Table 2.

Primary end points

	Baseline (2 weeks)		Postrandomization (final 11 weeks)
	CLC (n = 63)†	SAP (n = 61)†	CLC (n = 64)†	SAP (n = 61)†	Difference (95% CI)‡	P value‡
Hours of sensor data, median (IQR)	304 (267, 323)	311 (294, 322)	1,656 (1,464, 1,781)	1,683 (1,565, 1,780)	NA	NA
CGM-measured % below 70 mg/dL (3.9 mmol/L),¥ median (IQR)	4.3 (1.6, 7.4)	3.6 (1.7, 5.9)	2.1 (1.2, 3.5)	3.0 (1.7, 5.7)
Mean ± SD	5.0 ± 4.2	4.7 ± 4.9	2.4 ± 1.7	4.0 ± 3.4	−1.7 (−2.4, −1.0)	<0.0001
CGM-measured % above 180 mg/dL (10 mmol/L),¥ mean ± SD	40 ± 17	43 ± 18	34 ± 11	39 ± 15	−3.0 (−6.1, 0.1)	<0.0001 NI*

NA, not applicable; NI, noninferiority.

^†Excludes one participant from the CLC group and one participant from the SAP group who dropped out postrandomization and did not have any data 3–13 weeks postrandomization to be included in analyses. The baseline CGM data for another CLC participant were unavailable, so their follow-up CGM data were included in the regression models using the direct likelihood method (16).

^‡Adjusted for baseline values, age, prior CGM use, and clinical center (random effects).

^¥Small imbalances by chance were observed between the two treatment groups at randomization for BMI and type 1 diabetes duration (see Table 1). Results for both primary outcomes were similar after additionally adjusting for baseline BMI and duration (data not shown).

^*P value for NI (prespecified NI limit = 5%).