top of page

Once-Weekly Insulin Icoden vs Once-Daily Insulin Degludec in Adults w/ Insulin-Naive Type 2 Diabetes

The ONWARDS 3 Randomized Clinical Trial

June 24, 2023. doi:10.1001/jama.2023.11313 (Excerpts)

Question: How does once-weekly insulin icodec compare with once-daily insulin degludec in glucose-lowering efficacy (hemoglobin A1c [HbA1c]) in people with insulin-naive type 2 diabetes?

Findings: In this randomized, double-masked, double-dummy trial that enrolled 588 people with type 2 diabetes, estimated mean HbA1c change from baseline to week 26 was noninferior with insulin icodec (−1.6 percentage points) compared with insulin degludec (−1.4 percentage points), with confirmed statistical superiority (estimated treatment difference, −0.2 percentage points).

Meaning: The results demonstrate the efficacy of once-weekly insulin icodec treatment in people with insulin-naive type 2 diabetes.


Importance Once-weekly insulin icodec could provide a simpler dosing alternative to daily basal insulin in people with type 2 diabetes.

Objective To evaluate the efficacy and safety of once-weekly icodec vs once-daily insulin degludec in people with insulin-naive type 2 diabetes.

Design, Setting, and Participants Randomized, double-masked, noninferiority, treat-to-target, phase 3a trial conducted from March 2021 to June 2022 at 92 sites in 11 countries in adults with type 2 diabetes treated with any noninsulin glucose-lowering agents with hemoglobin A1c (HbA1c) of 7%-11% (53-97 mmol/mol).

Trial Registration Identifier: NCT04795531


As type 2 diabetes progresses, insulin therapy may help to optimize glycemic control.1,2 Basal insulin initiation is recommended when noninsulin glucose-lowering agents are insufficient for glycemic control.

Insulin icodec is a once-weekly basal insulin that may improve treatment acceptance and adherence by reducing the number of basal insulin injections from at least 365 per year to 52 per year.

Clinical trials have demonstrated that icodec has a long half-life of approximately 1 week, with phase 2 trials in people with type 2 diabetes demonstrating similar glycemic efficacy and safety profiles for icodec and once-daily insulin glargine U100. ONWARDS is a phase 3a clinical trial program that comprises 6 trials investigating the efficacy and safety of icodec in various clinical scenarios.7 In the ONWARDS 4 and ONWARDS 2 trials, among participants with insulin-treated type 2 diabetes, hemoglobin A1c (HbA1c) reductions were similar or better with icodec than once-daily insulin glargine U100 or insulin degludec, with no statistically significant differences in level 2 or 3 hypoglycemia rates.

This study reports the results of the ONWARDS 3 trial, which assessed the efficacy and safety of once-weekly icodec vs once-daily degludec in people with insulin-naive type 2 diabetes.


Study Design and Participants

A detailed description of the ONWARDS program has been published previously. ONWARDS 3 was a randomized, double-masked, double-dummy, active-controlled, treat-to-target, phase 3a trial conducted between March 24, 2021, and June 23, 2022, across 92 sites in Argentina, Austria, Brazil, Canada, China, the Czech Republic, Denmark, France, Mexico, Taiwan, and the US. The trial comprised a 2-week screening period, a 26-week treatment period, and a 5-week follow-up period

Study Treatment

Participants were randomized in a 1:1 ratio using an interactive web-response system to receive either once-weekly icodec and once-daily placebo (icodec group) or once-daily degludec and once-weekly placebo (degludec group), with stratification by region (Asia, North America, South America, Europe) and use of sulfonylureas or glinides. Trial treatments were administered by subcutaneous injections via a pen device (PDS 290), with starting dosages of 70 U per week (icodec, 700 U/mL) or 10 U per day (degludec, 100 U/mL) . Dosages were adjusted weekly using the mean of 3 daily prebreakfast self-measured blood glucose (SMBG) values, measured on the 2 days before and on the day of titration using a blood glucose meter that was factory-calibrated to display plasma-equivalent glucose values. The mean prebreakfast SMBG target range was 80 to 130 mg/dL. Basal insulin dosages were increased by 20 U per week (icodec) or 3 U per day (degludec) if the mean SMBG value was above the target or decreased by the same amounts if the lowest SMBG value was under 80 mg/dL. Participants were instructed to measure their prebreakfast SMBG daily from week 0 to the end of the trial and any time they experienced symptoms suggestive of hypoglycemia. Unlimited glucose strips were provided to all participants and they were encouraged to take additional SMBG measurements at any time.



Of 737 screened participants, 588 (294 in the icodec group and 294 in the degludec group) were randomized (Figure 1). All randomized participants received at least 1 dose of study treatment, except 1 participant in the icodec group. Of those exposed to treatment, 12 participants in the icodec group and 11 in the degludec group permanently discontinued treatment; 96% of participants in each group completed the week 26 visit while receiving their assigned treatment. Demographic and baseline characteristics were comparable between treatment groups, although a higher percentage of participants in the icodec than in the degludec group were receiving sodium-glucose cotransporter 2 inhibitor or glucagon-like peptide 1 receptor agonist medications. The most common concomitant medications were metformin, sulfonylureas, and sodium-glucose cotransporter 2 inhibitors.


In this randomized controlled trial, once-weekly icodec was demonstrated to be both noninferior and superior to once-daily degludec in HbA1c reduction in people with type 2 diabetes who had not previously used insulin therapy, although the difference in HbA1c reduction of 0.2 percentage points after 26 weeks of treatment may have marginal clinical significance. Episodes of hypoglycemia were numerically higher with icodec than degludec in the context of overall low rates in both treatment groups, with no difference in weight change. When considering treatment with icodec insulin in clinical practice, the small added glycemic benefit and convenience of the once-weekly administration should be weighed against the small absolute risk of hypoglycemia.

The observed glucose-lowering efficacy of icodec and the statistical superiority in HbA1c reduction of icodec to degludec aligns with phase 2 trial results for icodec. Of note, the current trial extended the findings from the double-masked, double-dummy, phase 2 trial in several ways. First, the comparator insulin in the current trial was degludec, which has a lower risk of hypoglycemia than insulin glargine U100, the phase 2 trial comparator. Second, the population studied was expanded from people using metformin, with or without dipeptidyl peptidase 4 inhibitors, to a broader population of people eligible for basal insulin initiation, including those receiving a wider range of noninsulin glucose-lowering agents. Third, the target glucose level used for insulin titration was aligned with current recommendations from the American Diabetes Association for preprandial targets (80-130 mg/dL), compared with the more stringent titration target of 70 to 108 mg/dL in the phase 2 trial.

This trial demonstrated statistical superiority of icodec to degludec in HbA1c reduction from baseline to week 26. In contrast to the HbA1c reduction findings, fasting plasma glucose change at week 26 was similar in both treatment groups. Of note, prebreakfast SMBG values were similar between treatment groups, leading to comparable dose titration and total weekly insulin dose during the last 2 weeks of treatment. Further evaluation of continuous glucose monitoring data from other ONWARDS trials may clarify these findings.

In similar treat-to-target trials of current basal insulins in people with type 2 diabetes who were insulin-naive, level 2 or 3 hypoglycemia rates were higher (0.4-6.2 events per patient-year exposure) than those observed in this trial, although the different designs of these trials should be considered.4-6,11-15 The difference in hypoglycemia is in line with the greater HbA1c reduction observed in the icodec group compared with the degludec group. Previous evidence suggests that once-weekly insulin dosing regimens are viewed positively by people with type 2 diabetes16; a study of injectable glucagon-like peptide 1 receptor agonists demonstrated that once-weekly treatments for type 2 diabetes may improve treatment adherence and persistence compared with once-daily treatments.17 As such, icodec treatment may reduce delays in insulin initiation and may improve treatment adherence and convenience, given that it would reduce basal insulin injections from at least 365 per year to only 52 per year. Another once-weekly insulin, efsitora alfa, is also in clinical development18-20; phase 3 studies are currently ongoing.

Icodec has a similar safety profile to degludec, a basal insulin that has a low hypoglycemia risk and is approved for use in type 2 diabetes.21,22 Observed body weight changes were modest in both groups and aligned with expected increases for participants who are insulin-naive and initiate basal insulin.23 The number of injection site reactions was numerically higher with icodec than degludec. Concern over injection site reactions has not been raised in other icodec trials to date, and this finding remains unexplained.3-6,8,9 Additionally, although the occurrence of diabetic retinopathy was higher with icodec than degludec, the absolute numerical difference was small. No difference in retinopathy events between groups has been observed in other icodec trials to date. Strengths of this study included the double-masked, double-dummy design in people with type 2 diabetes who were insulin-naive and required treatment intensification with degludec as a comparator. Sulfonylureas and glinides were continued, although reduced by 50% to reflect potential real-world icodec use. The target sample size was achieved, and trial completion rate was high, ensuring that the trial was well-powered to assess its primary end point. Finally, randomization was stratified by sulfonylurea or glinide use at baseline, and analyses were adjusted for use of these treatments.


This trial has several limitations. First, the study had a 26-week duration and sustained effects were not studied. Second, there was the potential for selection bias owing to the number of injections; however, compared with regimens for daily basal insulin, the regimen in this trial only required 1 extra injection per week (ie, 7 once-daily and 1 once-weekly injections compared with 7 once-daily injections). Third, this trial did not collect continuous glucose monitoring data or data on patient-reported outcomes. Fourth, to allow comparison between the treatment groups, both insulins were titrated weekly; consequently, the dose adjustment frequency was not tailored to the dosing frequency of each treatment type. Fifth, the trial was powered to assess the primary outcome; therefore, any lack of statistically significant differences between groups for secondary outcomes, including hypoglycemia, do not necessarily reflect a lack of clinical effect.


Among people with insulin-naive type 2 diabetes, once-weekly icodec demonstrated superior HbA1c reduction to once-daily degludec after 26 weeks of treatment, with no difference in weight change and a higher rate of combined level 2 (clinically significant) or level 3 (severe) hypoglycemic events in the context of less than 1 event per patient-year exposure in both groups.

To View Report: Published Online: June 24, 2023. doi:10.1001/jama.2023.11313

bottom of page