve itself on conduction pathways, resulting in both compression and edema. A recently published meta-analysis has shown an important impact of PPI after TAVI on long-term outcomes and an association with higher mortality and cardiac hospitalizations.⁴⁶ The implantation depth and the baseline right bundle branch block were found to be independent predictors of PPI in SE valves⁴⁷ while in a recent Brazilian registry⁴⁸ predilation was associated with a higher rate of new onset persistent left bundle branch block in this subset.

Remarkably, when performed alone, BAV is associated with only 1% rate of PPI⁴⁹ so that a “two-hit” pathophysiological model was proposed by Lange et al⁵⁰ for TAVI related conduction disturbances. The authors suggested that the first hit on conduction system during predilation with BAV would promote the persistence of advanced AV-blocks after final valve deployment. Nevertheless, differently from previous meta-analysis,⁵¹ and according to our results, the 30-day PPI OR was similar between the two approaches. Interestingly, some authors supported the use of “moderate predilation” in BEs in order to facilitate the final deployment without increasing the PPI risk.⁵⁰ However, we reported no benefit for predilation in the SE group, but a positive trend was seen in BE valves.

Study limitations. There are some limitations to our study. First, this meta-analysis includes mostly observational adjusted studies with only three RCTs included. Because of lack of randomization, each observational study carries intrinsic selection bias invariably associated with patients’ own characteristics, operators’ preferences at the time of TAVI in individual centers. Secondly, the wide variability in sample sizes across the studies represents another limitation of the present study. Thirdly, device success definition was not uniform across studies, as only the most recent studies used VARC-2 definitions. Furthermore, no information regarding amount of aortic valve calcium was included in the studies included herein. Since these CT data are of paramount importance for analysis of PPI and AR rates, conclusions regarding these clinical outcomes have to be drawn with caution.

Direct TAVI is feasible and appears safe compared to predilation with BAV. A direct approach is associated with comparable device success and performance. Direct implantation might be a reasonable first-choice approach with new-generation valves, leaving BAV predilation for selected cases. Further studies are needed to define clinical and anatomic features that would favor systematic BAV as part of TAVI procedure.

• Urologic Complications in Patients Receiving Indwelling Urinary Catheters During Transcatheter Aortic Valve Replacement
• The Prognostic Utility of Invasive Right Heart Catheterization in Patients Prior to Transcatheter Aortic Valve Replacement
• Outcomes of Transcatheter Aortic Valve Replacement in Patients Treated With Systemic Steroids