Page 324 - Glucose Monitoring Devices

Stopping the plunge: suspend before low systems 331

Similar to the European analysis of pump uploads, a retrospective review of
nearly 30,000 patients accounting for 758,382 patient-days when the LGS feature
was used were compared to 166,791 patient-days when the feature was not enabled.
On days when the LGS feature was on, there was a 69% reduction in sensor glucose
values under 50 mg/dL (2.7 mmol/L) [24]. Most suspend events occurred during the
night. At the time of insulin suspension, the median sensor glucose level was 60 mg/
dL (3.3 mmol/L) and rose to 87 mg/dL (4.8 mmol/L) 2 h after the suspension, and
further increased to 164 mg/dL (9.1 mmol/L) 4 h after basal insulin resumed [24].
Importantly, hyperglycemia exceeding 300 mg/dL was also reduced by 5.6%
when the LGS feature was used [24].
Some of the initial concern and delay in regulatory approval of LGS systems in the
United States stemmed from concerns over what would occur if insulin was suspended
based on inaccurate sensor glucose readings (i.e., when sensor glucose was not low).
To address this, a study assessed the effect of random 2-h suspensions in the overnight
period as long as blood glucose levels before bed were less than 300 mg/dL
(16.7 mmol/L) and blood b-hydroxybutyrate levels were less than 0.5 mmol/L.
Comparing over 100 nights of random insulin suspensions to control nights where
usual basal rates were maintained, results showed fasting glucose levels
were w50 mg/dL (2.7 mmol/L) higher but there were no clinically meaningful differ-
ences in blood b-hydroxybutyrate levels [25]. This supported the contention that LGS
could be safely employed in the face of inaccurate sensor glucose readings.

Stopping the plunge: suspend before low systems
Although reducing the severity and duration of hypoglycemia was an important first
step, the preference would be to prevent hypoglycemia altogether. To achieve this
aim, the suspension of insulin in response to a predicted low glucose value was
necessary. Danne and colleagues conducted a study using in silico modeling to
assess differences between responses of usual basal delivery without insulin suspen-
sions, LGS, and a predictive low glucose suspend (PLGS) system to challenges
intended to induce hypoglycemia. This demonstrated superiority of the PLGS
system [26]. Subsequently, the group studied participants using exercise to induce
hypoglycemia, and of the 15 evaluable episodes where the hypoglycemic threshold
was met and PLGS was triggered, 80% resulted in avoided hypoglycemia [26].
In-home testing of another PLGS system demonstrated efficacy in both pediatric
and adult populations by significantly reducing overnight hypoglycemia [27e29].
To date, two companies have marketed PLGS systems: Medtronic with its 640G
and 670G systems (Medtronic, Northridge, California) and Tandem with its t:slim
X2 with Basal IQ Technology (Tandem, San Diego, California). An in-clinic assess-
ment of the Medtronic 640G was performed by Buckingham and colleagues in
participants aged 14e75 years old. In this study, basal rates were increased in a
standardized fashion to induce nocturnal hypoglycemia with the PLGS feature set
at 65 mg/dL (3.6 mmol/L) [30]. The system is designed to interrupt basal insulin