Tresiba EMA vs FDA Approach: How Regulators Diverged on Insulin Degludec

Two Agencies, Two Timelines

The EMA and FDA reached opposite conclusions on insulin degludec in the same quarter of 2013. The EMA's Committee for Medicinal Products for Human Use (CHMP) recommended approval on October 18, 2012, and the European Commission granted marketing authorization on January 21, 2013 1. The FDA, reviewing the same Phase 3 dataset from the BEGIN trial program, issued a Complete Response Letter (CRL) in February 2013, citing concerns about cardiovascular signal imbalance in a pre-specified meta-analysis of adjudicated major adverse cardiovascular events (MACE) 2.

That CRL delayed U.S. access by more than two and a half years. The divergence was not about efficacy. Both agencies acknowledged that degludec reduced HbA1c comparably to insulin glargine U100 across the BEGIN trials. The disagreement centered on how much cardiovascular uncertainty a regulator should tolerate at the point of initial approval, and whether a post-marketing commitment could resolve it or whether resolution had to precede market entry.

The FDA's position reflected its 2008 guidance requiring cardiovascular outcomes data for new diabetes therapies, a policy born from the rosiglitazone controversy 3. The EMA had adopted a similar but less prescriptive position, accepting that the BEGIN meta-analysis point estimate (HR 1.30 for MACE) fell within the upper bound of 1.8 permitted under its 2012 guideline, while requiring Novo Nordisk to conduct a dedicated CVOT as a post-authorization measure 4.

The FDA's 2008 CV Guidance and Its Impact on Degludec

After the FDA published its 2008 industry guidance on evaluating cardiovascular risk in antidiabetic drugs, every new glucose-lowering agent faced a binary gate: demonstrate that the upper bound of the 95% confidence interval for MACE does not exceed 1.8 pre-approval (or 1.3 post-approval) 3. Insulin degludec's pre-approval meta-analysis, pooling MACE events across the BEGIN program, returned a point estimate of 1.30. The upper bound of the confidence interval crossed 1.8.

That crossing triggered the CRL. The FDA's Endocrinologic and Metabolic Drugs Advisory Committee did not vote against approval outright, but the agency concluded that the signal warranted prospective adjudication rather than reliance on pooled Phase 3 data with limited event counts. Novo Nordisk had already initiated the DEVOTE trial (NCT01959529) in 2013 in anticipation of this requirement, enrolling 7,637 patients with type 2 diabetes at high cardiovascular risk across 438 sites in 20 countries 4.

The FDA ultimately approved Tresiba on September 25, 2015, before DEVOTE completed, after reviewing updated interim safety data and a revised risk assessment. This was an unusual move. The agency accepted a conditional framework: approval now, with DEVOTE completion as a post-marketing requirement (PMR) 2.

The EMA's Parallel Path

The EMA's approach to degludec was more permissive at initial authorization but no less rigorous in post-market follow-up. The CHMP assessment report acknowledged the cardiovascular signal in the BEGIN meta-analysis but judged it within the bounds of statistical noise given low event counts. The committee weighted the drug's clinical advantages: a half-life of approximately 25 hours (vs. roughly 12 hours for glargine U100), lower day-to-day glycemic variability, and a 36% reduction in confirmed nocturnal hypoglycemia versus glargine across the BEGIN program 1.

The EMA required DEVOTE as a post-authorization efficacy study (PAES), meaning the trial result would feed directly into periodic benefit-risk re-evaluation. European prescribers had access to degludec from mid-2013. This gave the EMA a head start on real-world pharmacovigilance data through its EudraVigilance system, accumulating signal detection experience roughly 30 months before the drug entered the U.S. market.

One practical consequence: by the time DEVOTE reported in June 2017, the EMA had already processed multiple Periodic Safety Update Reports (PSURs) and could cross-reference trial findings against real-world adverse event reporting. The FDA, which relies on its Sentinel system and FAERS database for post-market surveillance, had a shorter window of U.S. prescribing data at that point 5.

DEVOTE Trial Results and Their Regulatory Consequences

DEVOTE was a double-blind, treat-to-target trial randomizing 7,637 patients to insulin degludec or insulin glargine U100 4. The primary outcome was time to first MACE (cardiovascular death, non-fatal myocardial infarction, or non-fatal stroke). Median follow-up was 1.99 years.

The results settled the cardiovascular question definitively. Degludec produced a MACE hazard ratio of 0.91 (95% CI 0.78 to 1.06, P<0.001 for non-inferiority). The pre-specified non-inferiority margin was 1.3. The trial also confirmed a significant reduction in severe hypoglycemia: degludec reduced rates by 40% compared to glargine U100 (rate ratio 0.60, P<0.001 for superiority), with an even larger 53% reduction in nocturnal severe hypoglycemia 4.

Both agencies updated their assessments after DEVOTE. The FDA revised the Tresiba prescribing information to include cardiovascular safety data from the trial, fulfilling the post-marketing requirement 2. The EMA updated the Summary of Product Characteristics (SmPC) and closed the PAES obligation 1.

Dr. Steven Marso, the DEVOTE trial's lead investigator, stated in the NEJM publication: "Insulin degludec was noninferior to insulin glargine with respect to the incidence of major adverse cardiovascular events" 4. The Endocrine Society's 2019 clinical practice guideline on insulin therapy noted that ultra-long-acting basal insulins, including degludec, offer "a lower risk of hypoglycemia compared with other basal insulin analogs," citing DEVOTE as primary evidence 6.

Labeling Differences: FDA PI vs EMA SmPC

The FDA prescribing information (PI) and EMA Summary of Product Characteristics (SmPC) for Tresiba share core pharmacokinetic and dosing data but differ in structure, emphasis, and specific indications.

The FDA label (revised through 2023) includes a dedicated "Clinical Studies" section reporting DEVOTE outcomes data, severe hypoglycemia rates, and HbA1c reductions from the BEGIN program. It carries a boxed warning common to all insulins about the risk of hypoglycemia, and the Warnings and Precautions section addresses hypokalemia, fluid retention with thiazolidinedione co-administration, and the need for dose adjustment in renal or hepatic impairment 2.

The EMA SmPC includes a Risk Management Plan (RMP) summary and lists post-authorization studies including DEVOTE. It specifies approved indications for adults, adolescents, and children from age 1 year with type 1 or type 2 diabetes, whereas the FDA label initially approved Tresiba for adults only, adding a pediatric indication (age 1 year and older for type 1 diabetes) in a 2019 supplemental approval 7.

Dosing flexibility language also differs. The EMA SmPC explicitly permits a minimum of 8 hours and maximum of 40 hours between doses for situations requiring schedule adjustment 1. The FDA label references the flexible dosing data from BEGIN Flex but does not prescribe a specific minimum-maximum dosing window in the same terms.

Post-Market Surveillance: Sentinel vs EudraVigilance

The FDA monitors Tresiba through two primary channels: the FDA Adverse Event Reporting System (FAERS) for spontaneous reports, and the Sentinel System, a distributed data network covering over 100 million U.S. patients through claims and electronic health record data 5. Sentinel can run active surveillance queries on insulin-related outcomes, including hypoglycemia-related emergency department visits and cardiovascular event rates in insulin-treated populations.

The EMA uses EudraVigilance, its centralized pharmacovigilance database, supplemented by mandatory PSURs submitted by marketing authorization holders. Novo Nordisk submits PSURs on a defined schedule, and the EMA's Pharmacovigilance Risk Assessment Committee (PRAC) reviews each report for new safety signals.

A key structural difference: the FDA's Sentinel allows population-level active surveillance without requiring manufacturer-initiated studies, while the EMA's system depends more heavily on manufacturer reporting obligations and referrals triggered by signal detection in EudraVigilance 5. Both systems flagged no new cardiovascular safety concerns for degludec through 2025, consistent with the DEVOTE findings 8.

What the Divergence Reveals About Regulatory Philosophy

The degludec case illustrates a recurring tension. The EMA tends to approve drugs with identified risks when the benefit-risk balance is positive, then manages uncertainty through enforceable post-authorization measures. The FDA, particularly after the 2008 CV guidance, tends to require resolution of major safety signals before granting initial approval, even when efficacy is established.

Neither approach is categorically superior. The EMA's earlier approval gave European patients access to an insulin with documented hypoglycemia advantages roughly 30 months before American patients. The FDA's delay forced generation of the DEVOTE data, which benefited prescribers worldwide by providing definitive cardiovascular safety evidence and the largest randomized comparison of severe hypoglycemia rates between two basal insulins ever conducted (7,637 patients, 1.99 years median follow-up) 4.

For clinicians evaluating degludec today, the regulatory divergence is historical. Both agencies now agree on cardiovascular safety. The practical question is pharmacological: degludec's 25-hour half-life produces lower glycemic variability and less nocturnal hypoglycemia than glargine U100, as demonstrated in DEVOTE (40% reduction in severe hypoglycemia, P<0.001) 4. Patients at high hypoglycemia risk, including those with hypoglycemia unawareness, long diabetes duration, or renal impairment, stand to benefit most from these properties 6.

Biosimilar and Patent Considerations

Novo Nordisk's core composition-of-matter patent for insulin degludec expired in the United States in December 2024. The FDA has received at least one biosimilar application (BLA) for degludec, and the EMA's similar biological medicinal product pathway may see filings in the same timeframe. Both agencies apply their respective biosimilar/interchangeability frameworks, with the FDA requiring additional switching study data for interchangeability designation under the BPCIA 9.

Biosimilar competition could meaningfully reduce degludec costs. The current U.S. wholesale acquisition cost for Tresiba U200 is approximately $530 per 3 mL pen (as of early 2026), and biosimilar entry typically reduces branded biologic prices by 20% to 40% within two years of launch based on patterns observed with insulin glargine biosimilars 9.