Was DEVOTE designed to show that degludec is better for the heart than glargine?

No. DEVOTE was a non-inferiority trial required by FDA guidance. It was powered to rule out excess cardiovascular risk from degludec (upper 95% CI below 1.3), not to demonstrate superiority. The study found non-inferior CV outcomes but did not show a statistically significant benefit.

Why was glargine U100 the comparator instead of a newer insulin?

When DEVOTE was designed in 2013 to 2014, glargine U100 was the global standard basal insulin. Glargine U300 (Toujeo) had not yet completed its key EDITION program. The regulatory and clinical relevance of glargine U100 made it the appropriate reference for a cardiovascular outcomes trial.

How was hypoglycemia defined in DEVOTE?

Severe hypoglycemia required external assistance to treat. Nocturnal hypoglycemia was defined as an episode occurring between midnight and 5:59 AM, confirmed by self-monitored plasma glucose below 56 mg/dL. Both were prespecified secondary endpoints with multiplicity adjustment.

Did both groups achieve the same blood sugar control?

Yes. Mean HbA1c was 7.5% in both arms at trial end. The treat-to-target design intentionally equalized glycemic outcomes so that any differences in CV events or hypoglycemia could be attributed to pharmacokinetic properties of the insulins rather than glucose levels.

How long were patients followed in DEVOTE?

Median follow-up was 1.99 years. The trial was event-driven, meaning it continued until 633 MACE events accumulated. This is shorter than many cardiovascular outcome trials (ORIGIN followed patients for 6.2 years), which limits conclusions about long-term safety.

Does DEVOTE apply to patients with type 1 diabetes?

No. DEVOTE enrolled only patients with type 2 diabetes and high cardiovascular risk. Hypoglycemia patterns and cardiovascular risk profiles differ substantially in type 1 diabetes, so these findings should not be extrapolated.

Is the 40% reduction in severe hypoglycemia clinically meaningful?

In a population with established cardiovascular disease, severe hypoglycemia is associated with arrhythmia and increased mortality risk. A 40% relative reduction (rate ratio 0.60 to 95% CI 0.48, 0.76) achieved in a double-blind, treat-to-target trial represents a clinically significant safety improvement.

How does DEVOTE compare to other insulin cardiovascular outcome trials?

ORIGIN (2012) tested glargine vs standard care over 6.2 years and found neutral CV outcomes. DEVOTE is the only completed head-to-head insulin CVOT comparing two basal insulin analogs. No equivalent CVOT exists for glargine U300, insulin detemir, or concentrated insulins.

Did background diabetes medications affect the results?

DEVOTE allowed standard-of-care adjustments, but enrollment (2013 to 2016) preceded widespread use of SGLT2 inhibitors and GLP-1 agonists with proven CV benefit. Whether degludec's relative safety profile changes when combined with these newer agents has not been studied in a dedicated trial.

What did the FDA conclude from DEVOTE?

The FDA accepted DEVOTE as evidence of cardiovascular safety for insulin degludec. The prescribing information for Tresiba (degludec) references the DEVOTE data, and the label includes the demonstrated reduction in severe hypoglycemia compared to glargine U100.

Inside the DEVOTE Methodology: What Most Summaries Skip

By HealthRX Medical Team

Published May 25, 2026Updated May 25, 2026Last reviewed May 25, 2026

At a glance

| Detail | Value | |---|---| | N | 7,637 | | Intervention | Insulin degludec, once daily | | Comparator | Insulin glargine U100, once daily | | Duration | Median 1.99 years (event-driven) | | Primary endpoint | Time to first MACE (CV death, non-fatal MI, non-fatal stroke) | | Key result | HR 0.91 (95% CI 0.78, 1.06); non-inferiority confirmed at 1.3 margin | | Hypoglycemia finding | 53% lower rate of severe nocturnal hypoglycemia with degludec |

Why the Trial Design Matters More Than the Headline

Most summaries of DEVOTE stop at two bullet points: non-inferior for MACE, less nocturnal hypoglycemia. That framing obscures the methodological choices that determine what the data can and cannot tell clinicians. This page unpacks those choices.

The Regulatory Backdrop: FDA's 2008 CVOT Mandate

After the rosiglitazone controversy, the FDA issued guidance in 2008 requiring new glucose-lowering therapies to rule out excess cardiovascular risk. The standard: the upper bound of the 95% confidence interval for the hazard ratio must fall below 1.3 for a composite of major adverse cardiovascular events. DEVOTE was designed specifically to satisfy this threshold. That context is essential. The trial was powered to exclude harm, not to demonstrate benefit. A hazard ratio of 0.91 looks favorable, but the confidence interval (0.78, 1.06) includes 1.0. The study was never sized to prove superiority.

Population: Who Got In and Who Didn't

DEVOTE enrolled adults aged 50 or older with established cardiovascular disease, chronic kidney disease, or at least one CV risk factor. Patients needed to have been on at least one insulin for 90 days or more, or to be insulin-naive with an HbA1c above 7.0%. The trial ran across 438 sites in 20 countries, as reported in the primary publication.

Key exclusion criteria included recent acute coronary or cerebrovascular events within 60 days, planned coronary or carotid revascularization, and New York Heart Association class IV heart failure. These exclusions removed the sickest patients from the denominator, which improves event adjudication consistency but limits how far the results generalize to real-world insulin users with acute decompensation.

Roughly 85% of enrolled participants had established CV disease. The remaining 15% qualified on risk factors alone. This skew toward secondary prevention means DEVOTE tells us more about insulin safety in patients who have already had a cardiovascular event than it does about primary prevention populations.

Randomization and Blinding

Patients were randomized 1:1 to degludec or glargine U100 using an interactive voice/web response system, stratified by site. Both insulins were supplied in identical prefilled pens to maintain double-blinding, a meaningful logistical achievement given that degludec and glargine have different pen devices in commercial packaging.

HealthRX Methodological Integrity Checklist: DEVOTE

| Design Element | DEVOTE Implementation | Strength / Limitation | |---|---|---| | Blinding | Double-blind, identical pens | Eliminates prescriber and patient bias in dose titration | | Randomization | 1:1, stratified by site | Adequate; no demographic stratification reported | | Treat-to-target | FPG target <71 to 90 mg/dL for both arms | Equalizes glycemic exposure, isolates non-glycemic effects | | Event-driven stopping | 633 MACE events required | Efficient but compresses follow-up in high-event populations | | Adjudication | Independent blinded committee | Standard for CVOTs; reduces ascertainment bias | | Comparator | Active (glargine U100), not placebo | Clinically relevant but raises the bar for separation | | Non-inferiority margin | Upper 95% CI <1.3 | FDA-mandated; does not test for benefit | | Hypoglycemia assessment | Prespecified secondary, confirmed by SMPG <56 mg/dL | Objective threshold, but relies on patient reporting of events |

This framework highlights a recurring tension in insulin CVOTs. The treat-to-target protocol is essential for isolating cardiovascular effects from glycemic confounding, yet it also constrains the trial's ability to detect differences that emerge only when insulin pharmacokinetics influence real-world dosing patterns. In clinical practice, patients do not titrate with trial-level precision.

The Treat-to-Target Protocol and Its Implications

Both arms titrated basal insulin to a fasting plasma glucose (FPG) target of 71 to 90 mg/dL. This is critical for interpretation. If both groups reach similar HbA1c levels (which they did: mean HbA1c was 7.5% in both arms at end of trial), any difference in cardiovascular or hypoglycemia outcomes cannot be attributed to glucose control itself. The design deliberately removes glycemic separation so that the comparison reflects pharmacokinetic and pharmacodynamic differences between the two molecules.

The achieved HbA1c parity also means that if a clinician's primary question is "will switching from glargine to degludec improve my patient's glucose control," DEVOTE does not answer that question. It answers a different one: "at similar glucose targets, is degludec at least as safe from a CV standpoint, and does it cause less hypoglycemia?"

The Primary Endpoint: Three-Point MACE

DEVOTE used the standard FDA three-point MACE composite: cardiovascular death, non-fatal myocardial infarction, and non-fatal stroke. Events were adjudicated by an independent, blinded committee using standardized definitions.

The trial was event-driven, requiring 633 confirmed MACE events before the database could lock. This approach is statistically efficient. It also means that the duration of follow-up is a consequence, not a design input. The median follow-up of 1.99 years is relatively short for a cardiovascular outcome study. For context, the ORIGIN trial comparing insulin glargine to standard care followed patients for a median of 6.2 years.

A shorter observation window may undercount slow-developing processes like atherosclerotic plaque progression or the cumulative metabolic impact of recurrent hypoglycemia. The DEVOTE authors acknowledged this limitation. The trial reliably excludes acute cardiovascular toxicity from degludec but says less about decade-scale safety.

Results: What the Numbers Actually Show

Primary Outcome (MACE)

| Outcome | Degludec (n = 3,818) | Glargine (n = 3,819) | HR (95% CI) | |---|---|---|---| | First MACE | 325 events (8.5%) | 356 events (9.3%) | 0.91 (0.78, 1.06) | | CV death | 136 (3.6%) | 142 (3.7%) | 0.96 (0.76, 1.21) | | Non-fatal MI | 152 (4.0%) | 179 (4.7%) | 0.85 (0.68, 1.06) | | Non-fatal stroke | 57 (1.5%) | 58 (1.5%) | 0.98 (0.68, 1.41) |

Non-inferiority was confirmed (p <0.001 for non-inferiority at the 1.3 margin). Superiority was not demonstrated (p = 0.25). The point estimates favor degludec across all three components, but none of the individual components reached statistical significance. This pattern is consistent with a true null effect or with a small benefit that the study was underpowered to detect on a component level.

Hypoglycemia

| Category | Degludec | Glargine | Rate ratio (95% CI) | |---|---|---|---| | Severe hypoglycemia | 187 events in 123 patients | 252 events in 156 patients | 0.60 (0.48, 0.76) | | Severe nocturnal hypoglycemia | 19 events in 16 patients | 45 events in 37 patients | 0.47 (0.31, 0.73) |

Severe hypoglycemia was reduced by 40%. Severe nocturnal hypoglycemia was reduced by 53%. Both were prespecified secondary endpoints with multiplicity-adjusted significance, as detailed in the trial's statistical analysis plan. The nocturnal finding aligns with the known ultra-long, flat pharmacokinetic profile of degludec, which produces less peak-trough variability than glargine U100.

These hypoglycemia findings carry clinical weight. Severe hypoglycemia in patients with established cardiovascular disease is associated with arrhythmia, QT prolongation, and excess mortality in observational analyses. A 40% relative reduction in a double-blind, treat-to-target design is a meaningful separation.

The Comparator Choice: Why Glargine U100 and Not U300

DEVOTE compared degludec to glargine U100 (Lantus), not to the concentrated glargine U300 (Toujeo) formulation. This matters because glargine U300 has its own data suggesting flatter pharmacokinetics and lower hypoglycemia rates compared to U100, as seen in the EDITION trials. Had DEVOTE used glargine U300 as the comparator, the hypoglycemia separation might have been smaller.

The choice of U100 was pragmatic. At the time of trial design (2013 to 2014), glargine U100 was the dominant basal insulin worldwide and the standard reference for regulatory comparison. The FDA label for insulin degludec reflects DEVOTE data against this specific comparator.

Statistical Approach: The Cox Model and What It Assumes

The primary analysis used a Cox proportional-hazards model with treatment as the sole covariate. The proportional-hazards assumption (that the relative risk of MACE remains constant over time) was not violated based on Schoenfeld residual testing reported in the supplementary materials.

The non-inferiority analysis used the full intention-to-treat population. A per-protocol sensitivity analysis yielded similar results. There was no hierarchical testing for superiority on the primary endpoint conditional on non-inferiority success. The protocol specified that if non-inferiority was met, superiority would be tested, but the p-value of 0.25 made this a formality.

For the hypoglycemia endpoints, the analysis used a negative binomial regression model with log-transformed exposure time as an offset, accounting for the variable follow-up inherent in an event-driven design.

Limitations the Authors Acknowledged

The DEVOTE investigators noted several limitations explicitly. The trial population was enriched for CV risk, limiting generalizability to lower-risk patients with type 2 diabetes. The treat-to-target design, while essential for isolating non-glycemic effects, does not reflect real-world insulin management where titration adherence varies. The relatively short follow-up (median 1.99 years) leaves open questions about longer-term safety signals.

One limitation the authors did not emphasize: the background medication profile. By 2017, SGLT2 inhibitors and GLP-1 receptor agonists with proven CV benefit (from EMPA-REG OUTCOME and LEADER) were entering widespread use. DEVOTE enrolled between 2013 and 2016, so use of these agents in the trial population was lower than it would be today. Whether degludec's CV profile holds when layered on top of modern cardioprotective regimens remains unstudied in a dedicated trial.

What DEVOTE Does and Doesn't Tell Prescribers

DEVOTE confirms that switching from glargine U100 to degludec does not increase cardiovascular risk. It provides strong evidence that such a switch reduces severe and nocturnal hypoglycemia at equivalent glycemic targets. It does not prove cardiovascular superiority. It does not address the comparison with glargine U300 or biosimilar glargines. It does not answer whether degludec's hypoglycemia advantage persists in patients already on modern cardioprotective therapies.

For prescribers weighing a formulary switch, the actionable signal from DEVOTE is the hypoglycemia reduction, not a CV benefit. The American Diabetes Association Standards of Care cite DEVOTE when recommending degludec as a basal insulin option for patients at high risk of hypoglycemia.

Frequently asked questions

›

References

Marso SP, McGuire DK, Zinman B, et al. Efficacy and safety of degludec versus glargine in type 2 diabetes. N Engl J Med. 2017;377(8):723-732. PubMed
FDA Guidance for Industry: Diabetes Mellitus, Evaluating Cardiovascular Risk in New Antidiabetic Therapies to Treat Type 2 Diabetes. FDA.gov
Tresiba (insulin degludec) prescribing information. FDA Label
ORIGIN Trial Investigators. Basal insulin and cardiovascular and other outcomes in dysglycemia. N Engl J Med. 2012;367(4):319-328. PubMed
Riddle MC, Bolli GB, Ziemen M, et al. New insulin glargine 300 units/mL versus glargine 100 units/mL in people with type 2 diabetes using basal and mealtime insulin: glucose control and hypoglycemia in a 6-month randomized controlled trial (EDITION 1). Diabetes Care. 2014;37(10):2755-2762. PubMed
American Diabetes Association. Standards of Care in Diabetes, 2024. Diabetes Care. 2024;47(Suppl 1). ADA Standards