Calcification characteristics of porcine stentless valves in juvenile sheep

Abstract

Objective: To compare calcification characteristics of two porcine stentless valves (Toronto SPV and Freestyle) with different designs, fixation and antimineralization techniques using a juvenile sheep model of valve implantation inside the circulation. Methods: The stentless valves (n=2×6) were implanted in juvenile sheep in the pulmonary artery as an interposition, while the circulation was maintained with a right ventricular assist device. The model was validated by the implantation of, clinically well-known, porcine (Hancock II) and pericardial (Pericarbon) valves. Half of the valves were explanted after 3 months, the rest after 6 months. Valves were examined macroscopically, by X-ray, light microscopy (HE, Masson, Von Giesson, Von Kossa, PTAH stains), and transmission electron microscopy. Quantitative determination of the calcium content of the cusps was performed with atomic absorption spectrometry. Results: After 3 months, the Freestyle had an extensively calcified aortic wall, most prominent at the outflow side of the porcine valve. After 6 months, calcification increased transmurally, but the valve cusps were free of calcification, and the inflow side was only slightly calcified. The Toronto SPV valve also started to calcify at the inflow side of the valve after 3 months with increased calcification after 6 months. The base of the Toronto SPV valve cusps showed slight calcification after 6 months of implantation. Conclusions: The pattern of calcification of the porcine aortic wall differs between the two studied stentless valves, with calcification located predominantly at the outflow side in the Freestyle valve, but also at the inflow side in the Toronto SPV valve. The cusps of the Freestyle valve were less prone to calcification than those from the Toronto SPV valve.

Experimental, Juvenile sheep, Stentless aortic valve, Bioprosthesis, Cardiac surgery, Calcification

Introduction

Stentless bioprosthetic valves potentially have important advantages due to the low gradients over the valve and the fact that formal anticoagulation is not necessary, enabling the patient to regain a fully active life within a short time period and causing fast regression of the left ventricular hypertrophy [1]. The durability of freehand-sewn aortic valve homografts used for aortic valve replacement in humans is greater than for stented aortic homografts [2]. In analogy with this, it is expected that the durability of stentless heterografts will be superior to that of its stented counterparts. This has been shown experimentally by Hazekamp et al. [3], with the stented Intact bioprosthesis (Medtronic, Irvine, CA), and its clinically unavailable stentless counterpart in growing pigs.

Thus far, only the first medium- and long-term studies are being published concerning the clinical results of stentless aortic bioprosthetic xenograft implantations [4]. Experimentally, juvenile sheep are a known model for bioprosthetic valve implantation, since they resemble the human situation concerning annular sizes, heart rate, cardiac output and intracardiac pressures. Most important, however, is the fact that bioprostheses implanted for a few months in juvenile sheep show changes comparable with those that take several years to develop in bioprostheses implanted in patients [5]. Bioprosthetic valve implantation in aortic [6], left ventricle apico-aortic conduit [7], mitral [8], or tricuspid [5],[8] position in juvenile sheep is most frequently used to test the durability and calcification characteristics of bioprosthetic valves. These models, however, are complicated, because full extracorporeal circulation is necessary in most models, or important blood loss occurs, with a mortality rate far over 50% [7],[8]. The extracorporeal circulation, the long-term housing of the sheep for such experiments, and the high mortality make these models very expensive. Probably for this reason, no experimental data are available at present comparing degeneration in several stentless bioprosthetic valves directly. Therefore, we decided to build an easier and cheaper model of in vivo bioprosthetic valve testing.

Several types of stentless valves are clinically available. They differ considerably in terms of design, composition of materials, tissue preservation and anticalcification treatment. We have chosen to compare the calcification characteristics in two clinically widely available stentless aortic porcine valves, the Toronto SPV valve (St. Jude Medical, St. Paul, MN), and the Medtronic Freestyle valve (Medtronic, Irvine, CA).

Materials and methods

All animals received humane care in compliance with the European Convention on Animal Care. The study was approved by the ethics committee of the Katholieke Universiteit Leuven. Juvenile sheep (less than 6 months old) breeded specifically for this purpose were selected.

Valves studied

Two clinically available stentless valves were selected for this study, the Toronto SPV valve (St. Jude Medical, St. Paul, MN) and the Freestyle valve (Medtronic, Irvine, CA). Both are porcine aortic roots, fixed with glutaraldehyde, but the Toronto SPV valve has its three sinuses scalloped, whereas, the Freestyle is an intact root with both coronary stumps ligated. Furthermore, the outside of the Toronto SPV is fully covered with Dacron, whereas, the Freestyle valve is only partially covered at the base of the root over the muscle bar that is most prominent near the right coronary cusp. The Freestyle valve is fixed without transvalvular gradient, with the aortic root under 40 mmHg, the Toronto SPV under low pressure (about 2 mmHg). The concentration of the glutaraldehyde used for fixation is different. The Freestyle valve is treated after fixation with alpha aminooleic acid (AOA), an agent known to reduce calcification [7],[9].

Implantation

The sheep were fasted for 48 h. The animals were premedicated with ketamine (10–20 mg/kg intramuscularly). Anesthesia was induced with increasing concentrations of halothane in oxygen. Albipen LA (15 mg/kg, Mycofarm, Brussels, Belgium) was administered intramuscularly for antibiotic profylaxis. Anesthesia was maintained with halothane and N₂O. Fentanyl (Janssen, Beerse, Belgium) was administered in boluses as necessary. After endotracheal intubation, mechanical ventilation was instituted. All ventilation parameters were adjusted to keep the arterial blood gasses and pH within the physiological range. The left chest was shaved, prepped and draped, and a left thoracotomy was performed in the second interspace. After administration of 100 mg of lidocaine (Xylocard, Astra, Södertälje, Sweden) intravenously, the pericardium was incised taking care not to damage the phrenic nerve, and the heart suspended in a pericardial cradle. The main pulmonary artery was completely isolated. After administration of 3 mg/kg heparin (Novo, Bagsvaerd, Denmark) intravenously, a pneumatic right ventricular assist system (Medos HIA-VAD 54 ml ventricle, Medos-Helmholtz Institute, Aachen, Germany) was installed with the inflow cannula in the right atrium and the outflow cannula 1 cm before the pulmonary bifurcation. The pulmonary artery was clamped immediately above the pulmonary valve. A second clamp was placed immediately proximal to the outflow cannula. In the mean time, the chosen valve was rinsed as prescribed in the manual. The valves were implanted as an interposition with running 5/0 polypropylene sutures. After removal of the clamps, the native pulmonary valve was destroyed by tearing two cusps with a clamp introduced through a purse-string suture placed at the sinuses, and afterwards the Medos system was stopped. Careful hemostasis was performed. The chest was closed in layers with a chest drain in the left pleural space. After waking up, the animal was extubated and brought to the recovery room. Feeding was allowed immediately. Intravenous fluid administration was stopped after 2 h. The chest drain was removed after 6 h. The animals received analgesics (piritramide, Dipidolor, Janssen, Beerse, Belgium) for the first 2 days on regular schemes and diuretics, as necessary. Albipen LA and low molecular weight heparin (enoxaparine, 20 mg twice daily, Clexane, Rhône–Poulenc Rorer, Brussels, Belgium) were administered for 6 days. Afterwards, the sheep returned to the controlled animal facility where the general health of the sheep was checked daily.

Explantation and analysis

Half of the valves were explanted after 3 months, the other half after 6 months. Sheep were premedicated and anesthetized in the way described before. The left thoracotomy was reopened and the heart dissected free. Heparin 3 mg/kg was administered, and after exsanguination, the valve was excised together with a proximal and distal part of the sheep pulmonary artery.

Macroscopical examination

Valves were grossly inspected and color photographs were taken. Special attention was paid to retraction of the cusps, or any deformed or indurated parts of the valve. Afterwards the valve was longitudinally transected through the commissures. Each of the three specimens thus includes a pre- and postvalvular part of the sheep pulmonary artery, together with a part of the porcine aortic wall (wall of the stentless valve), and respectively, a right coronary cusp (RCC), a non-coronary cusp (NCC), or a left coronary cusp (LCC). Color pictures were taken again.

X-ray assessment

X-ray examination (face, profile) was performed under mammography conditions to demonstrate and localize macroscopical calcification.

Histology

For histology, a longitudinal section of the specimen through the middle of the LCC was embedded in paraffin. Four-micrometer thick sections were routinely stained with hematoxylin and eosin (HE), Masson's trichrome stain for collagen, an elastic Von Giesson stain, a phosphotungstic-acid-hematoxylin (PTAH) for fibrin, and a Von Kossa calcium staining.

Transmission electron microscopy

For transmission electron microscopy, the RCC was divided in a basal part, a middle part, and the free edge of the valve. Also the aorta inflow and outflow were sampled. From each of the five regions, three to ten samples (<1 mm) were embedded in Epon. The samples were not taken from parts with massive calcification, since this yielded, on the photographs, only large black areas without additional information. One micrometer-thick sections were stained with Toluidine blue and examined by light microscopy. From each of the three cusp fragments an area of the outflow side, the inflow side and the middle (deep) part was dot marked. By analogy, from the aortic inflow and outflow, also the intimal inner medial and outer medial wall and adventitia were thus sampled. Ultra-thin sections were cut, stained with uranylacetate and lead citrate. Sections were treated with 2% potassium pyroantimonate to demonstrate calcium. Grids were examined in a Philips CM 10 electron microscope. Random photographs were taken.

Quantitative calcium determination

Half of every segment was used for quantitative calcium determination. The cusps were divided in three parts: commissural area, basal part and free edge. After lyophilization, the tissue was pulverized, and desiccated to constant weight in an oven. Hydrolysates were made in 6 N HCl. Calcium content was measured by flame atomic absorption spectrometry, and expressed as mg/mg of dry cuspal weight.

Data management and statistical analysis

Quantitative data were expressed as median (range), since the data did not follow a normal distribution. Comparisons were made with the Mann–Whitney U-test for comparison between the two groups, and the Kruskal–Wallis ANOVA when more than two groups were compared. The level for statistical significance was put at 0.05. Data management and statistical analysis were done with Statistica 4.5 (Statsoft, Tulsa, OK).

Results

Two sheep died during or shortly after the implantation procedure. One died during the operation due to a tear in the pulmonary artery extending distal to the outflow cannula, with rapid exsanguination of the sheep. Another sheep died due to respiratory insufficiency within 6 h after the operation. One sheep died 20 days postoperatively due to endocarditis of the tricuspid valve and the bioprosthetic valve in pulmonary position. One sheep died 28 days postoperatively, with necropsy showing pneumonia without evidence of endocarditis. These four sheep were replaced by new experiments and excluded from the study. Global mortality was thus 25%.

Macroscopical examination

The explanted Freestyle valves showed massive calcification of the aortic wall, forming an entirely rigid tube already after 3 months, that remained macroscopically unchanged after 6 months. All cusps after 3 months were nicely pliable, without macroscopical signs of calcification. Occasionally, slight fibrous sheathing was seen near the commissures and at the base of the cusps after 3 months. After 6 months the fibrous sheathing increased, somewhat, at the inflow side of the cusps, however never complete fibrous overgrowth encapsulating the cusps was seen. All cusps were functioning.

The explanted Toronto SPV valves showed also extensive calcification of the porcine aortic wall. A comparable fibrous reaction as in the Freestyle valve was also seen in the Toronto SPV. Contrary to the situation in the Freestyle valves, some punctiform scattered calcification was seen near the commissures, and slight induration at the base of the cusps. This increased after 6 months.

X-ray examination

Already after 3 months, massive calcification of the Freestyle aortic wall was visible. This calcification was extensive, in large homogenous plaques in the part of the aortic wall above the cusps (outflow part). Only minor calcification could be seen in the inflow part of the aortic wall in one valve, whereas, in the two other valves it was completely absent. After 6 months, calcification in the outflow part of the Freestyle aortic wall became more dense. In two valves, minor calcification was seen in the inflow part. In two Freestyle valves, calcification at the inflow suture line was seen after 6 months (Fig. 1 ). No calcification of the cusps was visible at any time.

Fig. 1

A representative example is shown of an X-ray taken of a Freestyle valve explanted after 6 months. Large, continuous calcification of the outflow part of the valve is clearly visible. Some calcification is also seen at the inflow suture line.

In the Toronto SPV valve, calcification of the outflow part was also seen in large plaques. Remarkable was the much more extensive calcification however in the inflow part of the Toronto SPV when compared with the Freestyle valve (Fig. 2 ). This consisted of several types of calcification. A plaque of calcification was seen under the right coronary cusp, sometimes extending in the adjacent part of the inflow portion of the aortic wall under the left and non-coronary cusp. A second was extensive calcification of the suture line proximally, but also fully around the valve, marking the line of attachment of the Dacron fabric to the valve. A third kind was scattered irregular calcification, that occurred to a variable extent in all locations in the inflow part of the valve. Some very slight calcification of the base of the cusp was seen after 3 months, and became more evident after 6 months. No calcification was seen at any time in the center or at the free edge of the cusp.

Fig. 2

A representative example is shown of an X-ray taken of a Toronto SPV valve explanted after 3 months. Calcification of the outflow part is seen. Remarkably when compared with Fig. 1 is the extensive calcification of the inflow part, consisting of calcification of the muscle bar, the suture lines, and some additional irregular calcification.

Light microscopy

In the Freestyle valves after 3 months of implantation, severe calcification was visible in the outflow part of the aortic wall in all sections. This was most pronounced in the inner layers of the media. After 6 months, calcification increased and extended also more in the outer layers of the media (Fig. 3 ). Only in one valve some calcification was visible in the inflow part of the Freestyle valve. It was located then completely at the outside of the porcine aortic wall, very close to the cloth covering. No calcification was ever visible in the cusp. The cloth covering gave rise to a foreign body reaction with fibrosis and accumulation of giant cells. This foreign body reaction was equally present after 3 and 6 months. Thin fibrous tissue covered the inside of the aortic wall both in the inflow and outflow part. It also covered the inflow side of the base of the cusp. The thin fibrous tissue covering the inflow part had a higher cellularity than the fibrous tissue of the outflow part. The layer became thinner after 6 months when compared with the situation after 3 months.

Fig. 3

A representative example is shown of the light microscopical appearance (H and E staining, overview) of a Freestyle valve explanted after 6 months. Severe, almost transmural calcification is seen in the outflow part of the valve. No calcification is seen in the inflow part. Note also the well preserved wavy appearance of the fibrosa of the leaflet.

Different observations were made in the Toronto SPV valve. Calcification was visible in all valves both after 3 and 6 months of implantation in the outflow part, although somewhat less extensive than seen in the Freestyle valve, but also clearly in the inflow part of the aortic wall (Fig. 4 ). Here it was often located in the inner part of the wall. Furthermore, more calcification was seen close to the cloth covering which covers the entire outer surface of the valve. The foreign body reaction was comparable with the one seen in the Freestyle valve. The overgrowth with fibrous tissue was somewhat thicker on the Toronto SPV than on the Freestyle valve, and covered more surface of the cusp, often more than half of the outflow surface. Also in the Toronto SPV, this layer was thinner after 6 months when compared with the situation after 3 months.

Fig. 4

A representative example is shown of the light microscopical appearance (H and E staining, overview) of a Toronto SPV valve explanted after 6 months. Calcification is visible at the inflow and outflow side of the aortic wall. Note also the calcification associated with the cloth covering. Some calcification is also visible at the base of the leaflet. Remark the well preserved wavy appearance of the fibrosa of the leaflet.

Electron microscopy

Electron microscopical pictures were taken in parts that were not massively calcified, since this yielded only massive black areas without additional information. With electron microscopy, calcification was clearly seen in the inflow part of the aortic wall of the Freestyle valve after 3 months, especially in the muscle bar, within the sarcomeres, in and around the mitochondria and the nucleus, in areas without macroscopical calcification. In the outflow part of the aortic wall, calcification was evident around collagen and elastic fibers. In the cusps, calcification was barely visible (Fig. 5 ). Some slight calcification was seen intracellularly, and rare dots of calcification between the collagen fibers. The collagen fibers were well preserved, with a wavy appearance of the collagen bundles. No important qualitative differences were seen between valves implanted for 3 or 6 months.

Fig. 5

A representative example is shown of the transmission electron microscopical appearance of the basal part of a cusp of a Freestyle valve explanted after 3 months. Well arranged bundles of collagen fibers are seen. No signs of calcification.

In the Toronto SPV valves, the findings were very similar qualitatively concerning calcification of the aortic wall. Somewhat more calcification was seen in the cusps of the Toronto SPV when compared with the Freestyle valves in cellular elements, such as mitochondria, and also between the collagen bundles (Fig. 6 ).

Fig. 6

A representative example is shown of the transmission electron microscopical appearance of the basal part of a cusp of a Toronto SPV valve explanted after 3 months. Well arranged bundles of collagen fibers are seen. Scattered calcification is seen throughout the sample.

Calcium content

No statistically significant differences existed in calcification between the subsegments of the cusps, so that these subsegments were pooled for the rest of the analysis. Median calcium content in the cusp was 1.19 (0.005–17.04) μg/mg dry cuspal weight after 3 months in Toronto SPV valves versus 0.50 (0.005–7.36) μg/mg in the Freestyle valves (P=0.046). After 6 months this was 2.13 (0.005–96.36) versus 1.00 (0.125–33.36) μg/mg dry cuspal weight in Toronto SPV versus Freestyle valves, respectively (P=0.021).

Discussion

The first aim of this study was to build an easier and (thus) cheaper model to study bioprosthetic valves inside the circulation in juvenile sheep. Such a model should allow the scientific community to test, more thoroughly, new bioprostheses before the first clinical experiments, and thus avoiding potential clinical catastrophes. The model presented here is certainly cheaper and easier than the previously described ones [3],[5],[6],[7]. The need to establish full extracorporeal circulation with an oxygenator in the circuit is obviated. The 25% mortality in this series was much lower than that reported in other series, reaching values of over 50% [7],[8]. This is probably caused by the less invasive procedure, which can be performed through a small thoracotomy of about 10 cm, with minimal blood loss, with minimal hemodilution due to the low priming volume of the right ventricular mechanical assist system, and performed while the lungs are continuously ventilated and perfused. In more recent series (unpublished) mortality went further down to about 15%.

With this model, valves are implanted in right-sided, pulmonary position with evidently lower closing pressures and flow velocities over the tested valve than in left-sided position. The frequency of valve opening is certainly equal and the range of excursion of the cusps is expected to be nearly the same. The changed hemodynamic load might alter the rate and pattern of calcification and degeneration of the bioprostheses. Thiene et al. [8], however, were unable to find a significantly different rate of calcification when bioprostheses were implanted in tricuspid or mitral position. Alterations of bioprosthetic valves implanted in tricuspid position in juvenile sheep are clinically and pathologically very similar to those occurring in human beings [5]. To validate the model of implantation in pulmonary position further, two series of stented bioprosthetic valves with well-known clinical behavior were implanted (Hancock II, Medtronic, Minneapolis, MN, n=4; and Pericarbon, Sorin, Saluggia, Italy, n=4) for 3 and 6 months. Progressive degeneration of these valves took place with a pattern of calcification mimicking the clinical findings. Calcification was especially visible in the aortic wall component of the Hancock II valve, and in the basal parts of the cusps in both the Hancock II and Pericarbon, but in time progressing towards the middle of the cusps. A perforation in a cusp of a Hancock II was found immediately adjacent to a macroscopically indurated zone after 6 months of implantation.

The severe fibrous overgrowth, called fibrous sheathing [5],[8],[10],[11], encountered in bioprosthetic valves in sheep implanted in tricuspid position, causing retraction and sometimes even complete immobility of the cusps, was also seen in pulmonary position after implantation of Hancock II or stented Pericarbon valves (unpublished). This was, however, minimal in our series of right-sided implanted Freestyle and Toronto SPV valves. This sheathing never caused restriction in movements of the cusps. This might indicate a very low thrombogenicity of these valves, since it is believed that sheathing originates from fibrin deposition and thrombus organization [8].

In the literature, no detailed experimental data were available concerning calcification of the Toronto SPV valves. The initial report by David et al. [12] concerning the predecessor of the Toronto SPV valve (with a somewhat different fixation and anti-mineralization technology, and without cloth covering) implanted in aortic position in sheep for up to 6 months, stated that all leaflets were mobile without any sign of calcification macroscopically or by histology. No information is given concerning calcification of the aortic wall. In our model, calcification is seen in the aortic wall, but also in the base of the leaflets. No final explanation for the difference with the findings of David et al., can be given, but it might be the fact that he used somewhat older sheep that is responsible for this. Furthermore, it might be that addition of the cloth covering in the Toronto SPV, or the differences in fixation technology are responsible for this. An argument that favors this explanation is that the calcification in our model is often pronounced close to the cloth covering.

More data are available concerning the fate of Freestyle valves after implantation in juvenile sheep. In an apico-aortic conduit in seven juvenile sheep [7], all valve leaflets were soft and pliable after about 4 months without histological signs of important calcification, while the aortic wall was severely calcified. This is concordant with our findings. Unexplained, is the reason why the valve cusps in the control group of that series [7] were severely calcified, in the light of our results with the Toronto SPV valves or the findings of David et al. [12]. In aortic position in two growing pigs, comparable findings were reported for the Freestyle valve [13], with calcification of the aortic wall (although less extensive than seen in our juvenile sheep model), and without calcification of the cusps.

Our study is the first reported study to compare the durability of two clinically available stentless valves directly in the same experimental model. The outflow part of the aortic wall is severely calcified in both valves. The inflow part is considerably more calcified in the Toronto SPV valve when compared with the Freestyle valve. Calcification of the cusps is limited in both valve types, but slightly more in the Toronto valve. AOA treatment, less cloth covering, and/or differences in fixation techniques could be responsible for this.

The severely calcified aortic wall is not without importance. A stentless valve becomes in fact a stented valve again, with its implications concerning the dynamical behavior, and incomplete opening of the valve. This might have implications on transvalvular gradients and regression of ventricular hypertrophy. Furthermore, calcification of the aortic wall might change the stress load on the cusps, with in turn its implications for calcification and degeneration [14]. Finally, would reoperation become necessary, extensive calcification of the aortic wall part, might technically complicate the operation.

We conclude that (1) this model is promising for preclinical evaluation of bioprosthetic heart valves, and (2) calcification is less in the valve cusps and the inflow part of the aortic wall in the Freestyle valve than in the Toronto SPV valve, but extensive calcification in the outflow part of the aortic wall remains problematic in both valves.

Appendix A. Conference discussion

Dr J.M. Hasenkam (Aarhus, Denmark): I noticed you did the implantations in the pulmonary artery where the hemodynamic load is obviously very low. Could you speculate about the impact of the higher hemodynamic load on the left side of the circulation?

Dr Herijgers: Right-sided implantation was a prerequisite to operate without any extracorporeal circulation. The oxygenator is a major problem in experiments with sheep. This causes a high mortality. So we wanted to avoid this. This is why we chose the pulmonary artery. We destroyed the native pulmonary valve to have a full load-bearing from the right side on these implanted valves.

There is a nice study published by Professor Thiene from Padua who compared the calcification in Pericarbon valves in tricuspid and mitral positions, stented valves, and he showed that there was no significant difference in the rate of calcification between the two implantation sites. If you see the absolute values in his publication, calcification is about two-thirds in the right-sided position than in the left side. So, probably pressure load has some effect but the difference was not statistically significant. The major problem in the right side is sheathing; this is why most of the implantations are done on the left side. We did not see this in our experiments, at least not with the Toronto or the Freestyle valve, but we see this also with stented valves in the right side, also in the pulmonary position.