The long-term biodegradation of various polyurethanes with and without surface modifications was evaluated by implanting small porous filamentous patches of these materials subcutaneously in the backs of dogs for one month. Data were compared to those obtained with spun polyurethane vascular grafts of similar materials implanted in the aorto-iliac position in dogs. The extremely high surface area of approximately 7 m2/cm3 of these porous filamentous patches provided numerous sites for surface cracking and the very fine filaments (10 microns in diameter) provided an easily identifiable structure to study the cracking phenomenon. Results from numerous one month implants clearly demonstrated that the subcutaneous implant model effectively reproduced the biodegradation behavior observed in vascular graft implants. The degradation was most pronounced in the softer Shore 80A polyurethanes and less pronounced in the harder 55D and 75D polyurethanes. The degradation could not simply be stopped by stress annealing the polyurethane and the degradation did not require the presence of metallic ions. Antioxidants, surface adsorbed albumin, poly(2-hydroxyethyl-methacrylate) grafting, silicone copolymerization, tetrafluoroethylene plasma discharge and the addition of urea linkages to the polymer were also shown to be ineffective in stopping the biodegradation process. In contrast, covalent bonding or grafting of silicone polymer to the surface of the urethane successfully inhibited the biodegradation process.