Simplified Switching to Once‐Weekly Insulin Icodec Without an Initial One‐Time Additional Dose Versus Once‐Daily Insulin Glargine U100 in Basal Insulin‐Treated Type 2 Diabetes (ONWARDS 10): A Randomised Controlled Trial

ABSTRACT

Aims

ONWARDS 10 assessed switching from daily basal insulins to once‐weekly insulin icodec without a one‐time additional dose at first injection.

Materials and Methods

In this treat‐to‐target, open‐label, 26‐week, phase 3b trial, adults with basal‐insulin‐treated type 2 diabetes (T2D) with glycated haemoglobin [HbA_1c] 7.0%–10.0% were randomised to once‐weekly icodec without an initial one‐time additional dose or to once‐daily insulin glargine U100. Primary endpoint was change in HbA_1c from baseline to week 26 (non‐inferiority margin: 0.3%‐points); secondary endpoints included change in time in range (TIR; 70–180 mg/dL [3.9–10.0 mmol/L]) from baseline (weeks −4 to 0) to weeks 22–26, and number of combined clinically significant or severe hypoglycaemic episodes (weeks 0–31).

Results

Overall, 412 participants were randomised to icodec (n = 206) or glargine U100 (n = 206). At week 26, estimated mean HbA_1c was 7.2% with icodec (baseline 8.1%) and 7.5% with glargine U100 (baseline 8.0%). Estimated treatment difference (ETD) for change in HbA_1c was −0.2%‐points (95% confidence interval [CI]: −0.4, −0.1), confirming non‐inferiority of icodec versus glargine U100 (p < 0.0001). Self‐measured blood glucose did not increase following icodec initiation. Change in TIR from baseline to weeks 22–26 was statistically significantly higher with icodec than with glargine U100 (estimated mean change: 18.0%‐points vs. 12.1%‐points; ETD: 5.9%‐points [95% CI: 2.6, 9.2], p = 0.0005). Combined clinically significant or severe hypoglycaemia rate was numerically lower with icodec than with glargine U100 (0.45 vs. 0.59 episodes per patient‐year of exposure).

Conclusions

Omitting the one‐time additional dose simplifies switching to icodec in adults with basal‐insulin‐treated T2D without compromising initial glycaemic control.

Trial Registration: ClinicalTrials.gov identifier: NCT06340854

1. Introduction

Insulin icodec (icodec) is a basal insulin analogue suitable for once‐weekly administration [1]. The efficacy and safety of icodec were investigated in the ONWARDS phase 3a clinical development programme in insulin‐naïve (ONWARDS 1, 3, and 5) and insulin‐experienced (ONWARDS 2 and 4) individuals with type 2 diabetes (T2D) [2, 3, 4, 5, 6, 7]. Across ONWARDS trials, icodec demonstrated non‐inferiority (ONWARDS 1–5) and superiority (ONWARDS 1–3 and 5) in glycated haemoglobin (HbA_1c) reduction versus once‐daily basal insulins [3, 4, 5, 6, 7].

In the ONWARDS 2 (basal switch) and 4 (basal‐bolus switch) trials involving switching from daily basal insulin to icodec, the weekly icodec dose was calculated by multiplying the pretrial daily basal insulin dose by seven [4, 6], In these trials, the first injection of icodec included an initial one‐time 50% additional dose to allow for faster achievement of steady state, given the long half‐life of icodec (~8 days), and to prevent potential transient hyperglycaemia immediately after the insulin switch [1, 8]. For the second injection (week 2), the icodec dose reverted to the initial calculated weekly dose (daily dose multiplied by 7). To date, the efficacy and safety of switching to icodec without using this additional dose approach in people with insulin‐treated T2D have not been explored. Despite the proven efficacy and tolerability of switching to icodec with a one‐time 50% additional dose, it is therefore of interest to investigate the effect of switching to icodec without the one‐time additional dose to further simplify its use. Omitting the one‐time additional dose can further simplify the protocol for switching to icodec, offering an alternative option allowing for individualised dosing, consistent with the recommended approach for all basal insulin treatments.

Modelling analysis based on ONWARDS 2 and 4 data predicted that switching to icodec without versus with a one‐time 50% additional dose would result in a mild, transient increase in prebreakfast self‐measured blood glucose (SMBG) levels in the first 1–2 weeks after treatment initiation; it was hypothesised that this increase would reduce to reach matching glycaemic control by week 4 [9]. We therefore wanted to test the modelling premise further.

The aim of ONWARDS 10 was to investigate for the first time the efficacy and safety of once‐weekly icodec when switched unit‐to‐unit from a daily basal insulin without an initial one‐time additional dose, versus once‐daily insulin glargine U100 (glargine U100) in adults with basal‐insulin‐treated T2D, with the aim of further simplifying the switch to weekly basal insulin icodec.

2. Research Design and Methods

2.1. Trial Design and Participants

ONWARDS 10 (NCT06340854) was a 26‐week, randomised, multicentre, treat‐to‐target, open‐label, phase 3b trial conducted at 70 sites across eight countries (Bulgaria, Germany, India, Japan, Poland, Spain, South Africa, and the USA). The trial consisted of a 2‐week screening period, a 4‐week run‐in period (during which participants remained on their pretrial basal insulin), a 26‐week treatment period, and a 5‐week safety follow‐up period (Figure S1). Eligible participants (aged ≥ 18 years; body mass index [BMI] ≤ 40.0 kg/m²) had T2D (HbA_1c 7.0%–10.0% [53.0–85.8 mmol/mol]) treated with once‐ or twice‐daily basal insulin with or without non‐insulin glucose‐lowering medications. See Table S1 for full inclusion and exclusion criteria.

The trial was conducted in compliance with the principles of the Declaration of Helsinki and in accordance with the Good Clinical Practice guidelines of the International Conference for Harmonisation. The protocol, consent form, and other relevant documents were reviewed and approved by the appropriate institutional review boards or independent ethics committees. All participants provided written informed consent before starting the trial.

2.2. Treatments

After the run‐in period, participants were centrally randomised (1:1) using an interactive web response system, ensuring concealed and unbiased assignment and reducing investigator influence. At randomisation, participants were switched from their pretrial basal insulin to once‐weekly icodec (700 U/mL; PDS290 prefilled pen injector) or once‐daily glargine U100 (100 U/mL; SoloSTAR prefilled pen injector), both administered subcutaneously. All personnel (including sponsor personnel) involved in data processing and interpretation were blinded to treatment allocations.

Switching from pretrial daily basal insulin to icodec was performed on a unit‐to‐unit basis (equivalent to pretrial total daily basal insulin dose multiplied by 7). In contrast to ONWARDS 2 (basal switch trial), an initial one‐time 50% additional icodec dose was not included at the first injection. In the comparator group, participants either switched to glargine U100 in accordance with the local label or continued their pretrial glargine U100 dose. Icodec and glargine U100 doses were adjusted weekly based on the three pre‐breakfast SMBG values measured 2 days before titration and on the day of titration, according to a prespecified algorithm (Table S2) to achieve a prebreakfast SMBG target of 80–130 mg/dL (4.4–7.2 mmol/L). Icodec dose was increased by 20 U if SMBG levels were above 130 mg/dL (7.2 mmol/L) and reduced by 20 U if they were below 80 mg/dL (4.4 mmol/L). Glargine U100 doses were adjusted by 3 U increments or decrements. Pretrial non‐insulin glucose‐lowering medications were continued throughout the trial, apart from sulfonylureas and glinides, which were discontinued at randomisation.

At randomisation, participants were provided with a blood glucose meter and instructed to measure their prebreakfast blood glucose levels daily throughout the 26‐week intervention period and the 5‐week safety follow‐up period. All values measured by the blood glucose meter indicative of hypoglycaemia were recorded as hypoglycaemic episodes in the participants' eDiary. Participants wore a blinded continuous glucose monitoring (CGM) device (Dexcom G6, Dexcom Inc., San Diego, CA, USA) at three prespecified periods: during the run‐in period (weeks −4 to 0), after treatment initiation (weeks 0–4), and during the last 4 weeks of treatment (weeks 22–26). CGM data collected during the run‐in period were used as baseline values. CGM readings were blinded to participants and investigators and were not used for insulin dose titration or reporting hypoglycaemic episodes. Participants using a personal real‐time CGM device were allowed to continue using this during the study CGM periods. If they chose to do so, the use was in addition to the study‐provisioned CGM and personal devices could not be used to generate study data.

2.3. Outcomes

The primary endpoint was absolute change in HbA_1c from baseline (week 0) to week 26. Supportive secondary efficacy endpoints were change in mean percentage of time in range (TIR; 70–180 mg/dL [3.9–10.0 mmol/L]) from baseline (weeks −4 to 0) to weeks 22–26 and change in Diabetes Treatment Satisfaction Questionnaire status version (DTSQs) total treatment satisfaction score (baseline to week 26).

Supportive secondary safety endpoints were mean percentage of time below range (TBR; < 54 mg/dL [< 3.0 mmol/L]) during weeks 22–26; change in mean percentage of time above range (TAR; > 180 mg/dL [> 10.0 mmol/L]) from baseline (weeks −4 to 0) to weeks 22–26; mean weekly insulin dose during weeks 24–26; change in body weight from baseline (week 0) to week 26; and number of clinically significant hypoglycaemic episodes (level 2; < 54 mg/dL [< 3.0 mmol/L]), severe hypoglycaemic episodes (level 3; associated with severe cognitive impairment requiring external assistance for recovery), and combined clinically significant or severe hypoglycaemic (level 2 or 3) episodes from baseline (week 0) to week 31.

Change in Treatment‐Related Impact Measure for Diabetes (TRIM‐D) total score (baseline to week 26) was an exploratory endpoint. Additional assessments were achievement of HbA_1c < 7.0% (< 53 mmol/mol) at week 26; achievement of this HbA_1c target at week 26 without clinically significant or severe hypoglycaemia in the previous 12 weeks; and achievement of clinically relevant improvements in DTSQs total treatment satisfaction score at week 26. Adverse events (AEs) were evaluated throughout the trial.

2.4. Statistical Analysis

The primary hypothesis was that icodec was non‐inferior to glargine U100 in terms of change in HbA_1c from baseline to week 26, using a non‐inferiority margin of 0.3%‐points, in line with recommendations from US health authorities [10, 11]. This margin is considered to provide sufficient sensitivity based on the following considerations: that it does not represent an unacceptable loss of efficacy with icodec relative to treatment with a basal insulin, and that it represents less than 50% of a suitably conservative estimate of the expected treatment effect of glargine U100 on HbA_1c in a placebo‐controlled study. The primary estimand was defined as the between‐group difference in change in HbA_1c from baseline to week 26 among all randomised participants, irrespective of initiation of bolus insulin treatment for more than 2 weeks or discontinuation of the investigational intervention. Sample size calculations indicated that, based on a previous icodec trial [4], a sample size of 404 participants was required to have 90% power for confirming non‐inferiority of the primary hypothesis using a non‐inferiority margin of 0.3%.

All efficacy, patient‐reported outcome, and CGM endpoints were evaluated using the in‐trial period (all data obtained at or after randomisation until last participant contact) and the full analysis set (all randomised participants). Safety endpoints were evaluated using the on‐treatment period (all data obtained while participants were considered exposed to treatment), with descriptive statistics based on the safety analysis set (all randomised participants exposed to treatment) and statistical analyses based on the full analysis set.

The primary endpoint was analysed using an analysis of covariance (ANCOVA) model with randomised treatment and region as fixed factors, and baseline HbA_1c as a covariate. Missing HbA_1c values at week 26 were imputed by the baseline value adding a random error term, using multiple imputation. Change in body weight, TIR, TAR, and DTSQs total treatment satisfaction score were analysed using an ANCOVA model with the same fixed factors as mentioned previously, baseline value as covariate, and with return to baseline imputation. Mean weekly insulin dose from weeks 24–26 was estimated using a similar model, with log‐transformed mean weekly insulin dose from week 24 to week 26 and using the log‐transformed baseline weekly insulin dose as a covariate. Missing data were imputed by return to baseline multiple imputation.

The number of hypoglycaemia episodes was analysed using a negative binomial model with a log‐link function, fixed factors as mentioned previously, and the logarithm of the period for which the hypoglycaemic episodes were considered as offset. Missing data were imputed using multiple imputation. TBR was analysed using a negative binomial model on the number of recorded measurements below range with a log‐link function and the logarithm of the total number of recorded measurements as an offset. The model included the same fixed factors as mentioned previously, and the baseline number of recorded measurements below range as a covariate. No imputation of missing data was performed for TBR.

The threshold for achievement of clinically relevant improvements in DTSQs total treatment satisfaction score was calculated as the mean change in DTSQs total treatment satisfaction score from baseline to week 26 for participants with a one‐category improvement in anchor measure (Patient Global Impression of Status [PGI‐S] Diabetes Treatment Satisfaction). Achievement of clinically relevant improvements in DTSQs total treatment satisfaction score (yes/no) was evaluated using a logistic regression model, with the same fixed factors as mentioned previously, and baseline value as a covariate. Achievement of HbA_1c < 7.0% at week 26 and achievement of this HbA_1c target at week 26 without clinically significant or severe hypoglycaemia in the previous 12 weeks was analysed using a similar model as used for DTSQs but including baseline HbA_1c as a covariate instead of baseline DTSQs score.

All presented statistical analyses were prespecified; however, the statistical test for treatment difference in change in HbA_1c was not. For Figure 1, weekly SMBG values were averaged per participant, then across each of the treatment groups and plotted over time as two curves; however, week 0 only presents data collected at the baseline visit. All statistical analyses were performed using SAS (version 9.4).

FIGURE 1.

Key outcomes. (A) Observed mean HbA_1c from baseline over time. Error bars show standard error of the mean. Observed data include data obtained after premature discontinuation. ^aData shown at week 26 are the estimated mean values and the corresponding standard error at week 26 based on multiple imputations. (B) Full‐week mean fasting SMBG values assessed as observed means at weeks 1–4. Week 0 presents data from the baseline visit only. (C) CGM‐based glycaemic ranges assessed as observed means at weeks −4–0 and weeks 22–26. Time spent is defined as 100 times the number of recorded measurements in a given range divided by the total number of recorded measurements. (D) Observed mean weekly total insulin dose over time. The graph shows the geometric mean (symbol) and standard error of the mean (error bars). CGM, continuous glucose monitoring; CI, confidence interval; ETD, estimated treatment difference; glargine U100, insulin glargine U100; HbA_1c, glycated haemoglobin; icodec, insulin icodec; SMBG, self‐measured blood glucose; TAR, time above range; TBR, time below range; TIR, time in range; U, units.

3. Results

3.1. Participants

Participants were recruited between April 19, 2024 and June 13, 2025. Of 592 screened participants, 412 were randomised to icodec (n = 206) or glargine U100 (n = 206). All randomised participants are included in the full analysis set, and since all randomised participants received at least one dose of trial treatment, they were all also included in the safety analysis set. Overall, 99.0% (n = 204) of participants in the icodec group and 98.1% (n = 202) in the glargine U100 group completed the week 26 visit without permanently discontinuing trial treatment (Figure S2). No participants were using personal CGM devices at the time of the trial.

Participant demographics and clinical characteristics are shown in Table 1. Baseline characteristics were generally comparable between treatment groups, except for lower TIR and higher TAR in the icodec group than in the glargine U100 group. The most commonly used basal insulin at screening was glargine U100 (39%), and the most commonly used concomitant glucose‐lowering medications were metformin (83%) followed by sodium‐glucose co‐transporter‐2 (SGLT2) inhibitors (50%), sulfonylureas (31%), and glucagon‐like peptide‐1 receptor agonists (GLP‐1 RAs) (23%).

TABLE 1.

Baseline characteristics.

	Icodec (n = 206)	Glargine U100 (n = 206)
Sex, n (%)
Male	122 (59.2)	116 (56.3)
Female	84 (40.8)	90 (43.7)
Age (years)	62.4 ± 9.9	62.5 ± 9.6
Race, n (%)
American Indian or Alaska Native	1 (0.5)	0 (0.0)
Asian	69 (33.5)	74 (35.9)
Black or African American	8 (3.9)	8 (3.9)
White	127 (61.7)	122 (59.2)
American Indian or Alaska Native, White	0 (0.0)	1 (0.5)
Not reported	1 (0.5)	1 (0.5)
Diabetes duration (years)	16.0 ± 7.8	16.2 ± 7.5
Body weight (kg)	83.6 ± 18.6	82.9 ± 17.6
BMI (kg/m2)	29.7 ± 5.0	29.8 ± 5.0
HbA1c (%)	8.08 ± 0.90	8.01 ± 0.85
HbA1c (mmol/mol)	64.8 ± 9.8	64.0 ± 9.3
TIR (70–180 mg/dL [3.9–10.0 mmol/L]) (%)	44.2	49.0
TAR (> 180 mg/dL [> 10.0 mmol/L]) (%)	55.3	50.3
TBR (< 54 mg/dL [< 3.0 mmol/L]) (%)	0.2	0.2
Total daily basal insulin dose at screening (U) a	27.2 (65.3)	26.8 (65.2)
Once‐daily basal insulin at screening, n (%)
Insulin glargine U100	67 (32.5)	89 (43.2)
Insulin degludec	42 (20.4)	44 (21.4)
Insulin glargine	27 (13.1)	24 (11.7)
Insulin human isophane	27 (13.1)	24 (11.7)
Insulin glargine U300	13 (6.3)	15 (7.3)
Insulin detemir	9 (4.4)	1 (0.5)
Other a	1 (0.5)	2 (1.0)
Twice‐daily basal insulin at screening, n (%)
Insulin human isophane	9 (4.4)	4 (1.9)
Insulin glargine U100	4 (1.9)	1 (0.5)
Insulin detemir	2 (1.0)	0
Insulin isophane porcine	2 (1.0)	0
Other b	3 (1.5)	1 (0.5)
Concomitant glucose‐lowering medication at screening, n (%)
Metformin	169 (82.0)	172 (83.5)
SGLT‐2 inhibitors	102 (49.5)	103 (50.0)
Sulfonylureas	60 (29.1)	68 (33.0)
GLP‐1 RA	46 (22.3)	48 (23.3)
DPP‐4 inhibitors	48 (23.3)	46 (22.3)
Thiazolidinediones	14 (6.8)	12 (5.8)
Glinides	6 (2.9)	10 (4.9)
Alpha‐glucosidase inhibitors	6 (2.9)	3 (1.5)

Note: Data are mean ± SD unless specified otherwise. Baseline refers to week 0, except for the CGM parameters, which were recorded during weeks −4 to 0.

Abbreviations: BMI, body mass index; CGM, continuous glucose monitoring; DPP‐4, dipeptidyl peptidase‐4; glargine U100, insulin glargine U100; GLP‐1 RA, glucagon‐like peptide‐1 receptor agonist; HbA_1c, glycated haemoglobin; icodec, insulin icodec; SD, standard deviation; SGLT‐2, sodium‐glucose co‐transporter 2; TAR, time above range; TBR, time below range; TIR, time in range.

^a Data are geometric mean (coefficient of variation).

^b Other once‐daily basal insulins at screening include insulin glargine, lixisenatide, insulin isophane bovine, and insulin isophane porcine. Other twice‐daily basal insulins at screening include insulin degludec, insulin glargine, and insulin glargine U300.

3.2. Glycaemic Outcomes

The mean HbA_1c in the icodec group decreased from 8.1% at baseline to an estimated mean of 7.2% at week 26. In the glargine U100 group, mean HbA_1c decreased from 8.0% at baseline to 7.5% at week 26. Estimated mean change in HbA_1c from baseline to week 26 was −0.8%‐points with icodec and −0.6%‐points with glargine U100, with an estimated treatment difference (ETD) of −0.2%‐points (95% confidence interval [CI]: −0.4, −0.1), confirming non‐inferiority of icodec versus glargine U100 (p < 0.0001) (Figure 1A). The reduction in HbA_1c from baseline to week 26 was statistically significantly greater with icodec than with glargine U100. A tipping point sensitivity analysis for the primary endpoint confirmed the robustness of the non‐inferiority findings for all missing data assumptions explored (Figure S3).

Mean fasting SMBG values decreased throughout the treatment period in both treatment groups (Figure S4). Over the first 4 weeks of treatment following the switch, there was no fasting hyperglycaemic peak with icodec (Figure 1B). Icodec demonstrated a gradual reduction in SMBG levels, from 168 mg/dL (9.3 mmol/L) at baseline to 140 mg/dL (7.8 mmol/L) at week 4; equivalent values for glargine U100 were 166 mg/dL (9.2 mmol/L) and 136 mg/dL (7.6 mmol/L), respectively.

The improvement in TIR from baseline (weeks −4–0) to end of treatment (weeks 22–26) was statistically significantly higher with icodec than with glargine U100 (estimated mean change: 18%‐points [equivalent to an increase of ~4 h 19 min] vs. 12%‐points [equivalent to an increase of ~2 h 54 min]; ETD: 6%‐points [95% CI: 2.6, 9.2], p = 0.0005). During weeks −4 to 0, the observed mean percentage of TIR was 44% for the icodec group, increasing to 64% during weeks 22–26. In the glargine U100 group, the mean percentage of TIR increased from 49% during weeks −4 to 0 to 60% during weeks 22–26 (Figure 1C).

During weeks −4 to 0, the observed mean percentage of TBR < 54 mg/dL (< 3.0 mmol/L) was low (0.2%) and within the internationally recommended target of below 1% in both treatment groups [12]. During weeks 22–26, there was no statistically significant difference between treatment groups in mean percentage of TBR < 54 mg/dL (< 3.0 mmol/L) (observed values: icodec, 0.3%; glargine U100, 0.2%; estimated treatment ratio: 1.2 [95% CI: 0.8, 1.7], p = 0.456).

The reduction in percentage of TAR from baseline (weeks −4 to 0) to weeks 22–26 was statistically significantly higher with icodec than with glargine U100 (estimated mean change: −18.5%‐points [equivalent to a decrease of ~4 h 26 min] vs. −12.4%‐points [equivalent to a decrease of ~2 h 59 min]; ETD: −6.2%‐points [95% CI: −9.5, −2.8], p = 0.0004). During weeks −4 to 0, the observed mean percentage of TAR was 55.3% in the icodec group, decreasing to 34.5% during weeks 22–26 (Figure 1C), versus a decrease from 50.3% to 39.4% in the glargine U100 group.

There were no statistically significant differences between treatment groups in the odds of achieving an HbA_1c < 7.0% at week 26 (estimated odds ratio [EOR]: 1.2 [95% CI: 0.7, 1.8], p = 0.5284) or of achieving the same HbA_1c target without clinically significant or severe hypoglycaemia during the previous 12 weeks (EOR: 1.1 [95% CI: 0.7, 1.8], p = 0.6554) (Figure S5).

3.3. Patient‐Reported Outcomes

From baseline to week 26, change in DTSQs total treatment satisfaction score was statistically significantly greater with icodec than with glargine U100 (ETD: 1.7 [95% CI: 0.7, 2.8]; p = 0.0009) (Figure 2A). The odds of achieving a clinically relevant improvement in DTSQs total treatment satisfaction score at week 26 were statistically significantly higher with icodec than with glargine U100 (odds ratio: 2.0 [95% CI: 1.1, 4.0]; p = 0.0354) (Figure 2B). There was also a statistically significantly greater improvement in TRIM‐D total score from baseline to week 26 with icodec than with glargine U100 (ETD: 2.9 [95% CI: 0.9, 4.9]; p = 0.0053) (Figure 2C).

FIGURE 2.

Participant‐reported outcomes. (A) Estimated change in DTSQs total treatment satisfaction score from baseline to week 26. The total treatment satisfaction score can range from 0 to 36, with higher scores indicating greater treatment satisfaction. The ANCOVA model included randomised treatment and region as fixed factors, and baseline value as a covariate. (B) Estimated proportion of participants achieving a clinically relevant improvement in DTSQs total treatment satisfaction score at week 26. The binary response was evaluated using a logistic regression model, with randomised treatment and region as fixed factors, and baseline value as a covariate. (C) Estimated change in TRIM‐D total score from baseline to week 26. The total score can range from 0 to 100, with higher scores indicating a better health state. The ANCOVA model included randomised treatment and region as fixed factors, and baseline value as covariate. *Statistically significant difference in favour of icodec. ANCOVA, analysis of covariance; CI, confidence interval; DTSQs, Diabetes Treatment Satisfaction Questionnaire status version; ETD, estimated treatment difference; glargine U100, insulin glargine U100; icodec, insulin icodec; OR, odds ratio; TRIM‐D, Treatment‐Related Impact Measure for Diabetes.

3.4. Safety Endpoints

The observed mean weekly total insulin dose in both treatment groups throughout the trial is shown in Figure 1D. During weeks 24–26, the estimated mean weekly insulin dose was 274 U/week (~39 U/day) with icodec and 270 U/week (~39 U/day) with glargine U100.

From baseline to week 26, the change in body weight was statistically significantly larger with icodec than with glargine U100 (estimated mean change: +2.0 kg vs. +0.5 kg; ETD: 1.5 kg [95% CI: 0.6, 2.3]; p = 0.0006).

The numbers and rates of overall hypoglycaemic episodes, nocturnal hypoglycaemic episodes, and AEs are reported in Table 2. During the first 4 weeks of treatment, there were 2 episodes of clinically significant hypoglycaemia with icodec versus 10 episodes with glargine U100. Overall, there were 53 episodes of clinically significant (level 2) hypoglycaemia in 18 participants receiving icodec and 70 episodes in 18 participants receiving glargine U100. The incidence of severe (level 3) hypoglycaemia was low in both treatment groups (two episodes in two participants in the icodec group and one episode in one participant in the glargine U100 group). The rate of combined clinically significant (level 2) or severe (level 3) hypoglycaemia was less than one episode per patient‐year of exposure (PYE) in both treatment groups (0.45 episodes per PYE with icodec vs. 0.59 episodes per PYE with glargine U100). Rates of nocturnal combined clinically significant (level 2) or severe (level 3) hypoglycaemia were 0.08 episodes per PYE with icodec versus 0.26 episodes per PYE with glargine U100.

TABLE 2.

Hypoglycaemia and other adverse events.

	Icodec (n = 206)				Glargine U100 (n = 206)				Treatment rate ratio (95% CI); p value
	n	%	E	R	n	%	E	R
Overall hypoglycaemic episodes
Hypoglycaemia alert value (level 1)	114	55.3	828	6.76	88	42.7	473	3.91	—
Clinically significant (level 2) hypoglycaemia	18	8.7	53	0.43	18	8.7	70	0.58	0.58 (0.22, 1.51); p = 0.2616
Severe (level 3) hypoglycaemia	2	1.0	2	0.02	1	0.5	1	0.008	—
Combined clinically significant (level 2) or severe (level 3) hypoglycaemia	20	9.7	55	0.45	19	9.2	71	0.59	0.59 (0.24, 1.48); p = 0.2620
Nocturnal hypoglycaemic episodes
Hypoglycaemia alert value (level 1)	43	20.9	113	0.92	38	18.4	119	0.98	—
Clinically significant hypoglycaemia (level 2)	6	2.9	9	0.07	8	3.9	31	0.26	0.29 (0.06, 1.34); p = 0.1137
Severe hypoglycaemia (level 3)	1	0.5	1	0.008	0	—	—	—	—
Combined clinically significant (level 2) or severe (level 3) hypoglycaemia	7	3.4	10	0.08	8	3.9	31	0.26	0.31 (0.07, 1.33); p = 0.1145
Adverse events
Total adverse events	135	65.5	360	2.94	123	59.7	302	2.50	—
Serious adverse events	16	7.8	20	0.16	12	5.8	24	0.20	—
Severe adverse events	8	3.9	10	0.08	10	4.9	15	0.12	—
Adverse events probably related to basal insulin	4	1.9	6	0.05	6	2.9	6	0.05	—
Adverse events possibly related to basal insulin	10	4.9	12	0.10	8	3.9	11	0.09	—
Serious adverse events probably related to basal insulin	1	0.5	1	0.008	1	0.5	4	0.003	—
Serious adverse events possibly related to basal insulin	1	0.5	1	0.008	0	—	—	—	—
Deaths	0	—	—	—	0	—	—	—	—

Note: Investigators were responsible for the detection and documentation of all adverse events from randomisation until end of trial. Hypoglycaemia alert value (level 1): blood glucose < 3.9 mmol/L and ≥ 3.0 mmol/L. Clinically significant (level 2) hypoglycaemia: blood glucose < 3.0 mmol/L. Severe (level 3) hypoglycaemia: no specific glucose threshold but hypoglycaemia was associated with severe cognitive impairment requiring external assistance for recovery. Nocturnal hypoglycaemic episodes occurred between 00:01 and 05:59 h inclusive.

Abbreviations: CI, confidence interval; E, number of episodes; glargine U100, insulin glargine U100; icodec, insulin icodec; n, number of participants experiencing at least one episode; PYE, patient‐year of exposure (1 PYE = 365.25 days); R, rate per PYE.

Overall, AEs were reported by 65.5% and 59.7% of participants in the icodec and glargine U100 groups, respectively. Most AEs were mild or moderate in severity (Table 2). The incidence of serious AEs was low and comparable between treatment groups (20 events in 16 participants with icodec and 24 events in 12 participants with glargine U100). The most frequently reported AEs in both treatment groups were within the system organ class of infections and infestations (icodec: 103 events; glargine U100: 81 events) (Table S3). During the first 4 weeks of treatment, there was one episode of hyperglycaemia with glargine U100 and no episodes with icodec. No deaths occurred and no new safety issues were identified in relation to icodec.

4. Conclusions

This 26‐week, randomised, phase 3b trial, which is the first to investigate switching participants from daily basal insulin to once‐weekly icodec without an initial one‐time additional dose, demonstrated a non‐inferior and statistically significantly greater reduction in HbA_1c versus glargine U100 in adults with T2D. Rates of combined clinically significant or severe hypoglycaemia were numerically lower with icodec than with glargine U100, in the context of overall low rates (less than one episode per PYE) in both treatment groups.

In the icodec group, no hyperglycaemic episodes were observed during the first 4 weeks of treatment. These data, along with the absence of a transient hyperglycaemic elevation in mean SMBG following the switch to icodec, suggest that omitting the initial one‐time additional dose would not result in increased SMBG levels from baseline. Previous modelling analysis based on ONWARDS 2 data showed a predicted mild transient increase in prebreakfast SMBG levels (approximately 0.4 mmol/L, 1–2 weeks in duration) after treatment initiation in individuals switching to icodec without versus with a one‐time 50% additional dose, reducing to matching levels by week 4 [9].

With icodec, TIR increased by 18%‐points (equivalent to 4 h 19 min) versus a statistically significantly lower increase of 12%‐points (2 h 54 min) with glargine U100. These data should be considered in the context of recent consensus guidelines which state that a change of 3%‐points in mean TIR between treatment groups is clinically meaningful [13]. The larger TIR increase seen with icodec versus glargine U100 may be a result of the longer half‐life of icodec, which could provide more consistent coverage across a 24‐h period than glargine U100, resulting in additional benefits on postprandial glucose levels. The mean percentage of TBR < 54 mg/dL (< 3.0 mmol/L) remained low in both treatment groups and, during weeks 22–26, mean percentage of TBR < 54 mg/dL (< 3.0 mmol/L) was well within the internationally recommended target of below 1% [12], with a mean time of less than 4 min/day. These results were similar to those previously reported in insulin‐experienced adults with T2D receiving icodec (ONWARDS 2: 0.2%) [4] or once‐weekly efsitora alfa (QWINT 3: 0.4%) [14].

The increase in body weight seen with icodec in this trial was broadly in line with that reported in a similar population in the ONWARDS 2 trial, which reported an increase of 1.40 kg after 26 weeks with icodec [4].

Statistically significant improvements in TRIM‐D total score and DTSQs total treatment satisfaction score from baseline to week 26 were observed in favour of icodec, accompanied by statistically significantly higher odds of achieving a clinically relevant improvement in DTSQs total treatment satisfaction score with icodec than with glargine U100. These results align with those previously reported in ONWARDS 2 and 5, which showed greater treatment satisfaction with once‐weekly icodec than with once‐daily basal insulins [4, 7, 15], highlighting the adherence potential of once‐weekly insulin regimens [16].

Strengths of this trial included the collection of blinded CGM data to avoid confounding bias and ensure minimal impact on the behaviour of participants not previously using such devices (all participants in this trial), a run‐in period allowing for baseline CGM data to be collected, the high proportion of participants completing the trial without permanently discontinuing treatment, inclusion of a multinational cohort, and the assessment of both statistical significance and clinical significance of the change in DTSQs total treatment satisfaction score.

Possible limitations included the open‐label design, which may have contributed to participant bias in the treatment satisfaction results, and the absence of a treatment control group, which included participants receiving a one‐time additional dose. The latter prevented a direct comparison between the switch to icodec with or without a one‐time additional dose. Finally, as with any randomised clinical trial, the present findings need to be further confirmed in clinical practice using real‐world studies.

In conclusion, switching from daily basal insulin to once‐weekly icodec without an initial one‐time additional dose resulted in a non‐inferior and statistically significantly greater HbA_1c reduction at 26 weeks with icodec than with glargine U100. The rate of combined clinically significant or severe hypoglycaemia was numerically lower with icodec than with glargine U100. This was accompanied by a greater increase in TIR with icodec than with glargine U100. No transient increase in prebreakfast SMBG levels was observed following the switch to icodec without a one‐time additional dose. Mean prebreakfast SMBG levels gradually reduced after switching to icodec, with comparable SMBG levels across treatment groups by week 3. Patient‐reported outcomes (measured by DTSQs and TRIM‐D) improved significantly in favour of icodec, accompanied by a clinically relevant advancement in DTSQs scores. These data indicate that switching from daily basal insulin to once‐weekly icodec without a one‐time additional dose provides effective glycaemic control, with no initial increase in glucose levels and without increasing hypoglycaemia risk. This provides reassurance for a simplified switching strategy from daily basal insulins to once‐weekly icodec.

Author Contributions

All authors participated in reviewing and interpreting the data outputs and in developing the manuscript. All authors had full access to all the data in the study and had final responsibility for the decision to submit for publication.

Funding

ONWARDS 10 was funded by Novo Nordisk.

Conflicts of Interest

J.R. has received research grants from Amgen, Applied Therapeutics, AstraZeneca, Biomea Fusion, Boehringer Ingelheim, Carmot, Corcept, Eli Lilly and Company, Hanmi Pharm, Merck, Novartis, Novo Nordisk, Oramed, Pfizer, Regeneron, and Sanofi; has served on scientific advisory boards and received honoraria or consulting fees from Amgen, Applied Therapeutics, Biomea Fusion, Eccogene, Eli Lilly and Company, Hanmi Pharm, Novo Nordisk, Oramed, Regeneron, Regor, Roche, Sanofi, and Structure Therapeutics; and has received honoraria for lectures from Eli Lilly and Company, Novo Nordisk, and Sanofi.

C.M. serves or has served on the advisory panel for Abbott, Bayer, Biomea Fusion, Boehringer Ingelheim, Dexcom, Eli Lilly and Company, Insulet, Medtronic, Novartis, Novo Nordisk, Roche, SAB Bio, Sanofi, Vertex, and vTv Therapeutics. Financial compensation for these activities has been received by KU Leuven; KU Leuven has received research support for C.M. from Medtronic, Novo Nordisk, and Sanofi; C.M. serves or has served on the speakers bureau for Abbott, Boehringer Ingelheim, Dexcom, Eli Lilly and Company, Insulet, Medtronic, Novo Nordisk, Sanofi, and Vertex. Financial compensation for these activities has been received by KU Leuven.

J.B.F., P.R., and C.S. are employees of Novo Nordisk and may hold stock options.

A.Y.Y.C. has received consulting fees and honoraria for advisory board participation and speaking from Abbott, Amgen, Aspen, Astellas, AstraZeneca, Bausch, Bayer, Biomea Fusion, Boehringer Ingelheim, Dexcom, Eisai, Eli Lilly and Company, Gan & Lee, GSK, HLS Therapeutics, Insulet, Medtronic, MSD, Novo Nordisk, Pfizer, Sandoz, Sanofi, and Vertex.

Supporting information

Table S1: Inclusion and exclusion criteria.

Table S2: Titration algorithm for adjustment of icodec and glargine U100.

Table S3: Most frequent (≥ 5%) adverse events by system organ class and preferred term during the on‐treatment period.

Figure S1: Trial design.

Figure S2: Participant disposition.

Figure S3: Tipping point analysis for primary endpoint.

Figure S4: SMBG levels over time.

Figure S5: Achievement of glycaemic targets at week 26.

Rosenstock J., Mathieu C., Feinberg J. B., Rao P., Scavenius C., and Cheng A. Y. Y., “Simplified Switching to Once‐Weekly Insulin Icodec Without an Initial One‐Time Additional Dose Versus Once‐Daily Insulin Glargine U100 in Basal Insulin‐Treated Type 2 Diabetes (ONWARDS 10): A Randomised Controlled Trial,” Diabetes, Obesity and Metabolism 28, no. 9 (2026): 7791–7801, 10.1111/dom.70813.

Data Availability Statement

Bona fide researchers may request access to data by submitting a research proposal for approval by Novo Nordisk and an independent review panel. Data sharing requests may be considered after research completion and primary publication, or, if the trial is intended for regulatory submission, after approval of product and product use in both the EU and USA. Individual participant data will be shared in data sets in anonymised format. Information about the data sharing process can be found at novonordisk‐trials.com.