A Study of Insulin Efsitora Alfa (LY3209590) Compared With Insulin Degludec in Participants With Type 1 Diabetes Treated With Multiple Daily Injection Therapy

HbA1c is the glycosylated fraction of hemoglobin A. It is measured primarily to identify the average plasma glucose concentration over prolonged periods of time.

Least Squares (LS) mean was determined using Analysis of Covariance (ANCOVA) model with treatment + country + CGM use prior to study entry [yes/no]+ carbohydrate counting for prandial insulin[yes/no] + baseline value of the dependent variable (Type III sum of squares) as variables. Missing data at Week 26 were imputed by return-to-baseline multiple imputations approach.

HbA1c is the glycosylated fraction of hemoglobin A. It is measured primarily to identify the average plasma glucose concentration over prolonged periods of time.

LS mean was determined using ANCOVA model with treatment + country + CGM use prior to study entry [yes/no] + carbohydrate counting for prandial insulin[yes/no] + baseline value of the dependent variable (Type III sum of squares) as variables. Missing data at Week 26 were imputed by return-to-baseline multiple imputations approach.

Percentage of time in glucose range between 70 and 180 milligram per deciliter (mg/dL) [3.9 and 10.0 millimole per liter (mmol/L)], measured by continued glucose monitoring (CGM) from week 23 to week 26 inclusive over a 24-Hour Period. The time component of the time-in-range statistic was calculated using the display time recorded by the CGM device.

LS mean was determined using ANCOVA model with treatment + country + CGM use prior to study entry [yes/no] + carbohydrate counting for prandial insulin dosing [yes/no] + Hemoglobin A1c Stratum at Baseline and baseline value of the dependent variable (Type III sum of squares) as variables. Missing data at Week 23-26 were imputed by return-to-baseline multiple imputations approach

The event rate of participant-reported clinically significant nocturnal hypoglycemia (defined as blood glucose level <54 mg/dL (3.0 mmol/L) or severe hypoglycemia and occurs at night and presumably during sleep between midnight and 6:00 AM), measured during treatment phase up to week 52.

Group mean was reported and determined by Negative binomial model using Number of episodes = Baseline hypoglycemia rate + HbA1c at Baseline (%) + Treatment, with log (exposure in days/365.25) as an offset variable.

HbA1c is the glycosylated fraction of hemoglobin A. It is measured primarily to identify the average plasma glucose concentration over prolonged periods of time.

LS mean was determined using ANCOVA model with treatment + country + CGM use prior to study entry [yes/no]+ carbohydrate counting for prandial insulin[yes/no] + baseline value of the dependent variable (Type III sum of squares) as variables. Missing data at Week 52 were imputed by return-to-baseline multiple imputations approach

Change from baseline in fasting blood glucose measured by self-monitoring blood glucose (SMBG).

LS mean was determined using ANCOVA model with treatment, country, CGM use prior to study entry [yes/no], carbohydrate counting for prandial insulin dosing [yes/no]) and baseline value of the dependent variable as variables. Missing data at Baseline were imputed with multiple imputation with assumption of missing at random. Missing data were imputed by return-to-baseline multiple imputations approach.

Glucose variability measured as coefficient of variation (CV) for blood glucose during the CGM session over a 24-hour period, between Week 23 to Week 26 and Week 49 to Week 52 was reported.

LS mean was determined using mixed model repeated measures (MMRM) model with BASELINE + Country + Prior CGM use + Carbohydrate counting for Prandial Dose + Hemoglobin A1c Stratum at Baseline + Treatment + Time + Treatment*Time (Type III sum of squares) as variables. Unstructured variance-covariance structure was used.

Percentage of time spent within the blood glucose range of 70 to 180 mg/dL [3.9 to 10.0 mmol/L], as measured during the CGM session over a 24-hour period, from Week 49 to Week 52.

LS mean was determined using ANCOVA model with treatment, country, CGM use prior to study entry [yes/no], carbohydrate counting for prandial insulin dosing [yes/no]) and Hemoglobin A1c Stratum at baseline and baseline value of the dependent variable as variables. Missing data at Baseline were imputed with multiple imputation with assumption of missing at random. Missing data were imputed by return-to-baseline multiple imputations approach.

The insulin dose was recorded daily or weekly in an electronic diary. The average weekly basal insulin dose at Week 26 and Week 52 was reported.

LS mean was determined using MMRM model with BASELINE + Hemoglobin A1c Stratum at Baseline + Country + Prior CGM use + Carbohydrate counting for Prandial Dose + Treatment + Time + Treatment*Time (Type III sum of squares) as variables. Variance-covariance structure was set as compound symmetry.

The insulin dose was recorded daily or weekly in an electronic diary. The average daily bolus insulin dose at Week 26 and Week 52 was reported.

LS mean was determined using MMRM model with BASELINE + Hemoglobin A1c Stratum at Baseline + Country + Prior CGM use + Carbohydrate counting for Prandial Dose + Treatment + Time + Treatment*Time (Type III sum of squares) as variables. Variance-covariance structure was set as compound symmetry.

The average weekly total insulin dose is the sum of the average weekly basal and bolus doses. LS mean was determined using MMRM model with BASELINE + Hemoglobin A1c Stratum at Baseline + Country + Prior CGM use + Carbohydrate counting for Prandial Dose + Treatment + Time + Treatment*Time (Type III sum of squares) as variables. Variance-covariance structure was set as compound symmetry.

The basal/total insulin dose ratio is the average weekly basal dose divided by the average weekly total dose.

LS mean was determined using MMRM model with BASELINE + Hemoglobin A1c Stratum at Baseline + Country + Prior CGM use + Carbohydrate counting for Prandial Dose + Treatment + Time + Treatment*Time (Type III sum of squares) as variables. Variance-covariance structure was set as compound symmetry.

Rate of Composite Level 2 and 3 Hypoglycemia Events were reported. Hypoglycemia with glucose <54 mg/dL (Level 2) or Severe Hypoglycemia confirmed by the investigator to be an event that required assistance for treatment (Level 3) was reported. A severe hypoglycemic event is characterized by altered mental or physical status requiring assistance of another person to actively administer carbohydrate, glucagon, or other resuscitative actions for the treatment of hypoglycemia.

Group mean was reported and determined by Negative binomial model using Number of episodes = Baseline hypoglycemia rate + HbA1c at Baseline (%) + Treatment, with log (exposure in days/365.25) as an offset variable.

Change from baseline in body weight was reported. LS Mean was determined by MMRM model using Baseline + Treatment + Time + Treatment*Time (Type III sum of squares) as variables. Variance-covariance structure was set as compound symmetry.

Percentage of time spent in the hypoglycemia range with blood glucose < 54 mg/dL (3.0 mmol/L), as measured during the CGM session over a 24-hour period from week 23 to week 26 and week 49 to week 52 was reported. LS Mean was determined by ANCOVA model using treatment + country + CGM use prior to study entry [yes/no] + carbohydrate counting for prandial insulin dosing [yes/no] + Hemoglobin A1c Stratum at Baseline and baseline value of the dependent variable (Type III sum of squares). as variables. Missing data at Baseline were imputed with multiple imputation with assumption of missing at random. Missing data at week 23-26 and week 49-52 were imputed by return-to-baseline multiple imputations approach.

Percentage of time spent in the hyperglycemia range with blood glucose greater than (>) 180 mg/dL (10.0 mmol/L), as measured during the CGM session over a 24-hour period from Week 23 to Week 26 was reported.

LS Mean was determined by ANCOVA model using treatment + country + CGM use prior to study entry [yes/no] + carbohydrate counting for prandial insulin dosing [yes/no] + Hemoglobin A1c Stratum at Baseline and baseline value of the dependent variable (Type III sum of squares) as variables. Missing data at Baseline were imputed with multiple imputation with assumption of missing at random. Missing data at week 23-week 26 were imputed by return-to-baseline multiple imputations approach.

Percentage of time spent in the hyperglycemia range with blood glucose greater than (>) 180 mg/dL (10.0 mmol/L), as measured during the CGM session over a 24-hour period from Week 49 to Week 52 was reported.

LS Mean was determined by ANCOVA model using treatment + country + CGM use prior to study entry [yes/no] + carbohydrate counting for prandial insulin dosing [yes/no] + Hemoglobin A1c Stratum at Baseline and baseline value of the dependent variable (Type III sum of squares) as variables. Missing data at Baseline were imputed with multiple imputation with assumption of missing at random. Missing data at week 49-week 52 were imputed by return-to-baseline multiple imputations approach.

The Diabetes Treatment Satisfaction Questionnaire change (DTSQc) score is used to assess relative change in participant satisfaction from baseline. The questionnaire consists of 8 items, 6 of which measure Treatment Satisfaction, dealing with: 1. satisfaction with current treatment; 2. convenience of the treatment; 3. flexibility; 4. satisfaction with own understanding of participant's diabetes; 5. how likely to recommend their present treatment; and 6. how satisfied to continue with their present treatment. Each item is rated on a 7-point Likert scale (which ranges from -3 (much less satisfied) to +3 (much more satisfied). The scores from the 6 treatment satisfaction items are summed to a Total Treatment Satisfaction Score, which ranges from -18 (much less satisfied) to +18 (much more satisfied). Higher the score the greater the improvement in satisfaction with treatment. The lower the score the greater the deterioration in satisfaction with treatment.

The DTSQc score is used to assess relative change in participant satisfaction from baseline. The questionnaire consists of 8 items, 6 of which measure Treatment Satisfaction, dealing with: 1. satisfaction with current treatment; 2. convenience of the treatment; 3. flexibility; 4. satisfaction with own understanding of participant's diabetes; 5. how likely to recommend their present treatment; and 6. how satisfied to continue with their present treatment. Each item is rated on a 7-point Likert scale (which ranges from -3 (much less satisfied) to +3 (much more satisfied). The scores from the 6 treatment satisfaction items are summed to a Total Treatment Satisfaction Score, which ranges from -18 (much less satisfied) to +18 (much more satisfied). Higher the score the greater the improvement in satisfaction with treatment. The lower the score the greater the deterioration in satisfaction with treatment.

The SF-36v2 is a participant-reported measure designed to assess health status using 36 items across 8 domains: Physical Functioning, Role-Physical, Bodily Pain, General Health, Vitality, Social Functioning, Role-Emotional, and Mental Health. The 8 health domains are aggregated into two summary scores known as the physical component summary (PCS) score and the mental component summary (MCS) score. Scoring of each domain and both summary scores are norm based and presented in the form of T-scores, with a mean of 50 and a standard deviation of 10. Higher scores indicate better levels of function and/or better health. Range cannot be specified in norm-based scores.

LS Mean was determined by MMRM model with Baseline + Country + Carbohydrate counting for Prandial Dose + Prior CGM use + Hemoglobin A1c Stratum at Baseline + Treatment + Time + Treatment*Time (Type III sum of squares). as variables. Unstructured variance-covariance structure was used.