Assessment of Chronically Implanted Subcutaneous Glucose Sensors in Dogs: The Effect of Surrounding Fluid Masses

We developed a continuous, distributed-anode glucose sensor and now report on its use during subcutaneous implantation in dogs without diabetes. Using telemetry, we monitored sensor response to weekly administration of intravenous glucose. In a preliminary attempt to reduce fibrosis around the sensor, some sensors were designed to slowly release dexamethasone (DEX). Before implantation, in vitro sensor sensitivity was similar to values obtained after explanation (0.66 ± 0.09 vs 1.07 ± 0.19 nA/mM, n = 9, p = ns). Sensitivity in individual animals varied substantially over time. Average longevity of sensors was 32.1 ± 8.6 days. Device failure was caused by leakage of fluid into, or interruption of, circuitry. Lag time during glucose ascent averaged 5.8 ± 1.0 min. In devices that became surrounded by fluid masses, lag time during descent was greater than in devices without fluid (33.7 ± 4.5 vs 10.7 ± 1.6 min, p < 0.001). There was a nonsignificant tendency for longevity of the six sensors that contained DEX to be greater than the eight sensors without DEX (47.2 ± 18.7 vs 20.8 ± 3.6 days, p = ns). The development of fluid masses surrounding electrochemical glucose sensors prolongs lag time and probably contributes to the commonly observed instability of sensitivity over time. In future long-term implant studies, it is likely that avoidance of fluid masses will improve sensor function. ASAIO Journal 1999; 45:555–561.