Despite a considerable amount of research attributed to the development of an implantable glucose sensor, to date there is no clinically applicable concept for continuous glucose monitoring. Investigations to validate the subcutaneous tissue for continuous glucose sensing mostly comprised short-term implantations of glucose sensors. Most implanted glucose sensors showed a significant drift in sensitivity over the implantation period. This bioinstability was not to be expected from the in vitro performance of the sensors. In this paper, the influence of possible failure mechanisms on the poor in vivo performance of subcutaneously implanted glucose sensors is reviewed.