Page 269 - Glucose Monitoring Devices

276 CHAPTER 14 Predictive low glucose suspend systems

abnormalities in children who have experienced severe hypoglycemia [11]. It can thus
be seen that trading hyperglycemia for hypoglycemia is not beneficial. The solution
must be achieving normoglycemia.
In addition to its negative physiological implications, hypoglycemia also contrib-
utes to fear of hypoglycemia (FOH). FOH tends to develop after an episode of severe
hypoglycemia or recurrent severe hypoglycemia and can be viewed as a sort of
phobia of hypoglycemia [12e15]. Patients with FOH experience increased distress
related to diabetes [16]. In addition, they tend to maintain higher glycemic control in
an effort to avoid subsequent severe hypoglycemic episodes [17,18]. FOH thus
contributes to both worsened glycemic control as well as increased distress in
patients with diabetes and their families.
It is here where incorporation of diabetes technologies to advance an automated
care becomes essential. Continuous subcutaneous insulin infusion (CSII) pumps
were first developed in the late 1970s and began to see routine patient use in the
late 1990s and early 2000s [19]. CSII pumps administer subcutaneous insulin at
smaller dosing intervals than those achieved by multiple daily injection (MDI)
therapy, deliver variable basal insulin rates throughout the day, and enable engaged
patients or caregivers to set temporary basal rates or suspend basal insulin to further
fine-tune insulin dosing. The use of insulin pumps has been associated with HbA1c
values that are 0.3%e0.5% below those of MDI users [20e22]. Although some
studies have demonstrated hypoglycemia reduction with CSII technology, its use
has not been consistently associated with reduced hypoglycemia [23e25].
CGM first became commercially available in the early 2000s, and its use has
grown exponentially over the past several years [20,26,27]. Early CGM devices
suffered from mean absolute relative difference (MARD) values that were signifi-
cantly worse than commercially available blood glucose meters [26,28,29]. These
devices also relied heavily on threshold alarms to alert users about impending
hypoglycemia [30]. Although the use of early generation CGMs was associated
with hypoglycemia reduction, the alarms contributed significantly to patient fatigue
and ultimately discontinuation of the devices [31e33]. Due to an increased arousal
threshold overnight, CGM alarms have also not been helpful at eliminating severe
nocturnal hypoglycemia. This is well highlighted in a case published by Bucking-
ham where an adolescent with T1D slept through over 2 h of alarms overnight before
a hypoglycemic seizure (Fig. 14.1)[34]. The information from CGM systems could
thus be seen to contribute to improved glycemic control and hypoglycemia reduction
but the burden of utilizing this information via direct self-management may have
been overwhelming for most patients.
Automating hypoglycemia reduction via predictive algorithms and predictive
low glucose suspend (PLGS) technology allows for hypoglycemia reduction without
creating an increased burden on patients. PLGS systems could be considered a form
of automated insulin delivery beyond simply suspending insulin below a certain
hypoglycemic thresholddas is done in low glucose suspend devicesdand before
increasing insulin delivery to minimize hyperglycemiadas is done with hybrid
closed-loop (or Artificial Pancreas) technology [35]. In the next sections, we will re-
view the development of PLGS systems and the results from their use in various clin-
ical trials and real-world settings.