Prevention of the Decrease in Sensitivity of an Amperometric Glucose Sensor in Undiluted Human Serum

The ultimate aim of glucose sensor construction is to provide an accurate means for the measurement of glucose concentrations in vivo as part of a closed-loop insulin delivery system. Various types of sensors (mainly those based on amperometric measurement of hydrogen peroxide generated by enzymatic oxidation of glucose) have been implanted into the subcutaneous tissue of animals (1)(2) and human subjects (3). The researchers conducting these experiments routinely found that sensors in subcutaneous tissue had much lower sensitivities to glucose than those in buffer solutions (3)(4).The causes for this loss of sensitivity in vivo are not known; however, four hypotheses are frequently put forward to explain this phenomenon (3)(4): (a) the concentrations of glucose and/or oxygen in the surrounding tissue are lower than expected; (b) proteins or other substances are adsorbed onto the outer surface of the sensor, hindering the diffusion of glucose and/or oxygen through the membrane (“membrane fouling”); (c) the activity of glucose oxidase is inhibited by an unknown, presumably low-molecular weight substance; and (d) detection of hydrogen peroxide by the platinum anode is inhibited (“electrode fouling”). Interestingly, a loss of sensitivity to glucose of similar magnitude is observed in human serum and in serum ultrafiltrate (3). This makes the first hypothesis unlikely, provided that identical mechanisms are responsible for the inactivation of sensors in serum and in subcutaneous tissue. If it were possible to prevent sensor inactivation with a membrane placed between the electrode surface and the enzyme layer, this would provide evidence that the second and third hypotheses are of minor importance and favor, at least indirectly, the importance of the fourth hypothesis. The aim of the present study was to examine the effect of such a membrane on the sensitivity to glucose of a sensor exposed to …