Table 2 Primary and secondary endpoints during the closed-loop and control therapy periods

Endpoint	Closed-loop (n = 26)	Control (n = 25)^a	Mean difference (95% CI for treatment difference^b)	P value
Primary endpoint
Proportion of time with glucose 3.9–10.0 mmol l⁻¹ (%)	66.3 (14.9)	32.3 (24.7)	35.3 (28.0 to 42.6)	<0.001
Key endpoints^c
Proportion of time with glucose > 10.0 mmol l⁻¹ (%)	33.2 (14.8)	67.0 (25.2)	−35.2 (−42.8 to −27.5)	<0.001
Mean glucose (mmol l⁻¹)	9.2 (1.2)	12.6 (3.0)	−3.6 (−4.6 to −2.5)	<0.001
HbA1c (mmol mol⁻¹)	57 (9)	72 (13)	−15 (−20 to −11)	<0.001
HbA1c (%)	7.3 (0.8)	8.7 (1.2)	−1.4 (−1.8 to −1.0)	<0.001
Proportion of time with glucose < 3.9 mmol l⁻¹ (%)	0.44 (0.19–0.81)	0.08 (0.00–1.05)	−0.10 (−0.36 to 0.16)	0.43
Secondary endpoints
Proportion of time with glucose > 16.7 mmol l⁻¹ (%)	1.8 (0.6–3.3)	12.5 (3.6–31.3)	NA	NA
Proportion of time with glucose > 20.0 mmol l⁻¹ (%)	0.2 (0.0–0.5)	3.2 (0.2–9.7)	NA	NA
Proportion of time with glucose < 3.0 mmol l⁻¹ (%)	0.04 (0.01–0.08)	0.03 (0.00–0.32)	NA	NA
s.d. of glucose (mmol l⁻¹)	3.0 (0.8)	3.4 (1.0)	NA	NA
Coefficient of variation of glucose (%)	32.2 (5.7)	27.7 (8.5)	NA	NA
Total daily insulin dose (U per day)	108 (73–188)	84 (54–129)	NA	NA
Total daily insulin dose (U kg⁻¹ per day)	0.90 (0.72–1.63)	0.71 (0.56–1.26)	NA	NA
Proportion of time with sensor glucose availability (%)	98.1 (96.8–98.5)	92.6 (89.8–98.0)	NA	NA
Proportion of time spent with closed-loop active (%)	92.3 (87.6–96.4)	NA	NA	NA

Data are presented as mean (s.d.) or median (IQR). Endpoints are calculated data from all randomized participants for whom at least 48 h of sensor glucose data were collected in at least one period. Glucose data are based on sensor glucose measurements. A two-sample t-test on paired differences was used to compare normally distributed variables. For highly skewed residuals for key and secondary endpoints, a transformation (windsorization) was used. All P values are two-tailed, and P < 0.05 was considered statistically significant. s.d. of glucose is a measure of glucose variability. NA, not applicable.
^aOne participant randomized to closed-loop therapy first did not continue to control therapy.
^bClosed-loop minus control.
^cTested in hierarchy to control the type 1 error using the fixed-sequence method. A gatekeeping strategy was used, in which the primary endpoint was tested first and, if passing the significance testing, other key endpoints were tested in order. If a nonsignificant result was encountered, formal statistical hypothesis testing was terminated, and analysis of any key endpoints below the one in question (any that were lower in the hierarchy) was considered exploratory.

Search