Close relationship between strict blood glucose control, including suppression of blood glucose variability, and mortality reduction in acutely ill patients with glucose intolerance investigated by means of a bedside-type artificial pancreas.

The purpose of this study was to elucidate the relationship between strict control of blood glucose (BG) and mortality reduction with the use of an artificial pancreas (AP). Patients were evaluated in the (1) early phase (E phase: mean 3.3 ± 2.6 days after admission, n = 84) and in the (2) late phase (L phase: mean 9.9 ± 3.3 days, n = 88), and were classified into a (1) group with a higher daily mean BG level (BGm) (BGmXa: BGm above Xmg/dl) and a (2) group with a lower BGm (BGmXb: BGm below Xmg/dl). Each group was classified into a (1) subgroup with a higher daily standard deviation of the BG levels (BGsd) and a (2) subgroup with a lower BGsd. In the E phase, the (1) mortality of the BGm200a group was significantly higher than that of the BGm200b group (56 vs. 29%, p < 0.05), and (2) in the BGm200b group, the mortality of the subgroup with a BGsd above 14 mg/dl was significantly higher than that with a BGsd below 14 mg/dl (46 vs. 17%, p < 0.025). In the L phase, the mortality of the BGm175a group was significantly higher than that of the BGm175b group (50 vs. 28%, p < 0.05). In conclusion, (1) a higher BGm and (2) higher BGsd in the E phase were prognostic risk factors. Based on the findings, it was considered that the target for BG control should be set at (1) BGm below 200 mg/dl and BGsd below 14 mg/dl in the E phase, and (2) BGm below 175 mg/dl in the L phase.