Procedural Mortality With Transcatheter Aortic Valve Replacement – Balloon Inflation is Associated With Increased Risk

Objectives. To assess the impact of balloon use for predilation, valve implantation, or postdilation on in-hospital mortality among patients undergoing transcatheter aortic valve replacement (TAVR). Background. TAVR utilizes self-expanding, mechanically expanding, or balloon-expandable valves. Balloon inflation is inherent to deployment of balloon-expandable valves. Balloons may additionally be used with all valve types for pre- and postdilation. The relationships between valve mechanism, balloon use, and in-hospital mortality are not fully characterized.  Methods. Prospective data were collected on 4063 patients undergoing TAVR for aortic stenosis at 4 high-volume centers in the United Kingdom. In-hospital mortality was analyzed according to valve expansion mechanism, use of balloons for pre- and postdilation, and specific cause of death.  Results. Mean patient age was 83 ± 8 years. Implanted valves were self expanding (n = 2241; 55%), mechanically expanding (n = 1092; 27%), or balloon expandable (n = 727; 18%). In-hospital death occurred in 66 cases (1.6%). Thirty-six deaths (54.5%) were classified as implantation-related mortalities, with rates of 0.8%, 0.5%, and 1.7% (P=.04) among self-expanding, mechanically expanding, and balloon-expandable technologies, respectively. Patients who underwent balloon inflation at any stage of their procedure (n = 2556; 63%) had significantly higher implantation-related mortality than those who did not (1.3% vs 0.3%, respectively; P<.01). Balloon-expandable valve procedures were associated with significantly higher all-cause mortality (2.6% vs 1.4%; P=.02) and implantation-related mortality (1.7% vs 0.7%; P=.02) than non-balloon-expandable valve procedures. Balloon-related complications accounted for 18 cases (26%) of total in-hospital mortality, including all 12 cases (17.4%) of annular rupture and 5 cases (7.2%) of coronary occlusion. Conclusions. Balloon use for predilation, valve implantation, or postdilation was associated with an increased mortality risk. Balloon-related complications were the largest contributor to in-hospital mortality, comprising all cases of annular rupture and the majority of coronary occlusion cases.

J INVASIVE CARDIOL 2021;33(10):E761-E768. Epub 2021 September 8.

Key words: aortic stenosis, balloon valvuloplasty, death, TAVR

Since the first TAVR procedure in 2002, the field has experienced rapid growth and innovation. Following publication of the PARTNER 3¹ and Evolut Low Risk² trials, TAVR is now a viable option to treat aortic stenosis across all patient risk categories. Mortality rates have consistently been non-inferior to surgical aortic valve replacement. Other complications, including paravalvular leak and valve malpositioning, continue to improve with refinement of technology and technique.³

Currently available transcatheter heart valves (THVs) are either self expanding, mechanically expanding, or balloon expandable. Unique advantages and disadvantages have become evident between the various technologies based on their underlying mechanisms, influencing their compatibility with different patient anatomies. To date, only limited randomized trials have been performed directly comparing THVs. No difference between devices in terms of overall mortality has been identified,  although other clinical outcomes have shown variation. Rates of aortic regurgitation and the requirement for additional valve implantation were higher with the self-expanding CoreValve (Medtronic) compared with the balloon-expandable Sapien XT (Edwards Lifesciences).⁴ However, this was not identified in a subsequent comparison of the self-expanding Evolut R (Medtronic) with the balloon-expandable Sapien 3 (Edwards Lifesciences).⁵ The mechanically expanding Lotus valve (Boston Scientific) had less disabling stroke and aortic regurgitation compared with the CoreValve, at a cost of greater pacemaker requirement and valve thrombosis rates.⁶ The self-expanding Acurate Neo (Boston Scientific) was found to have higher rates of acute kidney injury and aortic regurgitation than the Sapien 3 valve.⁷

Balloon-expandable valves have consistently been shown to carry the greatest risk for annular rupture.^8,9 Balloon use for pre- and postdilation in TAVR is commonplace and may also result in aortic trauma. There are limited published data on the comparative rates and causes of in-hospital mortality according to THV implantation mechanics and the use of pre- and postdilation. In this study, we aim to further explore the influence of mechanism of valve expansion and balloon use on acute mortality.

Data were prospectively collected at 4 high-volume United Kingdom centers with wide experience in implantation of all 3 valve types. Valve selection in each case was according to clinician judgment as to the most optimal device for the individual patient. All patients provided consent for anonymized data collection. Records of all 4219 patients undergoing TAVR between January 2013 and March 2020 were analyzed. A total of 136 patients with primary aortic regurgitation were excluded, as were 20 patients with inadequate data regarding mortality status.

Information regarding patient demographics, procedural details, complications, and in-hospital mortality was obtained. Patients were grouped by implanted THV type and by the presence or absence of balloon use for predilation, valve deployment, or postdilation. In-hospital mortality was divided into all-cause and implantation-related cohorts. Implantation-related mortality was defined as death resulting directly as a result of valve introduction. This included cases with fatal complications following balloon predilation, valve positioning, valve deployment, and postdilation. The relationships between balloon use at each stage of TAVR and mortality were then analyzed.

Statistical analysis. Continuous variables were non-parametric and are reported as median ± interquartile range. Categorical variables are given as integers and percentages. Data analysis was performed with SPSS Statistics, version 26.0 (IBM Corporation) using the Kruskal-Wallis test for continuous variables and Pearson’s Chi-squared test for categorical variables. Fisher’s exact test was used in the event of cells with an expected count of <5.

In-hospital mortality data were available for 4063 patients who underwent TAVR for aortic stenosis. Baseline demographics are detailed in Table 1. Mean patient age was 83 ± 8 years; 56.1% were men. Most cases were performed via percutaneous femoral access (89.1%) and under conscious sedation (80.8%). Clinical outcomes are documented in Table 2.

A total of 12 valve types were included in the analysis. The self-expanding valves implanted were the Medtronic CoreValve (18.5%), Medtronic Evolut R (17.5%), Boston Scientific Acurate Neo (12.5%), Medtronic Evolut Pro (5.7%), Edwards Lifesciences Centera (0.4%), Abbott Portico (0.4%), and MicroPort Vitaflow (0.05%). Balloon-expandable devices were the Edwards Lifesciences Sapien 3 (14.4%), Sapien XT (3.4%), and Sapien 3 Ultra (0.1%). Mechanically expanding THVs comprised the Boston Scientific Lotus (25.9%) and Lotus Edge (1.0%).

The most commonly deployed valve group was self-expanding (n = 2241; 55.2%), followed by mechanically expanding (n=1092, 26.9%) and balloon expandable (n = 727; 17.9%). Predilation was performed in 50.8% of procedures and was significantly more frequent with self- and mechanically expanding valves compared with balloon-expandable valves (54.1% vs 51.3% vs 39.9%, respectively). Rates of postdilation were highest after self-expanding (24.1%) and balloon-expandable (14.8%) valve deployment. Postdilation was infrequent with mechanically expanding valves (4.3%). Both predilation and postdilation were performed in 11.8% of patients (Figure 1).

Mortality. Sixty-nine patients died in-hospital. Three procedural deaths occurred prior to valve implantation, 2 from annular rupture following balloon predilation. All-cause mortality was 1.7% and implantation-related mortality was 0.9%. No statistical difference was found in all-cause mortality rates between the self-expanding, mechanically expanding, and balloon-expandable valve groups (1.5% vs 1.3% vs 2.6%, respectively; P=.06). However, implantation-related mortality differed between valve classes (0.8% vs 0.5% vs 1.7%, respectively; P=.04) (Figure 2).

Requirement for balloon dilation was associated with increased mortality rates (Figure 3). The 2028 patients who underwent predilation had increased implantation-related mortality compared with the 1965 patients who did not (1.2% vs 0.6%, respectively; P=.05). When compared against non-balloon based valve types, balloon-expandable valves were associated with higher all-cause mortality (2.6% vs 1.4%; P=.02) and implantation-related mortality (1.7% vs 0.7%; P=.02). Postdilation was undertaken in 675 patients (16.6%) and was associated with higher implantation-related mortality (1.6% vs 0.7%; P=.04). Patients with exposure to a balloon at any stage of their procedure (n = 2556; 65.3%) had significantly greater implantation-related mortality (1.3% vs 0.3%; P<.01) than those who did not (n = 1356; 34.7%;).

Due to the significant differences in access routes and aortic stenosis etiology between valve groups, the subpopulation of patients with native valve aortic stenosis who underwent femoral percutaneous TAVR was separately analyzed. The use of balloon-expandable valves compared with other types showed numerically higher all-cause (2.2% vs 1.2%; P=.09) and implantation-related mortality (1.4% vs 0.6%; P=.09), although statistical significance was lost. The involvement of any balloon continued to result in higher implantation-related mortality than no-balloon procedures (1.2% vs 0.2%; P<.01).

The individual causes of mortality are listed in Table 3. Eighteen deaths (26.1%) were related to balloon inflation, comprising the largest group contribution to mortality. This included 12 cases of annular rupture (17.4%) and 5 cases of coronary occlusion (7.2%). Only 2 cases of coronary occlusion occurred in the non-balloon-expandable valve cohort without postdilation involvement, of which 1 was a valve-in-valve procedure. Other major causes of death included 10 vascular access complications (14.5%), 6 ischemic strokes (8.7%), and  8 cases of multiorgan failure (11.6%). Ventricular perforation secondary to the left ventricular wire or temporary pacing wire resulted in 3 deaths (4.3%).

This study demonstrates that balloon inflation at any stage of a TAVR procedure is associated with a small but measurable mortality risk. Increased implantation-related mortality was consistently seen in both pre- and postdilation groups and a significant increase in all-cause and implantation-related mortality was seen with balloon-expandable valves.

Predilation is performed to assist with advancement of THVs across stenotic valves and to facilitate valve expansion. However, as valve delivery systems have improved, studies have suggested that predilation may be unnecessary¹⁰ or possibly harmful.¹¹ Longer procedural times¹² and increased rates of acute kidney and myocardial injury¹³ have been reported. Randomized trials have shown that avoiding predilation had no impact on successful valve deployment with both the balloon-expandable Sapien 3 and self-expanding Medtronic valves.¹⁴ A non-randomized study of mechanically expanding valve implantation without predilation showed acceptable safety and efficacy outcomes.¹⁵ Known significant risks of balloon aortic valvuloplasty include material embolization, severe aortic regurgitation,¹⁶ and annular rupture (noted in 2 cases in our series). Operators must weigh the potential risks of predilation with the need to prepare the valve landing zone. Judicious balloon sizing relative to annular dimensions is likely to help limit risk.

Balloon-expandable valve technologies offer rapid, straightforward valve implantation. However, balloon inflation results in circular valve expansion and variable pressure delivery depending on annular resistance. Areas of heavy or spiculated calcification may therefore penetrate the aortic wall on expansion. Accordingly, annular rupture is seen predominantly with balloon-expandable valves.⁸ A significantly increased risk of annular rupture with balloon-expandable valves in the setting of moderate or severe left ventricular outflow tract calcification (4.0% vs 0.4%; P<.01) was recently reported.⁹ Avoidance of balloon-expandable THVs in patients at higher risk of rupture will likely be beneficial. Computed tomography analysis can identify patients with left ventricular outflow tract calcification and help avoid valve oversizing.¹⁷ Valve recapture or repositioning is not possible with current balloon-expandable devices, increasing the risk of irreversible coronary occlusion. The combination of these factors likely contributed to our finding of increased procedural mortality rates with this valve type. It remains important to note that mortality beyond hospital admission was not assessed, and that strengths of this valve design, including low pacing requirements and ease of future coronary access, may provide late advantages.

Postdilation is performed to reduce or eliminate paravalvular leak or to minimize transvalvular gradient. Residual aortic regurgitation is important as it is linked to adverse prognosis.¹⁸ However, postdilation itself carries a significant risk of mortality from both annular rupture and coronary occlusion. Selecting the appropriate valve type and size for the individual patient anatomy helps to limit any aortic regurgitation. When aortic regurgitation is present, balancing the benefit of postdilation against the risks of annular rupture or coronary occlusion is a matter of judgment in each individual case.

The strength of this study lies primarily within the large number of cases performed at tertiary centers familiar with all 3 device types. Valve selection was therefore guided by compatibility with patient anatomy after assessment with standard practice imaging modalities, rather than THV availability or operator skillset. This is reflective of real-world practice. Access to individual medical records allowed accurate assessment of the cause of in-hospital death. Identification of implantation-related mortality highlights causes of death that may be avoided with improved or alternative valve technology/implantation technique.

Study limitations. Categorization of mortality into implantation-related and unrelated causes is open to debate. Stroke was considered implantation related because there are data suggesting a difference in the size and quantity of tissue particles captured in cerebral protection devices between valve types.¹⁹ Additionally, valves that may be manipulated, repositioned, and redeployed offer increased opportunities for material embolization. Another weakness of this analysis is that patients were not randomly allocated to valve types. Significant differences were noted between groups for multiple demographic and procedural variables. Alternate access TAVR, especially transapical, has been known to incur higher mortality and this may confound the increased mortality seen with balloon-expandable THVs in this study. However, when cases with native valve aortic stenosis and femoral percutaneous access were analyzed alone, balloon involvement at any stage continued to demonstrate increased implantation-related mortality.

The requirement for balloon pre- and postdilation is often reflective of high-risk anatomy; therefore, increased procedural mortality may result from factors beyond balloon inflation alone. Potential confounders include a high burden of valvular calcification, bicuspid anatomy, and oval-shaped aortic valve annuli. Despite this, annular rupture and coronary occlusion in these settings are most often a direct and immediate consequence of balloon dilation. Given the strong link between balloons and these high-mortality complications, further refinement of patient selection for this technology is required. Despite standard-practice patient anatomical assessment and the availability of other THV types, our series demonstrated that balloon-expandable valves were still implanted in patients who proceeded to annular rupture. In our experience, the minimal rates of annular rupture and postdilation required with mechanically expanding valves should lead to increased use of these devices when facing high-risk anatomy.

Use of a balloon during TAVR for predilation, valve deployment, or postdilation is associated with an increased mortality risk. Balloon-related complications were the largest single contributor to in-hospital mortality in this series, accounting for all cases of annular rupture and the majority of coronary occlusion cases. These results should be kept in mind when considering balloon use.