Balloon Versus Computed Tomography Sizing of the Aortic Annulus for Transcatheter Aortic Valve Replacement and the Impact of Left Ventricular Outflow Tract Calcification and Morphology on Sizing

Abstract: Objectives. To evaluate the role of balloon annular sizing in transcatheter aortic valve replacement (TAVR). Background. Multidetector cardiac computed tomography (MDCT) is the gold standard for aortic annular sizing in TAVR. Balloon sizing is increasingly used in patients with borderline annular size and severe calcification. A comparison between these two techniques is needed. Methods. We retrospectively compared baseline characteristics and 30-day outcomes of patients undergoing balloon-expandable TAVR using annular MDCT or balloon sizing. Paravalvular leak (PVL) rates were compared adjusting for access site, valve generation, size, and valve calcification. Results. A total of 205 patients underwent TAVR with MDCT (n = 110) or balloon sizing (n = 95). Balloon-sized patients were older (83 years vs 81 years; P=.03), had more valve calcification (60.2% vs 30.9%; P<.001), and underwent more minimalist TAVR (61.1% vs 40%; P=.03). Although we found no difference between balloon and MDCT sizing in rates of acute renal failure (3.2% vs 0.9%; P=.34), annular rupture (1.1% vs 1.8%; P>.99), ≥ mild PVL by angiography (40% vs 35.5%; P=.57), or 30-day transthoracic echocardiography (40.7% vs 29.3%; P=.78), balloon-sized patients had a higher aortic regurgitation index (≥25) of 74.4% vs 54.1% (P=.01). Thirty-day rates of ≥ moderate PVL were 7.0% with balloon and 5.7% with MDCT sizing (P=.34). Balloon sizing recommended a different valve size in 34.0% of patients who underwent both methods (n = 50). A different recommendation occurred more often in patients with moderate/severe annular calcification (50.0% vs 33.3%; P=.01) and non-tubular left ventricular outflow tracts (LVOTs) (70.6% vs 30.3%; P=.01). Conclusion. Balloon sizing can be a complement to MDCT for annular sizing in TAVR, especially in patients with moderate/severe annular calcification, borderline annular size, and non-tubular LVOT. 

J INVASIVE CARDIOL 2016;27(7):295-304. Epub 2016 April 15.

Key words: balloon, transcatheter aortic valve replacement, aortic stenosis, computed tomography, annular sizing

Multidetector cardiac computed tomography (MDCT) has become the gold standard for aortic annular sizing in balloon-expandable transcatheter aortic valve replacement (TAVR);^1-4 however, interpretation may be challenging in patients with borderline annular sizes and heavy calcification. In such scenarios, many centers have started using balloon sizing as an adjunctive technique.⁵ Balloon sizing has been previously described and has been shown superior to 2D transthoracic (TTE) or transesophageal echocardiogram (TEE) for TAVR sizing.^6-8 However, balloon sizing has never been compared with MDCT. We aimed to evaluate the safety and effectiveness of balloon sizing compared with MDCT in the prevention of paravalvular leak (PVL) and sizing-related complications in a cohort of patients who underwent TAVR with a balloon-expandable valve.

Methods

We performed a retrospective, single-center study of patients who underwent balloon-expandable TAVR using Sapien and Sapien XT valves (Edwards Lifesciences) with MDCT or intraoperative balloon aortic annular sizing from January 1, 2013 to November 30, 2014. In order to focus the study on the comparison between the two sizing strategies, we excluded patients who needed valve-in-valve procedures, had previous mitral valve replacement, underwent implantation with a Sapien 3 device (Edwards Lifesciences), or had a misplaced initial valve. The study was approved and performed in accordance with the regulations of the hospital institutional review board at Emory University in Atlanta, Georgia. 

Multidetector cardiac computed tomography sizing technique. All MDCT exams were performed on either a dual-source 64 detector or dual-source 128 detector scanner (Somatom Definition and Somatom Definition Flash, respectively; Siemens Medical Solutions). Retrospective electrocardiographic gating was performed with a collimation of 0.6 mm, fixed pitch of 0.2, and injection of 75 mL Isovue 370 contrast (Bracco Diagnostics, Inc). Scans were performed at 120 kV with tube current modified by patient size, but electrocardiographic dose modulation was not used as systolic images are necessary for annulus measurement. Images were reconstructed in 10% R-R intervals with a soft-tissue convolution filter (B26F). The images were interpreted on a dedicated 3D workstation (Siemens syngo.Via platform) and interpreted by experienced cardiac CT readers. The systolic phase was manually selected by the reader at the time of interpretation, and measurements including annular diameter and area were manually traced in the double-oblique transverse plane as previously described.^3,9,10 Valve size selection was primarily based on annular area, with a goal of 8%-15% oversizing (3.14-4.14 cm², 4.15–5.29 cm², and 5.3–6.6 cm² for the 23 mm, 26 mm, and 29 mm valves, respectively), as outlined in the manufacturer’s instructions for use.

Severity of calcification was determined using a semiquantitative evaluation of calcium distribution from the leaflet tips to the left ventricular outflow tract (LVOT) and from the aortic annulus to the LVOT that has been previously described by experienced MDCT centers for TAVR.^4,11 Valve calcification (leaflet tips to LVOT) was classified into: none; mild (1 nodule of calcium extending <5 mm in any direction and covering <10% of the perimeter of the annulus); moderate (2 nodules of calcification or 1 extending >5 mm in any direction or covering >10% of the perimeter of the annulus); or severe (multiple nodules of calcification of single focus extending >1 cm in length or covering >20% of the perimeter of the annulus). Annular calcification (annulus to LVOT) was classified into: none; mild (small isolated calcification spots); moderate (multiple large calcification spots); or severe (extensive calcification of all cusps).^4,11 A semiquantitative scale was used because calcification was calculated with images from either MDCT or non-gated chest CT,^4,11 which would have made any comparisons using a quantitative score (ie, Agatston) inaccurate.  

Balloon sizing technique. The balloon sizing technique has been previously described.^6-8 In patients with an 18-21 mm annulus size by TTE, a 21-22 mm balloon was used for balloon aortic valvuloplasty (BAV) (Figure 1). A Z-med (Numed) or Edwards Lifesciences balloon was prepared on a sterile table and the volume in the inflation syringe was adjusted so that complete inflation was required to reach final balloon size confirmed by calipers. The balloon was then deflated, inserted retrograde through the aortic valve, and used for BAV during rapid right-ventricular pacing. At full inflation, contrast was power injected into the proximal ascending aorta through a pigtail catheter placed at the level of the sinotubular junction (15 mL/s for a total of 25 mL volume, 900 psi). If contrast was not observed leaking around the balloon, a 23 mm valve was implanted. The presence of contrast leaking around the balloon and into the left ventricle identified the need for a larger valve, and a 26 mm valve was implanted (Figure 2). Similarly, in patients with a 21-24 mm or 24-28 mm annulus, a 24-25 mm or 27-28 mm balloon was used for BAV, respectively. The case was aborted if contrast leak was seen with a 27-28 mm balloon, as the largest available Sapien XT valve is 29 mm. We favor the use of a 2 mm difference between balloon and valve size (consistent with ≥10% oversizing by area). 

Patient and procedural characteristics. Patients were classified as balloon sized or MDCT sized. Baseline characteristics, procedural details, and 30-day outcomes were queried from our local TAVR registry database. We dichotomized race into Caucasian and non-Caucasian, and New York Heart Association (NYHA) class into I/II and III/IV. Procedural details and complications were defined by Valve Academic Research Consortium-2 (VARC-2) criteria.¹² “Minimalist” TAVRs were those performed via transfemoral access in the catheterization laboratory under conscious sedation and TTE guidance.¹³ For statistical power, a composite outcome of “sizing complications” was created by combining the occurrence of ≥ moderate PVL, second valve implantation, annular rupture, and type A aortic dissection.

All TAVRs were performed with a Sapien or Sapien XT valve. Access route was classified into transfemoral and non-transfemoral (transapical, transaortic, transcarotid, transinnominate, or transcaval). Fluoroscopy images were reviewed to determine the balloon sizing technique, number of BAVs and postdilation, second valve implantation, and postprocedural angiographic PVL (classified by Sellers criteria into none/trace, mild, moderate, or severe).¹⁴ Total BAVs were the sum of preimplantation BAVs and postdilations. Aortic regurgitation (AR) index was calculated from recorded procedural pressures using the previously validated formula: [(diastolic blood pressure – left end-diastolic pressure) / systolic blood pressure] x 100, and was dichotomized into <25 and ≥25.^15,16 

Thirty-day TTEs were reviewed by a blinded echocardiographer experienced in TAVR to classify PVL into none (or trace), mild, moderate, or severe using VARC-2 criteria.¹² Because of the low prevalence of ≥ moderate PVL, we dichotomized PVL into none or ≥ mild for our logistic regression models.

Subgroup analysis. Fifty patients underwent TAVR using both balloon and MDCT sizing. The sizing method that ultimately determined the implanted valve size identified 39 patients as the balloon-sized group who satisfied the algorithm for balloon sizing (Figure 1) and 11 patients as the MDCT-sized group. The degree of agreement between sizing methods was determined by comparing the valve size recommended by the balloon and MDCT sizing on each patient. Because our center had extensive experience using the balloon-sizing algorithm and began using MDCT sizing in December 2012, the use of MDCT to confidently select the appropriate TAVR size was not consistently implemented until 6 months later (June 2013). Consequently, most of the patients who had both sizing methods were implanted based on balloon sizing.

Procedural details and complications were determined. To evaluate if any anatomical characteristics were associated with the differences between the two methods, a blinded radiologist further reviewed the MDCT images and measured the diameters, area, and circumference at the annulus and at 4 mm into the LVOT. Using these measurements, the annular ellipticity index was calculated (annular ellipticity index = longest diameter / shortest diameter)¹⁷ and LVOTs were classified as tubular and non-tubular, depending on whether the measured areas at the annulus and at 4 mm below recommended the same or different transcatheter heart valve (THV) size, respectively (Figure 3). Non-tubular LVOTs were further classified as type B when the LVOT was smaller than the annulus (funnel shape) and type C when the LVOT was larger than the annulus (trumpet shape). 

Statistical analysis. Statistical analysis was done using IBM SPSS statistics (IBM). Differences between groups were compared using a Chi-square or Fisher’s test (when appropriate) for categorical variables and a non-parametric Mann-Whitney independent sample test for non-normally distributed continuous variables (determined by Shapiro-Wilk test), which are reported as median (interquartile range). Logistic regression models were created to compare the occurrence of PVL measured by angiography, AR index, and 30-day TTE-PVL between both groups, adjusting for valve type, valve size, access site, and aortic valve calcification. P-values <.05 were considered statistically significant. 

Results

A total of 205 patients underwent TAVR using MDCT (n = 110) or balloon sizing (n = 95). Baseline demographics and echocardiographic characteristics are shown in Table 1. All patients were in NYHA functional class III/IV at the time of TAVR. Balloon-sized patients were older (83 years vs 81 years; P=.03), more immunocompromised (11.6% vs 24.3%; P=.02), had a lower glomerular filtration rate (61.0 mL/min/1.73 m² vs 67.3 mL/min/1.73 m²; P=.05), and had a higher degree of calcification (60.2 vs 30.9; P<.001) than MDCT-sized patients. 

Procedural details are shown in Table 2. The balloon-sized group had fewer 29 mm valves implanted (9.0% vs 29.1%; P=.01), more minimalist TAVR (61.1% vs 40.0%; P=.03), more use of the transfemoral access site (73.7% vs 50.0%; P=.01), shorter intensive care unit time (23 hours vs 27 hours; P=.02), and shorter length of stay (3 days vs 4 days; P=.02). Twelve patients in the MDCT-sized group underwent TAVR without preimplantation BAV. As expected, balloon-sized patients had more preimplantation BAVs (2 vs 1; P=.01), longer fluoroscopy time (25.6 minutes vs 20.3 minutes; P=.01), and greater intraprocedure contrast volume (130 mL vs 108 mL; P=.01) than MDCT-sized patients.

There were no differences between groups regarding postprocedural complications (Table 3), including the composite outcome of sizing complications (14.7% in the balloon group vs 13.9% in the MDCT group; P=.82) and annular rupture (1.1% in the balloon group vs 1.8% in the MDCT group; P>.99). Of the 3 patients with annular rupture, 2 patients had a severe focal annular calcification that tracked down the aortic-mitral curtain, and 2 patients (both MDCT-sized) received a 29 mm valve with 20% oversizing. There were no coronary obstructions, and 2 patients (2.1%) developed transient severe aortic regurgitation after balloon sizing, which resolved with valve implantation. The 3 deaths occurred during the procedure (1 patient in the balloon group and 2 patients in the MDCT group).  

Thirty-day TTE was available on 197 (97.5%) of the surviving 202 patients (91 balloon and 106 MDCT). There were no differences between balloon and MDCT patients in 30-day left ventricular ejection fraction (55.0% vs 57.5%; P=.95) or aortic valve mean gradient (9.0 mm Hg vs 8.7 mm Hg; P=.44).

Paravalvular leak. The overall rate of ≥ moderate PVL was 6.8% by angiography (6.3% in the balloon group vs 7.3% in the MDCT group; P=.69) and 6.6% by 30-day TTE (7.0% in the balloon group vs 5.7% in the MDCT group; P=.34). PVLs were then dichotomized into none/trace and ≥ mild to increase the statistical power of the analysis. There was no statistical difference on the dichotomized rates of PVL between the balloon and MDCT groups when measured by either angiography (40.0% vs 35.5%; P=.50) or 30-day TTE (40.7% vs 29.3%; P=.09), even after adjusting for valve type and size, access site, and aortic valve calcification (P=.57 and P=.78, respectively) (Table 3). 

Aortic regurgitation index was available for 188 patients (90 in the balloon group and 98 in the MDCT group). Compared with MDCT-sized patients, balloon-sized patients were more likely to have a postprocedure AR index ≥25% (74.4% vs 54.1%; P=.01). This difference remained statistically significant (P=.01) after adjusting for valve type and size, access site, and aortic valve calcification. 

Subgroup analysis. Balloon sizing recommended a different valve size in 17 of the 50 patients (34%) who underwent both sizing methods: a bigger valve in 7 patients (14%) and a smaller valve in 10 patients (20%). In the former group, all patients received the recommended larger valve. In the latter group, 6 patients received the valve size recommended by balloon sizing, and 4 patients received a valve size recommended by MDCT (3 with an underexpanded larger valve). One patient in whom the balloon recommended a larger THV size than MDCT developed an annular rupture and subsequently died. MDCT parameters, procedure details, and complications between these subgroups are reported in Table 4. We found that balloon and MDCT strategies recommended conflicting valve sizes in patients with moderate/severe annular calcification (50.0% vs 33.3%; P=.01) and non-tubular LVOTs (70.6% vs 30.3%; P=.01). 

Discussion

Our study supports that balloon sizing can be a safe and effective method to size the aortic annulus and can be a valuable adjunct to MDCT prior to balloon-expandable TAVR. There were similar rates of PVL or sizing-related complications between the two methods, even after adjusting for valve type, valve size, access site used, and aortic valve calcification. Although the numbers are small, more balloon-sized patients had an AR index score ≥25 than MDCT-sized patients, a value that has previously been associated with better 1-year survival after TAVR.^15,16

Two groups of patients may specifically benefit from balloon sizing during TAVR: patients who cannot have an MDCT for logistic reasons (ie, urgent procedure) and patients who have inconclusive valve size selection by TTE, TEE, or MDCT. In our study, balloon sizing was discordant with MDCT and changed the treatment strategy in one-third of patients who had both sizing techniques, leading to either an implantation of a different valve size or an “under-filled” larger valve. Balloon sizing was more likely to recommend a different valve size in those patients with moderate/severe annular calcifications and non-tubular LVOTs. The observed rate of moderate PVL (6.3%) in patients who had the balloon sizing choose a different THV than MDCT suggests that the balloon method is preferred in patients with controversial anatomy. 

Most patients with moderate to severe calcification can have a successful TAVR, but are at a higher risk for PVL and TAVR-related complications.^10,11,18 Although MDCT can identify the amount and location of calcium in the LVOT,^10,18 a threshold level of calcification for aborting the procedure remains unknown. Balloon sizing as an adjunct to MDCT may help recognize patients with severe focal calcification who will have a failed TAVR procedure. In our experience, calcium-related leak during balloon sizing and complete filling of the sinuses of Valsalva by the balloon may identify patients at higher risk for annular rupture (Figure 4). The patient who died in our balloon-sized cohort had these characteristics during angiography and we have since identified other cases that resulted in severe PVL or rupture. Currently, we abort cases with focal calcification that prevent sealing by balloon (Figure 4). 

We also found that differences in LVOT morphology may identify cases when balloon sizing is a valuable adjunct to MDCT. As shown in Figure 3, a tubular LVOT (type A) has a long and homogeneous sealing zone that is accurately determined by MDCT measurements at the annular level. MDCT may overestimate (type B) or underestimate (type C) the landing zone size in those LVOTs with a non-tubular morphology, explaining the differences observed between balloon and MDCT sizing in our subgroup analysis. In these cases, balloon sizing provides additional value because it allows intraprocedural testing of the sealing zone. We theorize that non-tubular LVOTs (type C) with moderate/severe annular calcification may be the most difficult anatomy to prevent significant PVL due to a much shorter sealing zone. Although studies are needed to evaluate this hypothesis, our results suggest that a multimodality imaging strategy should strongly be considered in patients with borderline, non-tubular, and heavily calcified LVOTs. 

Study limitations. First, this study has the inherent limitations of a retrospective study with reliance on the quality of medical records. We increased the accuracy of our data by validating and reviewing the MDCT, TTE, TEE, and fluoroscopy images by a dedicated and experienced radiologist and echocardiographer. Second, PVL after TAVR can be difficult to evaluate because of the eccentric regurgitant jets, the echogenic shadows caused by the valve and calcifications, and the subjective scales used to classify its severity. We purposely used multiple methods to evaluate PVL (angiography, TTE, and AR index) to ameliorate this difficulty. Third, our aortic valve and annular calcification were calculated from MDCT or non-gated CT. We tried to produce a consistent classification by using the previously reported semiquantitative scales^10,18 proposed by the TAVR group in Vancouver, Canada, although calcification may be overestimated by non-gated CT. Fourth, there were some differences in the baseline and procedural characteristics between our groups. Although these variables may affect long-term outcomes, our acute and 30-day study results remained unaltered even after adjusting for important factors that affect the occurrence of PVL. Finally, the balloon-sizing algorithm was not used for selection of self-expanding THV. In these valves, we believe a similar algorithm selecting a THV that is ≥3 mm than the balloon used for sizing would result in outcomes similar to those published by Barbati and colleagues.¹⁹

Conclusion

Balloon annular sizing is a safe and effective strategy for valve size selection during TAVR that can be used as a complement to MDCT sizing. Balloon sizing can be especially beneficial in patients with borderline annular size, severe calcification, and non-tubular LVOT. Further studies are needed to test our results in TAVR with newer-generation devices and different sizing algorithms.

References

1.    Tops LF, Wood DA, Delgado V, et al. Noninvasive evaluation of the aortic root with multislice computed tomography implications for transcatheter aortic valve replacement. JACC Cardiovasc Imag. 2008;1:321-330.

2.    Stolzmann P, Knight J, Desbiolles L, et al. Remodelling of the aortic root in severe tricuspid aortic stenosis: implications for transcatheter aortic valve implantation. Eur Radiol. 2009;19:1316-1323.

3.    Jilaihawi H, Kashif M, Fontana G, et al. Cross-sectional computed tomographic assessment improves accuracy of aortic annular sizing for transcatheter aortic valve replacement and reduces the incidence of paravalvular aortic regurgitation. J Am Coll Cardiol. 2012;59:1275-1286.

4.    Webb J, Gerosa G, Lefevre T, et al. Multicenter evaluation of a next-generation balloon-expandable transcatheter aortic valve. J Am Coll Cardiol. 2014;64:2235-2243.

5.    Kasel AM, Cassese S, Bleiziffer S, et al. Standardized imaging for aortic annular sizing: implications for transcatheter valve selection. JACC Cardiovasc Imag. 2013;6:249-262.

6.    Babaliaros VC, Liff D, Chen EP, et al. Can balloon aortic valvuloplasty help determine appropriate transcatheter aortic valve size? JACC Cardiovasc Interv. 2008;1:580-586.

7.    Babaliaros VC, Junagadhwalla Z, Lerakis S, et al. Use of balloon aortic valvuloplasty to size the aortic annulus before implantation of a balloon-expandable transcatheter heart valve. JACC Cardiovasc Interv. 2010;3:114-118.

8.    Patsalis PC, Al-Rashid F, Neumann T, et al. Preparatory balloon aortic valvuloplasty during transcatheter aortic valve implantation for improved valve sizing. JACC Cardiovasc Interv. 2013;6:965-971. 

9.    Willson AB, Webb JG, Labounty TM, et al. 3-dimensional aortic annular assessment by multidetector computed tomography predicts moderate or severe paravalvular regurgitation after transcatheter aortic valve replacement: a multicenter retrospective analysis. J Am Coll Cardiol. 2012;59:1287-1294.

10.    Buellesfeld L, Stortecky S, Heg D, et al. Extent and distribution of calcification of both the aortic annulus and the left ventricular outflow tract predict aortic regurgitation after transcatheter aortic valve replacement. EuroIntervention. 2014;10:732-738.

11.    Barbanti M, Yang TH, Rodes Cabau J, et al. Anatomical and procedural features associated with aortic root rupture during balloon-expandable transcatheter aortic valve replacement. Circulation. 2013;128:244-253. 

12.    Kappetein AP, Head SJ, Genereux P, et al. Updated standardized endpoint definitions for transcatheter aortic valve implantation: the Valve Academic Research Consortium-2 consensus document. Eur Heart J. 2012;33:2403-2418.

13.    Babaliaros V, Devireddy C, Lerakis S, et al. Comparison of transfemoral transcatheter aortic valve replacement performed in the catheterization laboratory (minimalist approach) versus hybrid operating room (standard approach): outcomes and cost analysis. JACC Cardiovasc Interv. 2014;7:898-904.

14.    Sellers RD, Levy MJ, Amplatz K, Lillehei CW. Left retrograde cardioangiography in acquired cardiac disease: technique, indications and interpretations in 700 cases. Am J Cardiol. 1964;14:437-447. 

15.    Sinning JM, Hammerstingl C, Vasa-Nicotera M, et al. Aortic regurgitation index defines severity of peri-prosthetic regurgitation and predicts outcome in patients after transcatheter aortic valve implantation. J Am Coll Cardiol. 2012;59:1134-1141.

16.    Vasa-Nicotera M, Sinning JM, Chin D, et al. Impact of paravalvular leakage on outcome in patients after transcatheter aortic valve implantation. JACC Cardiovasc Interv. 2012;5:858-865. 

17.    Buellesfeld L, Stortecky S, Kalesan B, et al. Aortic root dimensions among patients with severe aortic stenosis undergoing transcatheter aortic valve replacement. JACC Cardiovasc Interv. 2013;6:72-83.

18.    Feuchtner G, Plank F, Bartel T, et al. Prediction of paravalvular regurgitation after transcatheter aortic valve implantation by computed tomography: value of aortic valve and annular calcification. Ann Thorac Surg. 2013;96:1574-1580.

19.    Barbanti M, Sgroi C, Imme S, et al. Usefulness of contrast injection during balloon aortic valvuloplasty before transcatheter aortic valve replacement: a pilot study. EuroIntervention. 2014;10:241-247.

From the ¹Structural Heart and Valve Center, Divisions of Cardiology, ²Cardiothoracic Surgery, and ³Department of Radiology and Imaging Sciences, Emory University School of Medicine, Atlanta, Georgia; and ⁴Department of Radiology; Konkuk University Medical Center, Konkuk University School of Medicine, Seoul, Korea.

Disclosure: The authors have completed and returned the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Thourani reports grants from Edwards Lifesciences, Medtronic, St Jude Medical, Sorin Medical, Boston Scientific, Abbott Medical, and Directflow Medical; equity in Apica. Dr Block reports personal fees from Edwards Lifesciences; past stockholder in Directflow Medical. Dr Babaliaros reports grants and personal fees from Edwards Lifesciences; personal fees from Directflow Medical, Medtronic, St Jude Medical, Boston Scientific, and Abbott Medical. The remaining authors report no conflicts of interest regarding the content herein.

Manuscript submitted January 4, 2016, provisional acceptance given February 5, 2016, final version accepted February 10, 2016.

Address for correspondence: Vasilis Babaliaros, MD, Emory University Hospital, F606, 1364 Clifton Road, Atlanta, GA 30322. Email: vbabali@emory.edu