Comparison of a needle-type and a microdialysis continuous glucose monitor in type 1 diabetic patients.

OBJECTIVE We examined the reliability of two continuous glucose sensors in type 1 diabetic patients at night and during rapid glucose excursions and verified the hypothesized nocturnal hypoglycemic drift of the needle-type sensor (CGMSgold) and delay of the microdialysis sensor (GlucoDay). RESEARCH DESIGN AND METHODS Blood was sampled overnight twice per hour in 13 patients. Rapid-acting insulin was given subcutaneously 30 min after breakfast. Sampling once per minute started 45 min after breakfast and 75 min after insulin injection for 30 min, with the aim of determining peak and nadir glucose values. Mean absolute differences (MADs) between sensor and blood glucose values were calculated. Sensor curves were modeled for all patients using linear regression. Horizontal and vertical shifts of sensor curves from the blood glucose curves were assessed. A vertical shift indicates sensor drift and a horizontal shift sensor delay. RESULTS Drift was minimal in the needle-type and microdialysis sensors (-0.02 and -0.04 mmol/l). Mean +/- SD delay was 7.1 +/- 5.5 min for the microdialysis sensor (P < 0.001). MAD was 15.0% for the needle-type sensor and 13.6% for the microdialysis sensor (P = 0.013). After correction for the 7-min delay, the microdialysis MAD improved to 11.7% (P < 0.0001). CONCLUSIONS The microdialysis sensor was more accurate than the needle-type sensor, with or without correction for a 7-min delay. In contrast to the previous version, the current needle-type sensor did not exhibit nocturnal hypoglycemic drift. Continuous subcutaneous glucose sensors are valuable adjunctive tools for glucose trend analyses. However, considering the large MADs, individual sensor values should be interpreted with caution.