This completed single-center (Drug Clinical Trial Center, Peking University Third Hospital, Beijing, China), randomized, double-blind, three-period crossover study (ClinicalTrials.gov identifier NCT05435677; registration date 23 June 2022) was conducted according to the principles of the Declaration of Helsinki [19] and ICH Good Clinical Practice [20]. The protocol for this study was approved by a suitably constituted independent ethics committee (Ethics Committee of Peking University Third Hospital, Beijing, China). All participants provided informed consent.

Eligibility for participation in this study required individuals to be Chinese men or women between the ages of 18 and 64 years (both inclusive) and having been diagnosed with T2D ≥ 180 days prior to screening. Individuals should be insulin naïve and should receive stable dose(s) of any of the following oral non-insulin glucose-lowering medications: metformin, dipeptidyl peptidase 4 (DPP4) inhibitors, sodium-glucose cotransporter 2 inhibitors, and/or alpha-glucosidase inhibitors. Short-term insulin treatment for a maximum of 14 days prior to screening and previous insulin treatment for gestational diabetes were, however, allowed. Individuals should also have a body weight of ≥ 50 kg, HbA1c ≤ 9.0% (75 mmol/mol), and a body mass index between 18.5 and 34.9 kg/m2 (both inclusive). Screened individuals were excluded from participation if they had any disorder that could jeopardize their safety or compliance with the study protocol. Conditions associated with T2D were exempted if they were under stable treatment. Individuals were also excluded if they had supine blood pressure outside of 90–159 mmHg (systolic) or 50–99 mmHg (diastolic), calcitonin ≥ 50 ng/L, presence or history of acute or chronic pancreatitis, renal impairment defined as an estimated glomerular filtration rate < 60 mL/min/1.73 m2, hepatic impairment defined as an alanine aminotransferase level ≥ 2.5 times the upper normal limit, a bilirubin level > 1.5 times the upper normal limit, uncontrolled and potentially unstable diabetic retinopathy or maculopathy, recurrent severe hypoglycemia (> 1 severe hypoglycemic episode within the last 180 days), hypoglycemia unawareness, or hospitalization for diabetic ketoacidosis within the last 180 days. Women that were pregnant, breast-feeding, or had the intention to become pregnant were also excluded from participation.

The overall study design, including procedures and assessments, was largely comparable with a previous study conducted in white individuals with T2D [16].

The current study comprised a screening visit (up to 3 weeks before randomization), a DPP4 inhibitor washout period, three treatment periods, and a follow-up visit (Fig. S1). The DPP4 inhibitor washout period of 5–6 days prior to randomization was included for participants treated with any DPP4 inhibitor at entry into the study in order to discontinue the DPP4 inhibitor treatment. Participants were allowed to continue treatment with other oral glucose-lowering medications throughout the study at a stable dose. Each treatment period consisted of a single dose of study product followed by 5 weeks of pharmacokinetic blood sampling. The treatment periods were separated by 1–4 weeks from the end of pharmacokinetic blood sampling until the next single dose.

The single dose of study product received at the beginning of each treatment period consisted of IcoSema, icodec alone, or semaglutide alone in randomized sequence. The IcoSema dose contained 40 U icodec (equivalent to 240 nmol) and 0.114 mg semaglutide (equivalent to 27.7 nmol). The icodec dose administered alone contained 40 U, while the semaglutide dose administered alone contained 0.1072 mg (equivalent to 26.1 nmol). IcoSema and semaglutide were provided in 1.5-mL pre-filled PDS290 pen injectors, while icodec was provided in a 3-mL pre-filled PDS290 pen injector (Novo Nordisk, Bagsværd, Denmark). The IcoSema study product strength was 700 U/mL icodec and 2 mg/mL semaglutide (Novo Nordisk). The icodec study product strength was 700 U/mL (Novo Nordisk), and the semaglutide study product strength was 1.34 mg/mL (Novo Nordisk). The slight difference in semaglutide dose between administration in IcoSema (0.114 mg) and administration alone (0.1072 mg) was due to the difference in semaglutide concentration between IcoSema (2 mg/mL) and semaglutide alone (1.34 mg/mL) combined with the injection volume equivalent to each increment on the PDS290 pen injector. Semaglutide 1.34 mg/mL provided in a 1.5-mL pre-filled PDS290 pen injector was the only semaglutide product available in China at the time of conduct of the present study. All three study products were scheduled to be injected subcutaneously in the left thigh by trained site staff at approximately 08:00 hours. Participants remained in-house at the clinical study site until 84–96 h after each administration of study product.

Pharmacokinetic assessment was carried out as previously described [16]. In brief, blood samples for analysis of icodec and semaglutide concentrations were drawn as shown in Table S1. Serum icodec concentrations were measured following IcoSema and icodec administration using a validated luminescence oxygen channelling immunoassay (LOCI). Plasma semaglutide concentrations were measured following IcoSema and semaglutide administration using a validated liquid chromatography–tandem mass spectrometry assay.

Safety assessments included adverse events (AEs), hypoglycemic episodes, clinical laboratory assessments, vital signs, physical examination, and electrocardiogram.

Statistical analyses were carried out using SAS version 9.4 (SAS Institute, Cary, NC, USA).

The primary endpoints in the current study were the area under the serum icodec concentration–time curve and the dose-normalized area under the plasma semaglutide concentration–time curve from zero to last quantifiable observation after a single dose (AUC0–t,SD,Icodec and dose-normalized AUC0–t,SD,Semaglutide). In the previous study with similar overall design in white participants with T2D, the standard deviation of the within-participant differences was 0.162 for log(AUC0–t,SD,Icodec) and 0.164 for log(dose-normalized AUC0–t,SD,Semaglutide) [16]. Assuming a standard deviation for both endpoints of 0.18 and that 16 participants completed the study, then the 90% confidence interval (CI) for the estimated geometric mean ratio between IcoSema and one of the two components administered alone would range from 0.91 to 1.10 times the geometric mean ratio with 90% probability. This was considered sufficient for comparison of relative bioavailability between the study products. To ensure 16 completers, 20 individuals were randomized.

Calculation and statistical analysis of pharmacokinetic endpoints were conducted essentially as described previously [16]. AUC0–t,SD,Semaglutide and the maximum concentration after a single dose for semaglutide (Cmax,SD,Semaglutide) were, however, dose-normalized in the current study as a result of the slightly different dose of semaglutide administered in IcoSema versus semaglutide alone. Evaluation of AUC0–t,SD and Cmax,SD for both icodec and semaglutide was carried out on the basis of guidelines for bioavailability and bioequivalence studies [21, 22].

Safety parameters were presented by descriptive statistics. AEs and hypoglycemic episodes were defined as treatment emergent if occurring from each single dose administration until 35 days later. Hypoglycemic episodes were classified as level 1 (hypoglycemia alert value; plasma glucose < 3.9 mmol/L and ≥ 3.0 mmol/L), level 2 (clinically significant hypoglycemia; plasma glucose < 3.0 mmol/L), and level 3 (severe hypoglycemia; associated with severe cognitive impairment requiring external assistance for recovery) [23].

Pharmacokinetic modelling was used to simulate the pharmacokinetic profiles at steady state based on the observed pharmacokinetic results of the present single-dose study. The steady-state pharmacokinetic profile of icodec was simulated after IcoSema and icodec dosing, while the steady-state pharmacokinetic profile of semaglutide was simulated after IcoSema and semaglutide dosing. The pharmacokinetic models for icodec and semaglutide as well as the modelling approach have been described previously [16]. In brief, 10 weeks of once-weekly dosing of IcoSema (40 U icodec/0.114 mg semaglutide), icodec (40 U), or semaglutide (0.1072 mg) was simulated for each participant. Maximum icodec concentration at steady state (Cmax,SS,Icodec) was compared between IcoSema and icodec alone, while the dose-normalized Cmax,SS,Semaglutide was compared between IcoSema and semaglutide alone, using the same statistical analysis as described above for AUC0–t,SD and Cmax,SD.