Closed-loop system offers better control than pump with sensor, six-month trial finds

A closed-loop insulin delivery system appeared to offer better control than a sensor-augmented pump over six months in patients with type 1 diabetes, according to a recent randomized trial.

Researchers randomly assigned patients from seven U.S. university centers in a 2:1 ratio to treatment with a hybrid closed-loop system or to a control group, which received treatment with a sensor-augmented pump. The hybrid system used an algorithm with dedicated hypoglycemia safety module, automated correction boluses, and an intensified program of nocturnal insulin delivery to achieve near-normal glucose levels in the morning. Before assignment, all patients participated in a two- to eight-week run-in phase during which they were trained to use the devices and baseline data were collected.

The study's primary outcome was the percentage of time with blood glucose level in the target range, defined as 70 to 180 mg/dL (3.9 to 10.0 mmol/L) measured by continuous glucose monitoring. The main secondary outcomes were percentage of time that the glucose level exceeded 180 mg/dL (>10.0 mmol/L), mean glucose level, HbA1c level, and percentage of time that the glucose level was below 70 mg/dL (<3.9 mmol/L) or below 54 mg/dL (<3.0 mmol/L). The study was supported by a grant from the National Institute of Diabetes and Digestive and Kidney Diseases. Most of the authors reported relationships with industry, and Tandem Diabetes Care provided the closed-loop systems and technical support, as well as reviewed the manuscript. Results of the study were published Oct. 16 by the New England Journal of Medicine.

One hundred sixty-eight patients were randomly assigned to treatment between July 12 and Oct. 9, 2018. One hundred twelve were assigned to the closed-loop group, and 56 were assigned to the control group. Patients ranged in age from 14 to 71 years, HbA1c level ranged from 5.4% to 10.6%, and diabetes duration ranged from 1 to 62 years. Fifty percent of patients were women. Over the six-month period, mean (±SD) percentage of time spent in the target glucose range increased from 61%±17% to 71%±12% in the closed-loop group but remained the same in the control group (59%±14%). The mean adjusted difference was 11 percentage points (95% CI, 9 to 14 percentage points; P<0.001).

All of the main secondary outcomes met the prespecified hierarchical criterion for significance and favored the closed-loop system. The mean between-group difference in percentage of time that blood glucose levels were below 70 mg/dL (<3.9 mmol/L) was −0.88 percentage point (95% CI, −1.19 to -0.57 percentage points; P<0.001), and the mean adjusted difference in HbA1c level after the six-month study period was -0.33 percentage point (95% CI, −0.53 to −0.13 percentage point; P=0.001). Sixteen patients in the closed-loop group reported 17 adverse events, and two patients in the control group reported two adverse events. No patients developed severe hypoglycemia. One episode of diabetic ketoacidosis was seen in the closed-loop group due to pump infusion failure.

The authors pointed out that the closed-loop group had more unscheduled contacts with clinicians than the control group (68 vs. 13) and that the insulin pumps used in the control group were not able to suspend insulin use due to predicted hypoglycemia. However, they concluded that in their study, the closed-loop system increased the time spent in the target glucose range over six months and resulted in fewer episodes of hyperglycemia and hypoglycemia and better HbA1c levels versus a sensor-augmented pump. The results should be interpreted in the context of these specific patient and setting characteristics, the authors noted.

An accompanying editorial called the trial results “impressive and clinically relevant,” especially since the study population included patients of different ages and with different levels of diabetes control, but agreed that it had limitations. Barriers to a fully automated closed-loop system include slow subcutaneous insulin absorption and insufficient sensor accuracy, and whether such systems can be used in higher-risk patients is still uncertain, among other issues, the editorialist said. “It is clear that patients would appreciate wearing devices that require minimal interaction, leading to a more carefree lifestyle,” the editorialist wrote. “We are not there yet, but [this trial] offers an almost fingerstick-free option, providing a big step toward a brighter future for patients.”