Two of the 10 participants were female, mean age was 55.4 (44–70) years, and mean hemoglobin A1c was 8.0% (6.7–9.6%). Mean diabetes duration was 34.6 (18–50) years, and the mean duration of CSII use was 11.2 (5–25) years.
The venous glucose value before breakfast and before lunch was not different in OL compared to CL. The venous glucose before exercise was significantly lower in OL compared with CL ().
Primary and Secondary Outcome Measurementsa
One participant needed 1.5 IU intravenous insulin before OL started, and five participants needed intravenous insulin before CL started, with a mean administered dose of 2.7 (1.5–4.5) IU. One participant received 12 g oral carbohydrates before starting CL.
The outcomes measures are presented in and . Individual experiments are depicted in . The overall mean venous glucose in OL was 9 (5.4–13.5) mmol/liter versus overall mean venous glucose in CL, 8.7 (6.4–11.0) mmol/liter, p = .74. The postbreakfast glucose concentrations tended to be lower in OL than in CL: 9.5 (4.3–13.3) versus 11.4 (7–16.2) mmol/liter, p = .07. After lunch, the reverse was seen: 9.4 (6.0–14.9) versus 7.7 (5.5–9.0) mmol/liter, p = .15. No differences were seen in the postexercise glucose concentrations: OL 7.5 (4.6–13) versus CL 8.2 (5.5–13.1) mmol/liter, p = .45.
Mean venous glucose concentration of OL and CL. The insulin administration in OL and CL (IU) and glucagon administration (mg multiplied by 10) per 30 min appear in bars.
Individual experiments. Insulin concentration is 1 IU/ml. Glucagon concentration is 1 mg/ml.
The postprandial breakfast sensor AUC and the post-exercise sensor AUC were significantly lower in OL compared with CL. No differences were seen in post-prandial lunch sensor AUC. Likewise, no differences were seen in time spent in euglycemia, hypoglycemia, or hyperglycemia.
The overall insulin amounts administered and the amount of insulin administered in the postprandial breakfast period were significantly lower in OL than in CL, while this was equal for the postprandial lunch period. In those with baseline glucose below 7 mmol/liter (n = 4) before start of CL, no significant differences were seen in insulin administration between CL and OL (). In those with baseline glucose above 7 mmol/liter (n = 6), significant differences were seen in overall and postprandial breakfast insulin administration.
Primary and Secondary Outcome Measures in Subgroups with Glucose Concentration Below or Above 7 before Breakfast in Closed Loopa
No severe hypoglycemia occurred. During OL control, two hypoglycemic episodes occurred between exercise and lunch. In CL, there were four instances of a system-advised carbohydrate intake alert preventing hypoglycemia (counted as four hypoglycemia events in the outcome comparison) in three participants: two episodes in one participant were at the end of the exercise, one episode occurred between exercise and lunch, and one episode in the postprandial period after lunch.
Subcutaneous glucagon administration during CL was successful to prevent hypoglycemia in 10 of 14 instances; in 5 of these 10 instances, two glucagon boluses were needed to prevent hypoglycemia. In 2 of these 5 instances, a rescue glucagon bolus was used. When glucagon delivery was successful, the first changes in CGM glucose rate of change were seen at median time of 15 (15–20) min. The median time of the total effect of glucagon was 57.5 (50–60) min. The median effect on CGM glucose concentration was 0 mmol/liter (-1.0 to 3.0) and the median change in glucose slope was 6.6 (5.0–11.3) mmol/liter/h. Overall, the mean glucagon administration per subject in the postbreakfast period was 0.04 (0–0.13) mg, the postexercise period was 0.12 (0–0.36) mg, and the postprandial lunch was 0.07 (0–0.15) mg. No participant reported side effects after glucagon administration.
The overall sensor MAD in OL was 10.3% (4–26%) and during CL 15.3% (6.0–32%), p = .05. The median MAD in OL over the postprandial breakfast period was 8.5% (2.8–50%) versus 11% (3.2–29.6%) in CL, p = .96; over the postexercise period, the median MAD in OL was 10.1% (2.3–21.6%) compared with median MAD in CL 20.5% (5.4–39.7%), p = .013; and the median postprandial lunch MAD in OL was 8.4% (4.4–14.9%) versus median MAD in CL 12.8% (2–37.3%), p = .09. The mean heart rate during exercise was 99 (87–127) bpm.