The cultivation of autologous endothelial cells on the blood surface of artificial hearts might prevent their detrimental thromboembolic complications. To investigate the growth characteristics of endothelial cells on theoretically suitable biomaterials, we compared three polyurethanes (Pellethane, Biomer, Enka) and three silicone rubbers (Elastosil, 3145 RTV, Medical Adhesive). All synthetic surfaces were precoated with an extracellular matrix (group 1), fibronectin (group 2), or a glutaraldehyde-preserved cellular matrix (group 3). After the seeding of 2.5 X 10(4)/cm2 human endothelial cells onto the various surfaces, primary adherence, growth kinetics, and maintenance of monolayer integrity were studied for 13 days. On the three polyurethanes all precoating procedures resulted in endothelial cell proliferation and the formation of persistent monolayers. In contrast, on silicone rubbers a persistent coverage with a confluent endothelium could be achieved only on the glutaraldehyde-preserved cellular matrix. When endothelial cell growth was quantitatively assessed on all precoating substrates, the glutaraldehyde-preserved cellular matrix proved to be far superior on each of the synthetics (p < 0.001). These results demonstrate the theoretical feasibility of endothelialization of artificial hearts in vitro. Provided such an endothelium can withstand the mechanical forces within an artificial heart, in vitro endothelialization might contribute to a regained attractiveness of the elective long-term implantation of artificial hearts.

The importance of initial human trials with autologous endothelial seeding lies not only in the implementation of a promising idea but also in the fact that canine data are only partially applicable to humans. The surface area of jugular veins in humans is much smaller than in dogs and considerably longer grafts are needed. Moreover, the reproductive capacity of adult human endothelial cells under in vivo conditions, which probably determines the success of seeding more than the seeding density, is also uncertain. Therefore the efficiency of autologous endothelial seeding in humans was investigated in 18 patients undergoing distal femoropopliteal bypass surgery. The average surface area of the jugular veins was 4.9 +/- 1.7 cm2 with an average cell yield of 32.6 +/- 18.0 x 10(4). The mean number of seeded cells per square centimeter of graft surface was 3.1 x 10(3). In a follow-up extending for 14 weeks, plasma levels of platelet factor 4 and beta-thromboglobulin as well as the platelet function in the whole blood aggregometer showed significantly better results in the seeded group. Plasma thromboxane B2, uptake and survival of indium 111-labeled platelets, and Doppler ultrasound investigations also favored the seeded group, but the results were statistically insignificant. No difference at all was found for the platelet dense granule compounds, releasable adenosine triphosphate and platelet serotonin. Thus our findings did not indicate the development of a closed endothelialized surface after 14 weeks, which is a period three times as long as the one required for confluent endothelial cell coverage in dogs.(ABSTRACT TRUNCATED AT 250 WORDS)