Table 2.
Primary, secondary, and exploratory outcomes
| Baseline/standard therapy† | 13-Week trial period | Treatment effect | ||||
|---|---|---|---|---|---|---|
| Outcome | Intervention group(n = 132‡) | Control group(n = 62‡) | Intervention group(n = 132‡) | Control group(n = 62‡) | Adjusted difference, intervention − control (95% CI) | P |
| Primary outcome: TIR 70–180 mg/dL, %§ | 43.9 ± 14.0, 44.2 [34.7–54.8] (n = 129) | 41.3 ± 14.6, 41.2 [31.4–52.2] (n = 61) | 61.2 ± 11.2, 62.3 [55.2–68.8] (n = 131) | 43.8 ± 14.5, 45.1 [34.3–53.6] (n = 62) | 17.5 (14.0, 21.1) | <0.0001 |
| Secondary outcomes in prespecified hierarchical orderǁ | ||||||
| Time below glucose range <54 mg/dL (noninferiority), %¶ | 0.32 ± 0.53, 0.13 [0.00–0.39] (n = 129) | 0.42 ± 0.91, 0.10 [0.00–0.43] (n = 61) | 0.23 ± 0.23, 0.17 [0.07–0.28] (n = 131) | 0.37 ± 0.53, 0.16 [0.06–0.45] (n = 62) | −0.05 (−0.11, 0.00) | 0.0501 |
| Time above glucose range >180 mg/dL, %§ | 54.4 ± 14.7, 54.7 [42.7–64.4] (n = 129) | 57.1 ± 15.5, 56.3 [46.1–67.3] (n = 61) | 37.6 ± 11.4, 36.3 [29.8–43.9] (n = 131) | 54.5 ± 15.4, 54.0 [42.3–64.4] (n = 62) | −16.8 (−20.8, −12.8) | <0.0001 |
| Mean sensor glucose, mg/dL§ | 200 ± 30, 198 [177–220] (n = 129) | 204 ± 31, 198 [181–225] (n = 61) | 174 ± 20, 170 [161–183] (n = 131) | 200 ± 30, 195 [175–217] (n = 62) | −26 (−34, −18) | <0.0001 |
| Change in HbA1c, %# | 8.48 ± 0.79, 8.40 [7.95–8.95] (n = 132) | 8.57 ± 0.95, 8.50 [8.00–9.10] (n = 62) | 7.25 ± 0.76, 7.10 [6.80–7.70] (n = 119) | 7.84 ± 0.83, 7.90 [7.30–8.30] (n = 55) | −0.58 (−0.79, −0.37) | <0.0001 |
| Change in HbA1c, mmol/mol# | 69 ± 8.6, 68 [63–74] (n = 132) | 70 ± 10.4, 69 [64–76] (n = 62) | 56 ± 8.3, 54 [51–61] (n = 119) | 62 ± 9.1, 63 [56–67] (n = 55) | −6.3 (−8.6, −4.0) | <0.0001 |
| Time below glucose range <70 mg/dL, %¶ | 1.66 ± 1.79, 1.07 [0.52–2.22] (n = 129) | 1.66 ± 2.25, 0.96 [0.23–2.07] (n = 61) | 1.18 ± 0.86, 1.04 [0.57–1.59] (n = 131) | 1.75 ± 1.68, 1.14 [0.58–2.60] (n = 62) | −0.36 (−0.61, −0.11) | 0.0050 |
| Change in T1-DDS total score# | 2.01 ± 0.69, 1.93 [1.46–2.36] (n = 132) | 2.27 ± 0.71, 2.21 [1.71–2.71] (n = 62) | 1.72 ± 0.63, 1.54 [1.29–2.00] (n = 130) | 2.08 ± 0.68, 1.91 [1.63–2.45] (n = 60) | −0.18 (−0.32, −0.05) | 0.0094 |
| Change in HCS total score# | 3.24 ± 0.54, 3.24 [2.89–3.72] (n = 132) | 3.13 ± 0.52, 3.11 [2.88–3.56] (n = 62) | 3.40 ± 0.50, 3.56 [3.00–3.78] (n = 131) | 3.14 ± 0.58, 3.12 [2.78–3.67] (n = 60) | 0.20 (0.06, 0.34) | 0.0048 |
| Change in DQOL-Brief total score# | 3.75 ± 0.47, 3.73 [3.43–4.07] (n = 132) | 3.62 ± 0.49, 3.60 [3.33–3.93] (n = 62) | 4.11 ± 0.47, 4.17 [3.87–4.47] (n = 130) | 3.60 ± 0.47, 3.57 [3.24–4.00] (n = 60) | 0.43 (0.31, 0.55) | <0.0001 |
| T1-DDS–proportion of participants with a clinically meaningful change (≥0.19), % ** | — | — | 53.8 (45.3–62.4) | 45.0 (32.4–57.6) | 24.3 (6.0, 44.1) | 0.0145 |
| DQOL-Brief–proportion of participants with a clinically meaningful change (≥0.238), % ** | — | — | 59.2 (50.8–67.7) | 21.7 (11.2–32.1) | 52.7 (36.2, 67.9) | <0.0001 |
| HCS–proportion of participants with a clinically meaningful result (final score ≥3), %** | — | — | 81.7 (75.1–88.3) | 61.7 (49.4–74.0) | 18.9 (4.5, 34.7) | 0.0076 |
| Exploratory outcomes†† | ||||||
| Time above glucose range >250 mg/dL, %¶ | 25.4 ± 13.8, 24.1 [14.0–35.3] (n = 129) | 27.6 ± 14.6, 24.4 [15.7–37.0] (n = 61) | 13.5 ± 8.2, 11.6 [7.9–17.9] (n = 131) | 25.7 ± 14.3, 23.4 [14.4–33.4] (n = 62) | −10.8 (−12.8, −8.9) | <0.0001 |
| Time in tight glucose range 70–140 mg/dL, %§,‡‡ | 24.6 ± 9.8, 24.2 [17.7–31.3] (n = 129) | 23.2 ± 10.4, 22.2 [16.6, 30.6] (n = 61) | 36.3 ± 10.3, 36.3 [29.5–43.9] (n = 131) | 24.8 ± 10.5, 23.5 [16.4–33.1] (n = 62) | 11.5 (8.4, 14.5) | <0.0001 |
| Glycemia risk index§,‡‡ | 67 ± 20, 67 [52–83] (n = 129) | 71 ± 20, 75 [54–85] (n = 61) | 44 ± 15, 41 [33–52] (n = 131) | 68 ± 20, 67 [54–83] (n = 62) | −24 (−29, −19) | <0.0001 |
| Coefficient of variation of sensor glucose, %§ | 37.2 ± 5.3, 37.7 [33.5–40.6] (n = 129) | 36.8 ± 5.9, 36.8 [32.6–39.7] (n = 61) | 36.3 ± 4.1, 36.1 [33.5–39.1] (n = 131) | 37.3 ± 5.2, 37.7 [33.9–40.6] (n = 62) | −1.1 (−2.6, 0.4) | 0.1516 |
| TIR 70–180 mg/dL, daytime (0600–0000 h), %§ | 45.5 ± 14.8, 45.6 [35.5–57.4] (n = 129) | 42.1 ± 15.3, 41.6 [30.9–53.0] (n = 61) | 60.5 ± 11.9, 62.0 [53.8–67.8] (n = 131) | 44.4 ± 15.2, 44.8 [34.4–56.2] (n = 62) | 16.1 (12.3, 19.9) | <0.0001 |
| TIR 70–180 mg/dL, nighttime (0000–0600 h), %§ | 39.2 ± 16.0, 41.4 [27.0, 50.9] (n = 129) | 38.8 ± 17.4, 39.0 [26.2–49.1] (n = 61) | 63.3 ± 13.4, 65.1 [54.4–72.9] (n = 131) | 41.8 ± 14.7, 40.3 [31.2–53.7] (n = 62) | 21.7 (17.8, 25.5) | <0.0001 |
| BMI (kg/m2)¶ | 26.6 ± 4.8, 25.6 [23.2–29.3] (n = 132) | 25.4 ± 4.1, 24.5 [22.2–28.1] (n = 62) | 27.2 ± 5.2, 26.4 [23.5–30.2] (n = 126) | 25.7 ± 3.9, 24.7 [22.8–27.7] (n = 56) | 0.25 (−0.03, 0.52) | 0.0782 |
| Total daily insulin use (units/kg)# | 0.63 ± 0.20, 0.61 [0.50–0.74] (n = 130) | 0.59 ± 0.17, 0.54 [0.46–0.65] (n = 58) | 0.58 ± 0.19, 0.56 [0.46–0.68] (n = 132) | 0.57 ± 0.16, 0.56 [0.46–0.65] (n = 57) | −0.03 (−0.06, 0.00) | 0.0789 |
Data are mean ± SD and median [interquartile range] unless otherwise indicated. To convert the values for glucose to mmol/L, multiply by 0.05551. Analyses conducted on a modified intention-to-treat data set.
†Baseline and follow-up (end-of-trial) data were used for change in HbA1c, T1-DDS score, HCS score, DQOL-Brief score, BMI, and total daily insulin. The remaining outcomes compared the standard therapy phase with the trial period.
‡Three participants were randomized to the intervention and one participant randomized to control prior to complete collection of standard therapy CGM data.
§P value determined using repeated-measures linear mixed-effects model with treatment group, visit, treatment group-by-visit interaction, age, sex, and duration of diagnosis as fixed effects and country and site as random effects.
ǁA hierarchical approach was used to control the type I error. Hypothesis testing for secondary outcomes was performed sequentially in the order listed in the table. When a P ≥0.05 was observed, the outcomes below that finding on the list were not formally tested.
¶P value determined using robust regression model using M-estimation with Huber weight function with treatment group, age, sex, baseline value of the outcome, and duration of diagnosis as fixed effects. Missing data were handled using multiple imputation.
#P value determined using a linear mixed-effects model with treatment group, age, sex, duration of diagnosis, and baseline value of the outcome as fixed effects and country and site as random effects.
**P value determined using a logistic mixed-effects model with treatment group, age, sex, duration of diagnosis, and continuous baseline value of the outcome as fixed effects and country and site as random effects.
††There was no formal hypothesis testing for exploratory outcomes.
‡‡Post hoc analysis.