Author + information
- Received July 16, 2013
- Revision received October 15, 2013
- Accepted October 24, 2013
- Published online February 1, 2014.
- Vinayak N. Bapat, MD∗ (, )
- Rizwan Attia, MD and
- Martyn Thomas, MD
- Department of Cardiothoracic Surgery and Cardiology, Guy's and St. Thomas' Hospital, London, United Kingdom
- ↵∗Reprint requests and correspondence:
Mr. Vinayak (Vinnie) N. Bapat, 6th Floor East Wing, Department of Cardiothoracic Surgery, St. Thomas' Hospital, Westminster Bridge Road, London SE1 7EH, United Kingdom.
The goal of this study was to provide a measurement of the true internal diameter (ID) of various surgical heart valves (SHV) to facilitate the valve-in-valve (VIV) procedure. During a VIV procedure, it is important to choose the right of the transcatheter heart valve (THV). Most users use the stent ID of an SHV to select the appropriate THV size. Echocardiography and computed tomography measurements are not yet standardized for measuring the ID of a variety of SHVs. Hence, we measured the true ID of SHV to assess the effect of valve design on the stent ID. Thirteen types of stented and 3 types of stentless valves were evaluated. True ID measurements were obtained using calipers and Hegar dilators. These were compared with the stent ID measurements. Fluoroscopy was used to confirm the impact of SHV designs on the true ID. Caliper measurements were found to be inaccurate and are hence not recommended. Hegar dilator measurements revealed a trend of reduction in stent ID. Porcine valves were most affected by their design, with reduction in the stent ID by at least 2 mm; pericardial valves with leaflets sutured inside the stent had the stent ID reduced by at least 1 mm, and SHV with leaflets sutured outside the stent had no effect on stent ID. In the majority of SHV designs, there is a reduction in the stent ID as a result of leaflet tissue. This is important in borderline sizes to avoid problems associated with oversizing and also to confirm suitability for the VIV procedure in the smaller label sizes of SHV.
Transcatheter aortic valve implantation (TAVI) has established itself as an accepted therapy for inoperable and high-risk patients with calcific aortic stenosis (1). Clinical need has led to the use of this technology in treating degenerated bioprosthetic surgical heart valves (SHV) to avoid redo open heart surgery (2). Multiple reports of valve-in-valve (VIV) procedures have appeared in the literature over the last 2 years, with substantial experience being acquired in treating a degenerated SHV in the aortic and mitral positions, and lesser experience in treating degenerated SHV in pulmonary and tricuspid positions (2–11).
This therapy area continues to grow rapidly because VIV treatment appears promising when compared with a redo open heart operation, due to its less invasive nature. One of the important determinants of immediate and long-term success of this novel treatment is choosing the right size of the transcatheter heart valve (THV) for a given SHV type and size. In a native aortic valve, measurements are performed at the level of the aortic annulus to determine the size of the THV (12). When performing a VIV procedure, the majority of current users use the stent internal diameter (ID) of an SHV to select the appropriate THV size (2–5). However, the SHV design may have an impact on this measurement because of the leaflet tissue mounted within the stent frame. We evaluated the effect of valve design on stent ID and discuss the concept of the true ID of an SHV, which is the relevant ID for the VIV procedure. We also briefly discuss sizing considerations for a VIV procedure in the aortic versus the mitral position.
Thirteen types of stented and 3 types of stentless aortic SHV of all sizes were obtained from various manufacturers (Table 1, Fig. 1). The 3 types of stentless aortic SHV studied were those implanted as aortic roots (Fig. 2). Stentless valves, which are implanted in a subcoronary position within a native aortic root, were excluded because they essentially behave like native aortic valves and take up the dimensions of the root in which they are implanted (Fig. 2).
We have previously published the stent ID of all SHV along with other dimensions (13). The industry standard when reporting the stent ID is the ID of the stent frame when covered with fabric or pericardium but without the valve leaflets (Fig. 3).
True ID was defined as the ID of the inflow of the SHV.
An attempt was made to obtain the ID measurements with the use of a Vernier caliper (Fig. 4). The caliper was introduced through the inflow of the SHV, and measurements were obtained with minimal distortion. At least 2 measurements were obtained for each SHV, and 3 operators independently measured all sizes of each type of SHV.
Hegar dilator measurements
We used Hegar dilators with increments of 0.5 mm to measure the true ID. The Hegar dilator has a conical tip; thus, it can be easily introduced within an SHV (Fig. 5) from the valve inflow. The main body of the dilator is a perfect circle in cross section and, hence, gives an exact diameter. Measurements were performed using incremental sizes of Hegar dilators, and the largest size of the dilator that could be placed within an SHV was noted for each SHV (Fig. 5). Measurements were performed by 3 operators independently.
Confirmation of hypotheses using fluoroscopy
SHV are essentially of 3 types depending upon the leaflet material and placement of the leaflets: 1) porcine valve with leaflets sutured inside of the stent; 2) pericardial valve with bovine pericardial leaflets sutured inside of the stent; and 3) pericardial valve with bovine pericardial leaflets sutured outside of the stent. To assess the effects of valve design on the stent ID, 1 SHV of each type with a radio-opaque stent frame was chosen. These were the Carpentier-Edwards (CE) standard, Perimount (Edwards Lifesciences, Irvine, California), and Trifecta (St. Jude Medical, St. Paul, Minnesota), respectively. A CoreValve THV (Medtronic, Minneapolis, Minnesota) was implanted within each of these, and the degree of separation between the radio-opaque stent frame and the CoreValve stent frame (inflow portion) was observed; this essentially is the difference between the stent ID and the true ID.
True ID measurement with calipers
Because the SHV can be easily distorted with the lateral push of the caliper arms, there was a large variability in these measurements (from 0.5 to 2 mm) (Fig. 4). This was particularly true for SHV designs such as the CE porcine valve, which has an asymmetric leaflet structure (Fig. 4) and lacks a complete polymer/metal ring at the base of its frame. This was also observed with the stentless valves because they lack a rigid base. An attempt was made to obtain measurements without distorting the valve, but the interoperator variability was still large; thus, we did not use this method for measurement of the true ID.
True ID measurement with Hegar dilators
Stented Aortic SHV
The true ID was smaller than the stent ID in the majority of SHV (Table 2). There was no interoperator variability found in the Hegar dilator measurements except for the CE porcine valve, where 0.5-mm variability was noticed only in the larger sizes for the reasons mentioned in the preceding text. In these cases, we have taken the largest measurement for analysis. We found the following trends in the reduction of the true ID depending on the design of the SHV:
1. Porcine SHV: The porcine valve leaflets are always sutured inside of the stent frame, and the true ID is at least 2 mm less than the stent ID (Fig. 6A). Examples: Hancock II (Medtronic), Mosaic (Medtronic), Aspire (Vascutek, Inchinnan, United Kingdom), CE porcine standard, CE porcine S.A.V. (Edwards Lifesciences), Epic/Biocor (St. Jude Medical), and Epic/Biocor Supra (St. Jude Medical).
2. Pericardial SHV with leaflets sutured inside of the stent: The effect of the pericardial leaflets was less than that of the porcine leaflets, and the difference between true ID and stent ID was 1 mm (Fig. 6B). Examples: Perimount (Edwards Lifesciences), Perimount 2700 (Edwards Lifesciences), Magna/Magna Ease (Edwards Lifesciences), and Soprano (Sorin, Milan, Italy).
3. Pericardial SHV with leaflets sutured outside of the stent: Because the leaflets were sutured outside of the stent, the stent ID and the true ID were similar (Fig. 6C). Examples: Mitroflow (Sorin) and Trifecta.
Stentless Aortic SHV
Although there is variability between various manufacturers, the true ID was always smaller than the labeled size (which corresponds to the root diameter) (Table 2).
Fluoroscopy confirmed the aforementioned findings (i.e., the effect of leaflet mounting on the reduction of stent ID). The degree of separation of a fully deployed CoreValve THV was greater in the CE porcine standard valve when compared with the Perimount SHV. In Trifecta, the stent frame and CoreValve inflow were in approximation, confirming the Hegar dilator findings (Fig. 7).
TAVI is now well established as a treatment modality for aortic stenosis. A large body of experience is accumulating with the VIV procedure using Sapien and Sapien XT valves (Edwards Lifesciences) and CoreValve and Evolut valves (Medtronic) (2–11). Because both of the valves were designed for the aortic position, the largest experience within the VIV field has been in the treatment of degenerated aortic SHV (2–8). Because of favorable early results, the indication has slowly progressed to treating degenerated SHV in the mitral, tricuspid, and pulmonary positions (9–11). However, recent data reported in the global registry have raised doubts about the utility of VIV in treating smaller label sizes of a variety of SHV (14). This is due to the higher residual gradients observed in SHV with label sizes ≤21 mm leading to “patient–prosthesis mismatch.” Mismatch is well known to be associated with a smaller reduction in gradients and less satisfactory benefits in terms of left ventricular systolic and diastolic function and left ventricular mass regression. The global registry also highlighted a 15.3% incidence of malposition and a 3.5% incidence of coronary obstruction (14).
Part of the problem is a limited understanding of the various designs of the SHVs implanted in the last 2 decades and their impact on the VIV procedure. One must take into account the design of an SHV, the design of a THV, and their interaction to achieve optimal results. Two important determinants of immediate and long-term success of this novel treatment are choosing the right THV size and securing the THV in an optimal position within the SHV. With limited sizes of THV available, one must choose a correctly sized THV to match an SHV. Undersizing will lead to a large paravalvular leak and/or embolization, whereas oversizing will lead to incomplete expansion of the THV, which can contribute to improper functioning and/or higher residual gradients (14).
In a native aortic valve, the ID measurements are performed at the level of the aortic annulus to determine the size of the THV (1,12). These are obtained either by transthoracic echocardiography, transesophageal echocardiography, computed tomography (CT) scan, or balloon sizing during the procedure (12,15). The THV size is then chosen with a degree of oversizing to achieve secure fixation. The majority of TAVI users have extrapolated the same principle when performing a VIV procedure (2–11). It is well known that for a given native aortic annulus diameter, there is a certain amount of variability in measurements obtained by various modalities, and hence, the ID of an SHV measured by these modalities may not be the same and may reflect neither the stent ID nor true ID. Hence, the most common measurement used today is the stent ID of an SHV provided by the manufacturer (2,5).
It is well known that the same label size of various stented SHV will differ in their stent ID measurements, which has led to confusion in the past. However, multiple publications have addressed this issue by compiling data from the manufacturers to provide precise stent ID for all label sizes of a variety of SHV (5,13). By convention, however, the stent ID represents the ID of a bare stent covered with fabric or pericardium only. It does not take into account the effect of artificial leaflets sutured within the stent (13). In a porcine SHV, the porcine leaflets are always sutured inside the stent frame and, because of their natural hinge point mechanism, tend to encroach maximally inside the stent. Hence, the true ID is reduced by at least 2 mm. We found that the range of reduction could be 2 to 4 mm and was largest in larger label sizes. The variability is due to the fact that these leaflets are hand sewn, and porcine leaflets from different pigs, even when matched, can vary in their thickness. In a pericardial SHV, the thickness of the bovine pericardium for the leaflets is controlled and matched, and the way the pericardium is sutured inside the stent leads to a reduction in the stent ID by only 1 mm. In designs such as that of the Trifecta and Mitroflow valves, because the sheet of pericardium is wrapped outside the stent, the leaflets have no effect on the stent ID.
The implications of these findings are 4-fold:
1. When treating borderline sizes, especially with porcine SHV, use of the true ID instead of the stent ID should lead to choosing a smaller THV size than initially planned. For example, label size 25 of Hancock II has a stent ID of 22.5 mm and, hence, would point toward the use of Sapien/Sapien XT size 26. However, the true ID is 20.5 mm, and hence, the correct size of Sapien/Sapien XT would be 23. A similar example for CoreValve usage concerns label size 27 of the Epic, where the stent ID (25 mm) would direct one to the use of a size 29 CoreValve, but the true ID (22.5 mm) points to the use of a size 26 CoreValve.
2. When treating labeled SHV sizes ≤21 mm, although the label size and/or stent ID appear adequate, a look at the true ID may reveal them to be unsuitable for VIV treatment. For example, Hancock II size 21 has a true ID of 16.5 mm, and hence, if treated with either Sapien size 23 or Evolut size 23, will result in a size mismatch and high residual gradients. This is also true for pericardial valves with a small true ID such as Perimount size 19 (true ID 17 mm). This finding also explains why the incidence of high gradients in the global registry (mean gradient ≥20 mm Hg) was 28.4%, and the gradients were higher after a VIV in SHV sizes ≤21 mm. THV are available in multiple sizes and have a lower and upper limit of diameter tolerance recommended by the manufacturer. We have used the same sizing guidelines for recommending use of a THV in a given SHV (Table 2). Although clinical cases have been performed, we felt that caution should be exerted when evaluating these cases for VIV. If implanted in a smaller diameter, either native aortic annulus or within a degenerated SHV, it may result in suboptimal function and higher residual gradients.
3. Oversizing can lead to outward deflection of the calcified leaflets, which will increase the chance of coronary obstruction. This risk will be higher if the sinuses are shallow and with certain designs in which the leaflets are outside of the stent frame (14,16,17).
4. Larger-sized SHV may still be suitable for VIV and may, in practice, end up with the best results because of having the least chance of a mismatch. For example, size 29 Magna and Trifecta have a true ID of 27 and 26 mm, respectively, and can be treated with either size 29 Sapien/Sapien XT or CoreValve size 31.
The stentless SHV design also poses certain challenges during a VIV procedure. The lack of radio-opaque markers increases the chances of malposition (2,14). The lack of support for anchoring, compared with stented SHV, can also contribute to embolization of the THV because the distensibility of the aortic root is maintained (18). Stentless valves are also associated with a higher risk of coronary obstruction, because one may need to err on the side of choosing a larger THV for a given diameter to ensure secure anchoring. This was observed in the global registry, which reported the highest incidence of coronary obstruction in the stentless valves (14,19). However, we feel that the risk of coronary obstruction is higher in stentless valves implanted as a subcoronary implantation (e.g., the O'Brien [CryoLife, Atlanta, Georgia], Freedom solo [Sorin], and Toronto SPV [St. Jude Medical] valves), where the suture line is close to the native coronary ostia (Fig. 1A), than in those implanted as root (e.g., Edwards Prima, Medtronic Freestyle root, and Toronto SPV root valves), where the native coronary arteries are implanted high on the sinuses (Fig. 1B and 1C). One should always exercise caution when treating stentless valves, because the mortality associated with coronary obstruction is very high (14).
Three other issues need to be highlighted:
1. In day-to-day practice, a patient may present without details of the SHV implanted. In these patients, one can identify the type of SHV using the fluoroscopic guide, but it is not possible to identify the size of the SHV (13). Because various noninvasive methods, including fluoroscopy, echocardiography, and CT scan, may not accurately measure the true ID, one should use on-table balloon aortic valvuloplasty assessment if the size of the SHV is not known (20). We are currently investigating and characterizing each of the SHV designs to assess the impact of various noninvasive modalities on ID measurements.
2. Although ID is the most important measurement used to choose a size of THV when treating a native aortic valve and an SHV, the height of the SHV may also play an important role. This is relevant only for the Sapien XT THV because all sizes of CoreValve are taller than any size of SHV. It is not the stent height, but the leaflet height and type of leaflets (i.e., porcine or pericardial), that is important. Leaflet height of an SHV increases with its labeled size. Adequate leaflet cover is achieved in the majority of combinations with Sapien XT THV because for a larger SHV size, one chooses a large Sapien XT valve. This may, however, not be true for certain valve combinations such as Trifecta 21 and Sapien XT 23, where the respective heights are 16 and 14 mm. This may, theoretically, result in leaflet overhang. The hemodynamic relevance of this is, however, unknown. This needs further investigation and also highlights the complexities of a VIV procedure.
3. Although we have measured only the aortic SHV, the same principle should apply to SHV designed for mitral and tricuspid implantation. In the mitral position, however, a larger oversizing may be essential to prevent delayed embolization in the atrium (20,21) as a result of the larger closing pressures to which it is subjected. Also, larger oversizing may not be a problem in mitral implantations because the mitral SHV are usually of larger sizes, and there is no risk of coronary obstruction. We feel that when considering mitral VIV using Sapien/Sapien XT THV, one should use the stent ID rather than the true ID to achieve this oversize. This will lead to a conical/flared deployment, which prevents atrial migration (20).
We tested unused SHV, and hence, they lacked the calcification or thickening seen in degenerated SHV. This should be taken into account when deciding a size of THV, especially in an SHV with a borderline ID. We have observed heavy calcification in SHV implanted in tricuspid and pulmonary positions, and because the right-sided pressures are low, there could be an argument for using a smaller THV rather than larger in a borderline case. Another factor that can affect the true ID in a clinical scenario is pannus. However, both calcification and pannus can be easily evaluated by echocardiography and CT scan; at worst, the ID is going to be less than the true ID, but never more.
Design features of SHV may lead to a reduction in the stent ID in the majority of SHV. It is important to identify the SHV design so as to take this reduction into consideration when choosing the appropriate THV during a VIV procedure. The true ID chart in this paper provides realistic ID measurements for VIV treatment in the aortic position. One should use the true ID for selecting an appropriate THV device in the aortic position, but use the stent ID to choose the device for a mitral VIV procedure. If the SHV size is not known, balloon aortic valvuloplasty may be the best way to confirm the true ID, because ID measurements obtained by other modalities may be inaccurate.
The authors thank Miss Maya Guthrie and Mrs. Shalina Sunni for their help in obtaining fluoroscopic images required for the paper, and Mrs. Urmi Bapat for editorial assistance.
Mr. Bapat is a consultant to Edwards Lifesciences, Medtronic Inc., St. Jude Medical, and Symetis. Dr. Thomas is a consultant to Edwards Lifesciences, St. Jude Medical, and Mitralign. Dr. Attia has reported that he has no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- computed tomography
- internal diameter
- surgical heart valve(s)
- transcatheter aortic valve implantation
- transcatheter heart valve(s)
- Received July 16, 2013.
- Revision received October 15, 2013.
- Accepted October 24, 2013.
- American College of Cardiology Foundation
- Généreux P.,
- Head S.J.,
- Wood D.A.,
- et al.
- Gurvitch R.,
- Cheung A.,
- Ye J.,
- et al.
- Khawaja M.Z.,
- Haworth P.,
- Ghuran A.,
- et al.
- Piazza N.,
- Bleiziffer S.,
- Brockmann G.,
- et al.
- Gotzmann M.,
- Mugge A.,
- Bojara W.
- Ng A.C.,
- Delgado V.,
- van der Kley F.,
- et al.
- Dvir D.,
- Webb J.,
- Brecker S.,
- et al.
- Babaliaros V.C.,
- Jungadhwalla Z.,
- Lerakis S.,
- et al.
- Gonzalo B.C.,
- Jorge S.F.,
- Nicolas V.G.
- Chakravarty T.,
- Jilahawi H.,
- Nakamura M.,
- et al.
- Cheung A.,
- Webb J.G.,
- Barbanti M.,
- et al.