Clinical evaluation of blood glucose variability with continuous intravenous measurement in ICU patients (GLIC-study)

Critically ill patients frequently develop hyperglycemia, independent of a
history of diabetes. This concept of hyperglycemia (stress diabetes) as an
adaptive mechanism to stress in critical illness has been accepted for years.

Since 2001 studies showed that targeting plasma glucose levels of 4.4-6.1
mmol/L compared to acceptance of hyperglycemia significantly decreased the
mortality and morbidity in post-operative cardiosurgical patients and reduced
morbidity in medical ICU patients.
These Intensive Insulin Therapy (IIT) studies were performed by infusing
intravenous insulin continuously. The small target range was successfully
reached and maintained by a huge increase in intermittent glucose measurements
and the use of extensive and difficult insulin dosing-protocols.
Despite the protocol and frequent glucose measurements in the first IIT-study,
there was a significant increase in severe hypoglycaemic events (0.8% of the
patients in the control group versus 5.1% in the IIT group developed a glucose
< 2.2 mmol/l). Much worse was the incidence of severe hypoglycaemia in medical
IIT treated ICU-patients (18.7% of the patients).
The occurrence of hypoglycaemia in critically ill patients is associated with a
higher mortality.
The incidence of severe hypoglycaemic events during IIT could probably be
reduced by increasing the frequency of glucose measurement during insulin
therapy and / or by adjusting the target range. Recent recommendations on
measurement of blood glucose and glycemic control in critically ill patients
advise a suboptimal range of 4.0-10.0 mmol/L.
However, increasing the target range gives more opportunity to increase the
variation in glucose levels. Recent studies show that increased glycemic
variability, defined as a coefficient of variation (CV) of >20%, has a strong
independent association with increased risk of mortality in patients without
diabetes.

The current IIT-protocols in several ICUs prescribe continuous administration
of insulin with intermittent measurement of glucose with an interval between
glucose measurements up to four hours. Each patient, independent of a medical
history of diabetes and independent of admission type, is treated following the
same protocol. Several studies have shown that mortality in patients treated
with IIT is higher in patients with a medical history of diabetes mellitus and
in septic patients. Therefore the target range can differ between patients
admitted to the ICU.

The GlucoClear glucose-sensor is an accurate glucose-oxidase coated sensor
developed for continuous intravenous glucose measurement in ICU-patients.
Recently (May 2013) the ICU nurses and doctors of the St Antonius Hospital
Nieuwegein, The Netherlands, obtained experience with this glucose-sensor (GLIC
1-study). Its accuracy is already tested and CE marking is obtained.

Subsequently (October 2013), in the context of a study, we monitored 25 ICU
patients with the continuous intravenous glucose sensor, during IIT. The
patients (post-operative cardiothoracic surgical-, septic- and medium care
patients) were treated following the local IIT protocol (target range for
glucose 4.4-6.1 mmol/L) with intermittent arterial point of care (POC) blood
glucose measurements. The monitor was blinded and the data were analysed after
completion of the study (GLIC 2-study).

The mean blood glucose value was 6.3 mmol/l and the mean blood glucose CV was
22.9%. 96% of the patients had blood glucose levels outside the target-range
with a duration of 63% of the observed period, 32% of the patients had severe
hyperglycemia (glucose > 10 mmol/L) and 44 % of the patients had hypoglycaemia
(glucose < 3.9mmol/L). Despite extensive experience with IIT in critically ill
patients, it remains difficult to maintain the blood glucose levels in these
patients in the target-range. Almost all patients experience blood glucose
levels both sides outside the target range with harmful blood glucose
fluctuations.

Up till now, no study investigated the clinical implementation of ITT with
continuous intravenous glucose monitoring.

Our hypothesis is that continuous intravenous glucose monitoring is helpful in
ICU patients treated with IIT to:

a) reach and maintain the glucose-target of 4.4-6.1 mmol/L sooner
b) to increase safety, by eliminating the development of hypoglycaemia
c) reducing the blood glucose CV, especially in high risk groups like sepsis
patients and patients recovering on the ICU or the medium care unit (MCU) who
still are insulin-resistant and start to eat
d) to reduce workload for blood glucose sampling by the nursing staff.

Future continuous glucose measurements may reduce the workload associated with
frequent conventional glucose measurements. Moreover, major inaccuracy in Point
Of Care (POC, arterial blood sample) measurements can be attributed to
pre-analytical (sampling and handling of samples) variables which are expected
to be irrelevant when using the more accurate GlucoClear sensor.

The patients will probably reach the target range sooner while measuring the
blood glucose levels continuously compared to intermittend glucose
measurements. Especially in hyperglycaemic keto-acidotic diabetic patients,
continuous glucose measurement can be a valuable tool to reach normoglycemia
soon and safe, without development of severe hypoglycaemia.
Additional questions to be answered could be if there is a difference in
glucose variability between diabetic and non-diabetic patients and if this
could be related to postoperative morbidity or mortality. Future research of
fully automated closed loop insulin dosing guidelines, to reach the blood
glucose target range and without potentially harmful hypoglycemia, will be
challenging and help to improve patient care.

The objective of this study is to improve the implementation of IIT in
critically ill patients using continuous intravenous blood glucose monitoring.
Improvement is defined by preventing hypoglycaemia (glucose < 3.9 mmol/L),
reducing the blood glucose CV to less than 20% and reducing the percentage of
patients experiencing hyperglycemia after the target-range of 4.4-6.1 mmol/L is
reached.

Non-randomized, prospective, observational study to evaluate the added value of
continuous intravenous glucose measurement by GlucoClear.
No extra blood samples will be collected. Values of the blinded glucose trend
will be compared with the local IIT protocol glucose samples afterwards.
During the non-blinded part of the study, comparison with POC measurements is
no longer required.
Glucoclear measures automatically continuously (5 minutes-interval) maximally
72 hours.

Patients admitted to the ICU with severe sepsis: After admission to the ICU and
blood glucose levels are twice below 4.4 mmol/L or above 6.1 mmol/L the patient
receives a peripheral intravenous catheter. Subsequently, the sensor for
continuous glucose measurement will be placed into the catheter. Subsequently,
1 or 2 IE/ hour of insulin infusion will be started intravenously using
another venous access. The continuous glucose measurement starts after
positioning the sensor and calibration time of 20 minutes and will be continued
for at least 8 and maximally 72 hours in the ICU. For patients who at the ICU
and start to eat or are transported to the MCU and start to eat within 72 hours
after placement of the GlucoGlear, the GlucoClear remains attached and the
measurements and insulin infusion rate will be performed.

Patients admitted to the ICU after cardiothoracic surgery: Patients undergoing
a cardiothoracic surgical procedure and are expected to stay in the ICU
post-operatively for at least two days will be included in the study.
Postoperatively the GlucoClear will be attached and calibrated.
The target range is supposed to be 4.4-6.1 mmol/L. After surgery the patient
will be transported to the ICU and intensive insulin therapy will be
implemented. In the ICU insulin will be administered continuously, without
administration of a bolus.

Patients who are not yet included for GlucoClear which are transported to the
MCU and expected to stay at the MCU for three days and develop hypoglycaemia
(glucose < 4.4 mmol/L) or hyperglycemia (glucose > 6.1 mmol/L) twice within 30
minutes are included for continuous glucose monitoring with GlucoClear and
insulin treatment.

Every 5 minutes a 20 microliter blood sample will be obtained automatically
into the catheter, blood glucose level measured by the sensor and the sampled
blood will be infused back automatically intravenously through the same
catheter. The obtained measurements by GlucoClear will be visible on the
monitor immediately for nurses and medical staff. The local IIT-protocol with
intermittent measurements is a nurse driven protocol.

Therefore, the study IIT-protocol with continuous measurements will also be
nurse-driven. The nurses are instructed to evaluate the glucose-trend each 20
minutes (after 4 measurements) to decide if insulin dose has to be changed (the
obtained blood glucose level and the change of insulin-infusion rate is
recorded in Metavision and filled in a form for later analysis. The nurses are
allowed to change the insulin-infusion rate for reaching the target range, by
changes of 0.1-2.0 IE themselves. The nurse is not allowed to administrate a
bolus of insulin. In situations where the glucose-levels are extreme high or
when there is no response, the intensivist will be consulted.

The main goal is to reach the target-range of 4.4 - 6.1 mmol/L as soon as
possible and to maintain the glucose level in the target-range by changing the
insulin-infusion rate. If the medical staff is not confident with the sensor
readings or there are too much missing values, blood glucose measurements by
POC or bloodgas (both bloodsamples from the arterial line) are allowed. If the
sensor is not measuring correct or there is a dysfunction of the sensor which
is not recovered within one hour, a new sensor will be placed or the
observation will be aborted and the former IIT-protocol with intermittent
measurements will be followed.

The study ends when a patient is transported to the ward, or after 72 hours, or
when the GlucoClear is not measuring continuously (missing values) or if
continuous monitoring is still not possible after replacement of the sensor
and/or catheter.

All study patients have an extra peripheral venous catheter for continuous
glucose measurement.

Local phlebitis of the intravenous catheter for the sensor for the continuous
glucose measurements could be an adverse event. In that case we finish the
measurements and the catheter will be removed. The risks associated with this
study are not greater than the inherent risks associated with the procedure
involving the introduction of any venous catheter.