Search

Heart valve substitutes of biological origin often fail by degenerative mechanisms. Many authors have hypothesized that mechanical fatigue and structural degradation are instrumental to in vivo failure. Since the properties of the structural matrix at implantation may predetermine failure, we have examined the ultrastructure, fracture, mechanics, and uniaxial high-strain-rate viscoelastic properties of: (1) fresh, (2) cryopreserved, and (3) cellular extracted porcine aortic valve leaflets. The cellular extraction process is being developed in order to reduce immunological attack and calcification. Cryopreservation causes cellular disruption and necrotic changes throughout the tissue, whereas extraction removes all cells and lipid membranes. Both processes leave an intact collagen and elastin structural matrix and preserve the high-strain-rate viscoelastic characteristics of the fresh leaflets. Extraction does cause a 20% reduction in the fracture tension and increases tissue extensibility, with the percent strain at fracture rising to 45.3 +/- 4 (mean +/- SEM) from 31.5 +/- 3 for fresh leaflets. However, extraction does preserve matrix structure and mechanics over the physiological loading range. Glutaraldehyde fixation produces increased extensibility, increased elastic behavior, and, when applied to extracted leaflets, it causes a marked drop in fracture tension, to 50% of that for fresh leaflets. The combination of extraction and fixation may lead to early degenerative failure. The cellular extraction technique alone may be a useful alternative to glutaraldehyde fixation in preparing bioprosthetic heart valves.

We have developed a multistep detergent-enzymatic extraction process (involving hypotonic and hypertonic solutions, the detergents octyl-phenoxy-polyethoxyethanol and sodium dodecyl sulfate, as well as DNAse and RNAse) which, while inhibiting autolysis, removes all cells from tissues and, with them, cellular antigens together with lipids and more soluble glycosaminoglycans. What remains is acellular matrix with the structural proteins well conserved and normally arranged. Canine arteries extracted to acellular matrix were implanted as coronary artery bypass allografts in a canine model, without the use of cardiopulmonary bypass, and compared with autogenous saphenous veins. Of nine pilot acellular matrix implants, four were patent, as compared with four of seven saphenous vein grafts. All occlusions in both graft types occurred acutely soon after implantation, with almost all patent grafts followed up for 6 months. The acellular matrix allografts showed no inflammation and only minimal cellular repopulation. This model needs further development, but appears promising for preclinical evaluation. Canine aortic and pulmonic valves extracted to acellular matrix using a modification of our extraction process, eliminating the detergent sodium dodecyl sulfate, were implanted heterotopically as allografts in the left main pulmonary artery in dogs, a location chosen to avoid the need for cardiopulmonary bypass. At 1 month, two-dimensional echocardiography of six implants showed leaflet motion and 3- to 5-mm Hg transvalvular gradients. Explant histology of four valves at 1 month showed no inflammation, cellular repopulation at the base of the valve, and partial endothelialization.(ABSTRACT TRUNCATED AT 250 WORDS)

We have developed an acellular matrix vascular prosthesis (AMVP) made by detergent and enzymatic extraction of natural arteries, yielding a tissue framework of collagen and elastin from the original vessel, with preservation of the natural basement membrane at the blood flow surface. These biografts have excellent handling characteristics and suturability, as well as low thromboreactivity. Whole vessel static testing of circumferential compliance (8.9 +/- 1 [SEM] X 10(-2)% mmHg at 100 mmHg) revealed behavior virtually identical to the paired natural vessel from which each AMVP was derived in nine canine carotid arteries. We implanted 16 canine-origin AMVPs into nine dogs (12 femoral and three carotid arteries, and one infrarenal aorta) with no antithrombotic drugs. Angiographic patency was maintained in 15 of 16 (one occlusion within 3 days) for follow-up from 3 days to 6 years, with no aneurysm formation in three AMVP at over 4 1/2 years. Explant analysis revealed preservation of AMVP elastica and collagen with no inflammation or dystropic calcification of the AMVP, and almost total thrombus free flow surfaces. These results suggest that allograft AMVPs could achieve long-term patency equivalent to saphenous veins.

We have examined the tensile viscoelastic properties of circumferential and radial strips of porcine aortic valve leaflets following fixation in glutaraldehyde and formaldehyde, with or without pressure. After aldehyde treatment, the radial strips remained weaker and less stiff than circumferential strips and responded slightly differently to the treatment. After fixation, with or without pressure, the radial strips showed large changes in stress-strain and hysteresis responses due to initial loading, and there was a twofold reduction in tensile strength and final stiffness. For strips in both directions, fixation without pressure produced doubled extensibility and a ramping stress-strain curve. Permanent (plastic) deformation of 5-20% occurred as a result of cyclic loading, stress relaxation, and creep experiments. Pressure fixation, however, produced little change in stress-strain results other than a simple shift to lower strain and produced no plasticity. Both methods of fixation reduced stress relaxation and creep. Mechanical test results are consistent with a loss of ground substance matrix during fixation. Reductions in tensile strength after fixation may be due to "riveting" of collagen geometry, producing local stress concentrations.

The tensile viscoelastic properties of radial and circumferential strips of fresh porcine aortic valve leaflets have been examined and compared with the fiber architecture of the leaflets as seen in the scanning electron microscope. The leaflet material showed nonlinear viscoelasticity, largely independent of strain rate, and dependent on the degree of preconditioning by cyclic loading. Preconditioning to a stable stress-strain response could only be accomplished on 1-cm-long strips above a minimum width of 4 mm (circumferential) and 6 mm (radial). Preconditioning yielded a more elastic and extensible material with reduced hysteresis. The leaflets were markedly anisotropic. Circumferential strips were up to 8 times stronger and stiffer than radial strips, and displayed greater stress relaxation and less creep. The circumferential mechanical properties were due to well-aligned circumferential collagen bundles reinforcing the composite structure, while radial properties were due to a more random collagenous support throughout the leaflet. Despite the presence of a septal shelf on the right coronary leaflet, no mechanical differences could be discerned between leaflets.