BARCELONA – Use of a subcutaneous continuous glucose monitoring system translated into better glycemic control for critically ill patients in intensive care, a study has shown.
Compared with those monitored by point-of-care testing, patients with continuous monitoring spent significantly more time in their target blood glucose range, Daphne Boom reported at the annual meeting of the European Association for the Study of Diabetes.
The implantable device also markedly reduced the need for blood samples, said Ms. Boom, a medical student at the Onze Lieve Vrouwe Gasthuis, Amsterdam. Nurses obtained a mean of just two samples over a 24-hour period for these patients, compared with 12 over the same time for point-of-care patients.
Daphne Boom
Ms. Boom and her colleagues compared the two monitoring regimens in a group of 177 patients who were admitted to intensive care with an expected stay of at least 24 hours; 87 were randomized to continuous subcutaneous glucose monitoring and 90 to point-of-care monitoring. The target glucose range for each group was 5-9 mmol/L. The continuous monitoring system measured glucose every 10 minutes, and sounded an alarm whenever it registered outside this parameter. Blood samples were drawn every 2 hours for the point-of-care patients. A computer algorithm made insulin dosing recommendations for patients in both groups, based on the monitoring system level or the results of finger-stick samples.
In both groups, an arterial blood sample provided a reference glucose range every 6 hours; however, the samples were blinded to clinicians and used only for calibrating medications.
The primary endpoints were the incidence of severe hypoglycemia (below 2.2 mmol/L) and severe hyperglycemia (above 25 mmol/L). The measure of efficacy was the total time spent within the blood glucose target range. Nursing workload was assessed by the number of blood samples drawn every 24 hours.
Most patients were from the medical service; 13% were complicated cardiac surgery patients. The majority (92%) were mechanically ventilated. The investigators looked at 1,358 paired measurements.
There were five incidents of severe hypoglycemia in the intervention group and two in the control group, but the difference was not statistically significant. There were no cases of severe hyperglycemia in either group.
There was one case of mild hypoglycemia in each group. There was one case of mild hyperglycemia in the intervention group and two in the control group.
The mean study duration was 73 hours for the intervention group, of which 58 (79%) were spent within the target glucose range. The mean study duration for the control group was 59 hours, of which 43 (73%) were spent in the target range. This difference was statistically significant.
Significantly fewer blood samples were drawn from the intervention group (2 vs. 12 every 24 hours), showing that the continuous monitoring system was a more efficient use of nursing time, Ms. Boom added.
She said she had no relevant financial disclosures.