Customized closed-loop control boosts time in range in pregnant women with type 1 diabetes
Click Here to Manage Email Alerts
A small cohort of pregnant women with type 1 diabetes improved time in range with use of a closed-loop insulin delivery system with a zone model predictive control algorithm, according to findings from a small pilot study.
“Our study reports the first closed-loop control outcomes in pregnant women in the U.S. using a novel zone model predictive control design specifically customized to meet the continuous glucose monitoring time in range targets for pregnancy based on consensus guidelines,” Eyal Dassau, PhD, director of the biomedical systems engineering research group and senior research fellow in biomedical engineering at the Harvard John A. Paulson School of Engineering and Applied Sciences at Harvard University, and colleagues wrote in a study published in Diabetes Technology & Therapeutics. “Compared with the 1-week run-in, there was an increased percentage of time in the 63 mg/dL to 140 mg/dL range accompanied by decreased percentages above and below this range, during the supervised closed-loop control session. Moreover, nine of 11 participants had a greater time in range during closed-loop control.”
Researchers recruited 11 pregnant women with preexisting type 1 diabetes who used an insulin pump and were at 14 to 32 weeks gestation with a singleton pregnancy. Participants were recruited from three study sites from August 2020 to June 2021. During a 1-week run-in phase, participants used a Dexcom G6 CGM and their personal insulin pump at home. The run-in phase was followed by a 2-day supervised closed-loop control period where participants switched to an institution-supplied insulin pump (Tandem Diabetes Care). The CGM and insulin pump connected to an interoperable artificial pancreas system on an unlocked study phone. The system used the zone model predictive control algorithm with a tighter target range of 80 mg/dL to 110 mg/dL during the day and 80 mg/dL to 100 mg/dL from midnight to 4 a.m. Women chose meals from nearby restaurants and markets and ate them at the test site. Meal boluses were calculated based on each participant’s settings and the user-estimated carbohydrate intake. Self-monitored blood glucose testing took place 2 hours before and after each meal, and hypoglycemia treatments were initiated when glucose dropped below 63 mg/dL. Repeat measurements were taken every 15 minutes until the hypoglycemia was resolved.
The study cohort had an increase in time in range from 64% in the run-in phase to 81.5% during the supervised session (P = .007). Time in hyperglycemia with glucose higher than 140 mg/dL decreased from 30.8% during the run-in to 16.5% during the supervised session (P = .029) and time in hypoglycemia of less than 63 mg/dL dropped from 5.2% to 2% (P = .039). Glycemic variability was less during the supervised session compared with the run-in phase as measured by both standard deviation (31.2 mg/dL vs. 42.6 mg/dL; P = .017) and coefficient of variation (28% vs. 34.2%; P = .047).
Of 60 meals with data available, the mean carbohydrate content was 39.6 g. Mean estimated time in range was 86.4% for the first hour of the postprandial window and 80.4% for the first 2 hours.
Of all hypoglycemia treatments, 58% were required within 3 hours of a meal bolus and rescue carbohydrate intake ranged from 3 g to 16 g. The mean number of hypoglycemia treatments needed per participant was 4.7, and all treatments were self-administered by the participants. There were no hospitalizations, diabetic ketoacidosis episodes or severe hypoglycemia cases in the study.
“Our pilot study demonstrates that a customized closed-loop control system designed to specifically target glucose goals for pregnancy is feasible for women previously using insulin pumps and CGM,” the researchers wrote. “Further studies are needed to evaluate this system in a larger group of women and sustained use throughout pregnancy.”
Perspective
Back to Top
Elizabeth O. Buschur, MD, FACE
Glycemic management in pregnant women with type 1 diabetes is of utmost importance for maternal and fetal health. The goal of glycemic control in pregnancy includes near normal glucose with minimal hypoglycemia, and this is challenging for patients with type 1 diabetes. Furthermore, many women conceive while using continuous subcutaneous insulin infusion pumps and continuous glucose monitors that are not FDA-approved for use in pregnancy, and decisions need to be made whether to use these devices off-label in pregnancy.
The study by Dassau and colleagues evaluates the feasibility of closed-loop insulin delivery with a pregnancy-specific zone model predictive control algorithm in a small outpatient study of 11 pregnant women with type 1 diabetes. The study compares a 1-week run-in phase to a 2-day closed-loop control intervention. While overnight glucose control was similar between the run-in phase and the closed-loop control period, there was a significant increase in percentage of time in the pregnancy-specific time in range goal of 63 mg/dL to 140 mg/dL during supervised closed-loop control (81.5% vs. 64%; P = 0.007), less time spent with glucose above 140 mg/dL (16.5% vs. 30.8%; P = 0.029) and less time spent with glucose less than 63 mg/dL (2% vs. 5.2%; P = 0.039). There were no serious adverse events in either group.
This study is promising in that closed-loop control with a pregnancy-specific zone model predictive control algorithm improves time in range 63 mg/dL to 140 mg/dL. Activity and food choices were not restricted in this real-world study. However, one limitation of the study is that the participants had excellent glycemic control at enrollment (HbA1c 5.6 ± 0.5%). More research is needed on use of advancing diabetes technology, including closed-loop systems with customized targets, in pregnant women with type 1 diabetes.
Collapse
Click Here to Manage Email Alerts