Biosimilar Insulin Aspart Premix SAR341402 Mix 70/30 Versus Originator Insulin Aspart Mix 70/30 (NovoMix 30) in People with Diabetes: A 26-Week, Randomized, Open-Label Trial (GEMELLI M)

Abstract

Introduction: This study compared the efficacy, safety, and immunogenicity of biosimilar insulin aspart premix SAR341402 Mix 70/30 (SAR_Asp-Mix) with European-approved insulin aspart mix 70/30 - NovoMix® 30 (NN-Mix) in people with type 1 (T1D) or type 2 diabetes (T2D).

Methods: This 26-week, open-label, phase 3 trial enrolled 402 people with T1D (n = 105) or T2D (n = 297) previously treated with premix insulin, who were randomized (1:1) to SAR_Asp-Mix (n = 204) or NN-Mix (n = 198).

Results: After 26 weeks, the least squares (LS) mean [median] change in HbA1c from baseline was similar in both treatment groups (SAR_Asp-Mix - 0.55% [- 0.60%]; NN-Mix - 0.64% [- 0.60%]). The LS mean difference for SAR_Asp-Mix versus NN-Mix was 0.08%, with the upper bound of the two-sided 95% confidence interval (- 0.139 to 0.303) slightly above the prespecified noninferiority margin of 0.3%. Noninferiority of SAR_Asp-Mix over NN-Mix was not demonstrated in the primary intent-to-treat analysis, primarily because of one extreme outlying value impacted by the COVID-19 pandemic in the SAR_Asp-Mix group. Noninferiority was achieved in all secondary analyses, including prespecified per-protocol supportive and COVID-19 sensitivity analyses, as well as post hoc sensitivity analyses. Other efficacy endpoints, insulin dosages, anti-insulin aspart antibody response, hypoglycemia, and adverse events were similar between groups.

Conclusions: The totality of evidence indicates that SAR_Asp-Mix provides effective glycemic control with a similar safety and immunogenicity profile to NN-Mix in people with diabetes treated for 26 weeks.

Trial registration: EudraCT number 2017-000092-84.

Keywords: Biosimilar; GEMELLI M; Insulin aspart mix; Premix; SAR341402.

Fig. 1

HbA1c (% and mmol/mol) by study visit (a), least squares mean change in HbA1c (%) from baseline to week 26 in overall study population and by subgroup of diabetes type (T1D or T2D) using ANCOVA analysis (with return to baseline multiple imputation). The statistical model used for the analysis is described in Table 2 for the overall study population and in ESM Fig. S3 by subgroup of diabetes type. P value for treatment-by-subgroup interaction = 0.3199 (b), FPG (mmol/L and mg/dL) by study visit (c), and seven-point SMPG profiles (mmol/L and mg/dL) at baseline and week 26 (d) in the ITT population. Data are mean ± standard error. ANCOVA analysis of covariance, BL baseline, FPG fasting plasma glucose, HbA1c glycosylated hemoglobin, ITT intent-to-treat, SMPG self-monitored plasma glucose, T1D type 1 diabetes, T2D type 2 diabetes, W week

Fig. 2

Daily insulin doses (U/kg) by study visit during the on-treatment period in participants with T1D (a) and T2D (b) (safety population). Data are mean ± SE. Baseline insulin doses were defined as the median of daily doses available in the week prior to the first injection of study medication (corresponding to doses of the prestudy insulin). Insulin doses at day 1 were defined as the median of daily doses available in the week after the first injection of study medication. For weeks 4, 12, 20, and 26, insulin dose values were reported as the median of daily doses available in the week prior to the study visit. BL baseline, D day, SE standard error, T1D type 1 diabetes, T2D type 2 diabetes, W week

Fig. 3

Forest plot of the odds ratio of participants experiencing at least one episode of hypoglycemia (a) and rate ratio (events per patient-year) (b) by category of hypoglycemia during the 26-week on-treatment period for the overall population (safety population). Odds ratio results (a) based on a logistic regression model with fixed-effect terms for treatment group and randomization strata of geographical region (Indian, non-Indian), type of diabetes (T1D, T2D), screening HbA1c (less than 8%, 8% or higher), and prior use of NN-Mix (Yes, No). Rate ratio results (b) based on an overdispersed Poisson regression model with fixed-effect terms as described above for odds ratio results. If the model did not converge (e.g., because of sparse data), randomization strata were removed from the model. ^aOdds ratio SAR_Asp-Mix versus NN-Mix for participants with at least one hypoglycemic event. ^bRate ratio SAR_Asp-Mix versus NN-Mix for hypoglycemic events per patient-year. HbA1c glycated hemoglobin, n number of patients with at least one treatment-emergent event, % percentage of participants with at least one event, nE number of events, PY total patient-years, R rate per patient-year, NC model did not converge