Objective: To evaluate the efficacy and safety of domestic semaglutide injection in the treatment of type 2 diabetes mellitus (T2DM) inadequately controlled with metformin. Methods: This multicenter, randomized, positive-controlled phase 3 equivalence clinical trial enrolled patients with T2DM from 39 research centers in China between March 13, 2024, and April 9, 2025. Central randomization was performed using interactive response technology, and the subjects were stratified according to their glycated hemoglobin A1c (HbA1c) levels (<8.5% and ≥8.5%) and randomly divided into an experimental group (domestic semaglutide injection) and a control group (original semaglutide injection) at a ratio of 1∶1. The primary endpoint was the change from baseline in HbA1c at week 32, with an equivalent threshold of ±0.4%. The key secondary endpoints included HbA1c change from baseline at week 20, the proportion of patients with HbA1c<7.0% or HbA1c≤6.5% at 20 and 32 weeks, and the evaluation of safety. Efficacy endpoints were analyzed in the full analysis set (FAS), and safety endpoints were analyzed in the safety set (SS). Changes from baseline in HbA1c were analyzed using a mixed model for repeated measures; the proportions of subjects with HbA1c<7.0% or HbA1c≤6.5% were compared using the Cochran-Mantel-Haenszel χ2 test; and safety endpoints were analyzed using Fisher's exact test. Results: Both the FAS and SS included 481 patients, with 241 in the experimental group and 240 in the control group. After 32 weeks of treatment, the least squares mean (LSM) (95%CI) changes from baseline in HbA1c were -1.80% (-1.93%--1.68%) in the experimental group and -1.85% (-1.98%--1.72%) in the control group, with a difference of 0.05% (-0.13%-0.23%)(both one-sided P<0.001). Similarly, equivalence was demonstrated at week 20, with an intergroup difference (95%CI) of 0.08% (-0.10%-0.26%) for the change from baseline in HbA1c (both one-sided P<0.001). The proportions of patients achieving HbA1c<7.0% or HbA1c≤6.5% showed no statistically significant differences between the two groups. The incidence of overall adverse events was comparable between the two groups [85.06% (205/241) vs 83.33% (200/240)], and the incidences of drug-related hypoglycemic events were 0.41% (1/241) and 2.50% (6/240), respectively, with no statistically significant differences (both P>0.05). Conclusions: In patients with T2DM inadequately controlled by metformin, domestic semaglutide injection demonstrated glucose-lowering efficacy equivalent to original semaglutide injection over the 32-week short-term period, with generally similar overall safety profiles between the two groups.
