In this open-label multicentre randomised controlled trial, we investigated three peri-operative treatment strategies to lower glucose and reduce the need for rescue insulin in patients aged 18-75 years with type-2 diabetes mellitus undergoing non-cardiac surgery. Patients were randomly allocated using a web-based randomisation program to premedication with liraglutide (liraglutide group), glucose-insulin-potassium infusion (insulin infusion group) or insulin bolus regimen (insulin bolus group), targeting a glucose < 8.0 mmol.l . The primary outcome was the between group difference in median glucose levels 1 h after surgery. We analysed 150 patients (liraglutide group n = 44, insulin infusion group n = 53, insulin bolus group n = 53) according to the intention-to-treat principle. Median (IQR [range]) plasma glucose 1 h postoperatively was lower in the liraglutide group compared with the insulin infusion and insulin bolus groups (6.6 (5.6-7.7 [4.2-13.5]) mmol.l vs. 7.5 (6.4-8.3 [3.9-16.6]) mmol.l (p = 0.026) and 7.6 (6.4-8.9 [4.7-13.2]) mmol.l ) p = 0.006, respectively). The incidence of hypoglycaemia and postoperative complications did not differ between the groups. Six patients had pre-operative nausea in the liraglutide group, of which two had severe nausea, compared with no patients in the insulin infusion and insulin bolus groups (p = 0.007). The pre-operative administration of liraglutide stabilised peri-operative plasma glucose levels and reduced peri-operative insulin requirements, at the expense of increased pre-operative nausea rates.

Continuous Glucose Monitoring (CGM) systems could improve glycemic control in critically ill patients. We aimed to identify the evidence on the clinical benefits and accuracy of CGM systems in these patients. For this, we performed a systematic search in Ovid MEDLINE, from inception to 26 July 2016. Outcomes were efficacy, accuracy, safety, workload and costs. Our search retrieved 356 articles, of which 37 were included. Randomized controlled trials on efficacy were scarce (n = 5) and show methodological limitations. CGM with automated insulin infusion improved time in target and mean glucose in one trial and two trials showed a decrease in hypoglycemic episodes and time in hypoglycemia. Thirty-two articles assessed accuracy, which was overall moderate to good, the latter mainly with intravascular devices. Accuracy in critically ill children seemed lower than in adults. Adverse events were rare. One study investigated the effect on workload and cost, and showed a significant reduction in both. In conclusion, studies on the efficacy and accuracy were heterogeneous and difficult to compare. There was no consistent clinical benefit in the small number of studies available. Overall accuracy was moderate to good with some intravascular devices. CGM systems seemed however safe, and might positively affect workload and costs.

No abstract

No abstract