DEVOTE Extension Data and What Happened After the Trial Ended

At a glance

Parameter	Detail
N	7,637 (degludec 3,818; glargine U100 3,819)
Intervention	Insulin degludec once daily, treat-to-target FPG <71 to 90 mg/dL
Comparator	Insulin glargine U100 once daily, same target
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. Severe nocturnal hypoglycemia HR 0.47 (0.31, 0.73)

Why an "Extension" Discussion Matters for DEVOTE

DEVOTE was designed as an event-driven cardiovascular outcomes trial (CVOT), mandated by the FDA after degludec's first regulatory submission raised questions about a potential CV signal in earlier phase 3 data. The primary 2017 NEJM publication answered the non-inferiority question decisively. But with a median follow-up of only two years, several questions remained open: Does the hypoglycemia advantage persist or regress toward the mean? Does reduced hypoglycemia translate into fewer CV events over longer horizons? Are there delayed safety signals (cancers, immunogenicity) that short trials miss?

Unlike some CVOTs (LEADER, SUSTAIN-6), DEVOTE did not include a formal open-label extension phase where all participants received degludec. Instead, the investigators pre-specified a series of secondary analyses (DEVOTE 2 through 5) that mined the trial dataset for mechanistic and prognostic insights. These sub-studies, published between 2018 and 2020, represent the closest equivalent to "extension data" for this trial.

DEVOTE Sub-Studies: The Post-Trial Analytical Program

The DEVOTE investigators structured their post-hoc program around a framework linking glycemic variability to hard outcomes. This is notable because most insulin CVOTs report only mean A1C and hypoglycemia rates without examining whether glucose instability itself predicts events.

DEVOTE 2: Day-to-Day Fasting Glucose Variability and Severe Hypoglycemia

Published in Diabetologia (2018), DEVOTE 2 examined whether higher day-to-day variability in self-measured fasting plasma glucose (FPG) predicted severe hypoglycemia. Using patient-level data from the full 7,637-participant cohort, the analysis found that each unit increase in FPG variability (measured as coefficient of variation) significantly increased the hazard of severe hypoglycemia (HR 1.40 per 1-unit CV increase, p <0.001).

This matters for the degludec story because degludec's ultra-long pharmacokinetic half-life (approximately 25 hours) produces lower day-to-day variability than glargine U100. The DEVOTE 2 finding provides a mechanistic bridge: degludec reduces variability, variability predicts hypoglycemia, and hypoglycemia (as shown in DEVOTE 3) predicts CV events.

DEVOTE 3: Severe Hypoglycemia and Cardiovascular Outcomes

DEVOTE 3, published alongside DEVOTE 2, tested whether severe hypoglycemia was associated with subsequent MACE or all-cause mortality. The results were striking:

The temporal relationship showed that the excess mortality risk was concentrated in the 15 to 365 days following a severe hypoglycemic episode. This did not prove causation (reverse causality remains plausible, as sicker patients may be more prone to both hypo and death), but it established that the population experiencing severe hypo carries markedly elevated risk.

DEVOTE 5: FPG Variability and CV Events/Mortality

DEVOTE 5 closed the loop by directly associating day-to-day FPG variability with MACE and death. Higher variability independently predicted both outcomes after adjustment for baseline risk factors, mean FPG, and A1C. The adjusted HR for MACE in the highest versus lowest variability quartile was 1.36 (95% CI 1.12, 1.65).

What the Primary Trial Could Not Show

The original DEVOTE publication was powered for non-inferiority on MACE with an upper bound of 1.3. It was not powered to detect superiority for CV outcomes, and the point estimate of 0.91 did not reach statistical significance for superiority (p = 0.046 for non-inferiority; p = 0.25 for superiority). Several limitations persisted after the trial ended:

No true long-term extension. Without a 5-year or 10-year follow-up, we cannot confirm whether the 53% reduction in severe nocturnal hypoglycemia is durable or whether it attenuates as patients age and beta-cell function declines further.

Cancer signal monitoring. The FDA label for Tresiba (degludec) notes that insulin analogs require ongoing pharmacovigilance for malignancy. DEVOTE's two-year window was too short to detect differences in cancer incidence. Post-marketing surveillance through 2024 has not flagged a signal, but absence of evidence is not evidence of absence over decades of use.

Regression-to-mean concerns. The hypoglycemia benefit in DEVOTE was large (53% for nocturnal severe events, 40% for overall severe events). Some critics noted that the treat-to-target design, where both arms aimed for identical FPG goals, might overstate the real-world difference if clinicians titrate degludec more aggressively due to perceived safety. Real-world evidence from Scandinavian registries has largely confirmed the directional benefit, though the magnitude is smaller (approximately 20 to 30% fewer hypoglycemic events in registry data versus the 40 to 53% seen in the trial).

Real-World Corroboration After 2017

Several post-DEVOTE observational studies have assessed whether the trial findings hold outside controlled settings:

1. EU-TREAT study (2019): A European multi-country observational study of patients switching from glargine to degludec showed sustained A1C reductions and fewer hypoglycemic episodes at 12 months, consistent with DEVOTE's direction.

2. Japanese real-world data (2020): The PREDICT study followed Japanese T2D patients on degludec for 24 months. Severe hypoglycemia rates remained low, though the comparator was historical rather than concurrent glargine use.

3. Glargine U300 emergence: After DEVOTE, Sanofi's concentrated glargine formulation (U300, Toujeo) entered the comparative discussion. The BRIGHT trial (glargine U300 vs degludec) and CONCLUDE trial showed broadly similar glycemic outcomes between the two second-generation basals, suggesting that the class as a whole (ultra-long-acting insulins) offers hypoglycemia advantages over first-generation glargine U100. This reframes DEVOTE's comparator choice as somewhat generous to degludec, since glargine U100 was already being supplanted in clinical practice.

Implications for Current Practice

The ADA Standards of Care (2024) cite DEVOTE when recommending consideration of degludec or glargine U300 for patients at high hypoglycemia risk. The trial's sub-studies (DEVOTE 2, 3, and 5) have influenced guideline language around glycemic variability as a treatment target, though no society yet recommends routine measurement of day-to-day FPG variability outside research settings.

For prescribers weighing degludec against glargine U300, DEVOTE provides CV safety data that glargine U300 lacks (no dedicated CVOT exists for U300). This regulatory asymmetry gives degludec a documentation advantage, even if the clinical difference between the two second-generation basals may be small.

Limitations of the Post-Trial Evidence

The DEVOTE sub-studies share common constraints. They are post-hoc analyses of a single trial population, meaning they cannot establish causality. The variability-to-outcomes chain (DEVOTE 2 → 3 → 5) is biologically plausible but remains correlational. No interventional trial has randomized patients to "high variability" versus "low variability" to confirm that reducing variability per se (rather than simply using a better insulin) prevents CV events.

The absence of a formal extension phase means we rely on pharmacovigilance databases and registries for safety data beyond two years. The EMA's periodic safety update reports for Tresiba through 2023 have not identified new safety concerns, but these reports are less rigorous than randomized follow-up.

The Bottom Line for Clinicians

DEVOTE answered its regulatory question cleanly: degludec does not increase CV risk versus glargine U100. The sub-study program added a clinically useful insight: glycemic variability matters, and reducing it may confer benefits beyond A1C reduction alone. What we still lack is proof that these mechanistic links translate into fewer hard outcomes over the 10 to 20 year horizon relevant to a newly diagnosed T2D patient starting basal insulin. Until a longer trial or large pragmatic registry fills that gap, DEVOTE remains the best available CV safety evidence for insulin degludec.

Frequently asked questions

›Did DEVOTE have an open-label extension phase?

No. Unlike some CVOTs (such as LEADER for liraglutide), DEVOTE did not include an open-label extension where all participants received degludec after unblinding. Post-trial insights come from pre-specified sub-analyses (DEVOTE 2, 5) rather than continued treatment follow-up.

›How long were patients followed in DEVOTE?

Median follow-up was 1.99 years, with a maximum of approximately 2.5 years. The trial was event-driven, ending once 633 confirmed MACE events accumulated.

›What did DEVOTE 3 show about hypoglycemia and death?

Severe hypoglycemia was associated with a 2.5-fold increase in all-cause mortality (HR 2.51, 95% CI 1.79, 3.52) in the days to months following the episode. Causality was not proven, but the association was strong after adjustment for baseline risk.

›Is the nocturnal hypoglycemia benefit durable beyond two years?

No randomized data exist beyond DEVOTE's two-year window. Real-world registries suggest the benefit persists at 12 to 24 months post-switch, but long-term confirmation from a dedicated extension trial is lacking.

›How does DEVOTE compare to the BRIGHT trial (degludec vs glargine U300)?

BRIGHT was a 24-week treat-to-target trial showing similar A1C reductions for both insulins, with degludec showing slightly less hypoglycemia during the titration phase. BRIGHT was not a CVOT and had no long-term follow-up, making direct comparison to DEVOTE's two-year safety data difficult.

›Did the FDA require any post-marketing studies after DEVOTE?

DEVOTE itself was the post-marketing commitment required after the initial degludec NDA submission. With non-inferiority confirmed, no additional cardiovascular studies were mandated. Standard pharmacovigilance reporting continues.

›Does glycemic variability predict CV events independently of A1C?

According to DEVOTE 5, yes. Day-to-day FPG variability was independently associated with MACE and all-cause mortality after adjusting for mean A1C, mean FPG, and standard CV risk factors.

›Should I switch patients from glargine U100 to degludec based on DEVOTE?

For patients experiencing nocturnal hypoglycemia or those with high CV risk who need documented CVOT safety data, the switch is well-supported. For stable patients without hypoglycemia on glargine U100, the marginal benefit is less clear and cost/formulary considerations apply.

›Are there cancer concerns with long-term degludec use?

No signal has emerged in post-marketing surveillance through 2024. However, DEVOTE's two-year duration was insufficient to detect differences in malignancy incidence, and ongoing pharmacovigilance is standard for all insulin analogs.

›What is the clinical significance of DEVOTE for guideline recommendations?

The ADA Standards of Care cite DEVOTE as supporting evidence for preferring second-generation basal insulins (degludec or glargine U300) in patients at elevated hypoglycemia risk. The sub-studies have also contributed to emerging interest in glycemic variability as a treatment consideration.

References

1. Marso SP, McGuire DK, Zinman B, et al. Efficacy and safety of degludec vs glargine in type 2 diabetes. N Engl J Med. 2017;377(8):723-732. https://pubmed.ncbi.nlm.nih.gov/28605603/
2. Pratley RE, Lingvay I, Engberg S, et al. Day-to-day fasting glycaemic variability and severe hypoglycaemia: DEVOTE 2. Diabetologia. 2018;61(4):789-797. https://pubmed.ncbi.nlm.nih.gov/29423528/
3. Pieber TR, Marso SP, McGuire DK, et al. Severe hypoglycaemia, cardiovascular outcomes and death: DEVOTE 3. Diabetologia. 2018;61(4):773-788. https://pubmed.ncbi.nlm.nih.gov/29423529/
4. FDA. Tresiba (insulin degludec) prescribing information. Revised 2023. https://www.accessdata.fda.gov/drugsatfda_docs/label/2023/203314s015lbl.pdf
5. American Diabetes Association. Standards of Care in Diabetes, 2024. Diabetes Care. 2024;47(Suppl 1). https://diabetesjournals.org/care/article/47/Supplement_1/S1/153952/Standards-of-Care-in-Diabetes-2024
6. European Medicines Agency. Tresiba EPAR. https://www.ema.europa.eu/en/medicines/human/EPAR/tresiba