Efficacy and safety of once-weekly semaglutide versus once-daily sitagliptin as add-on to metformin in patients with type 2 diabetes in SUSTAIN China: A 30-week, double-blind, phase 3a, randomized trial

Abstract

Aim: To evaluate the efficacy and safety of once-weekly subcutaneous semaglutide, a glucagon-like peptide-1 (GLP-1) analogue, versus once-daily sitagliptin as add-on to metformin in patients with type 2 diabetes (T2D) in a multiregional clinical trial.

Materials and methods: In the 30-week, randomized, double-blind, double-dummy, active comparator SUSTAIN China trial, 868 adults with T2D inadequately controlled on metformin (HbA1c 7.0%-10.5%) were randomized to receive once-weekly semaglutide 0.5 mg (n = 288), semaglutide 1.0 mg (n = 290) or once-daily sitagliptin 100 mg (n = 290). The primary and confirmatory secondary endpoints were change from baseline to week 30 in HbA1c and body weight, respectively.

Results: The trial enrolled ~70% (605/868) of the patients in China, and the remaining patients from four other countries, including the Republic of Korea. Both doses of semaglutide were superior to sitagliptin in reducing HbA1c and body weight after 30 weeks of treatment. The odds of achieving target HbA1c of less than 7.0% (53 mmol/mol), weight loss of 5% or higher, or 10% or higher, and the composite endpoint of HbA1c less than 7.0% (53 mmol/mol) without severe or blood glucose-confirmed symptomatic hypoglycaemia no weight gain, were all significantly higher with both semaglutide doses compared with sitagliptin. The safety profile for semaglutide was consistent with the known class effects of GLP-1 receptor agonists (RAs). Consistent efficacy and safety findings were seen in the Chinese subpopulation.

Conclusions: Once-weekly semaglutide was superior to sitagliptin in improving glycaemic control and reducing body weight in patients with T2D inadequately controlled on metformin. The safety and tolerability profiles were consistent with those of semaglutide and other GLP-1 RAs. Semaglutide is an effective once-weekly treatment option for the Chinese population.

Keywords: GLP-1 analogue, glycaemic control, incretin therapy, phase III study, randomized trial, type 2 diabetes.

Conflict of interest statement

LJ reports receiving personal fees from Novo Nordisk during the conduct of this study. ZN is an employee of Novo Nordisk. SR is an employee of and stockholder at Novo Nordisk. TVS is an employee of and shareholder at Novo Nordisk. XD, YiL, YuL, SL, ML and GY have nothing to disclose. FGE reports receiving grants from Novo Nordisk outside the submitted work.

© 2020 The Authors. Diabetes, Obesity and Metabolism published by John Wiley & Sons Ltd.

Figures

FIGURE 1

Flow of participation through the trial. Note: Data for randomized patients (n = 868) are from the full analysis set. Number of patients randomized in the China region, which includes China, Hong Kong and Taiwan, (N = 605), the Republic of Korea (N = 110), Brasil (N = 75), South Africa (N = 45) and Ukraine (N = 33)

FIGURE 2

Primary and secondary efficacy endpoints (HbA1c and body weight) change from baseline to week 30. Change in A, mean HbA1c by week, B, change in mean HbA1c after 30 weeks, C, change in mean body weight by week, D, change in mean body weight after 30 weeks, E, proportion of patients achieving the HbA1c target of less than 7.0%, F, proportion of patients achieving HbA1c less than 7.0% without severe or BG‐confirmed symptomatic hypoglycaemia and no weight gain and G, the proportion of patients achieving a weight loss of 5% or more. For all estimated change data (panels A–D), data are mean estimate (± standard error) ‘On‐treatment without rescue medication’ data. The post‐baseline responses were analyzed using a mixed model for repeated measurements with treatment and the China region/other as fixed factors and baseline value as covariate, all nested within visit. Mean estimates are adjusted according to observed baseline distribution. All site visits, except screening visit, were to be completed in fasting state. In panel B, the non‐inferiority p‐value was calculated as two times one‐sided p‐value from a t‐distributed test statistic comparing the treatment contrast with 0.3 rather than zero as in a superiority test. For panels E–G, data are ‘On‐treatment without rescue medication’ data. The binary endpoints were analyzed using a logistic regression model with treatment and the China region/other as fixed factors and baseline values as covariates. Before analysis, missing data for individual components were imputed from a mixed model for repeated measurements with treatment and the China region/other as fixed factors and baseline values as covariates, all nested within visit, and subsequently dichotomized. All site visits, except screening visit, were to be completed in fasting state. BG, blood glucose; ETD, estimated treatment difference; CI, confidence interval; OR, odds ratio; *statistically significant