Back to Research & Science
Research & Science•••11 min read

Stopping GLP-1 Therapy in Type 2 Diabetes: What a Large Italian Cohort Study Reported

An observational study of 80,302 people with type 2 diabetes linked GLP-1 RA discontinuation with higher rates of heart attack, stroke, and death. Educational summary, limits, and tracker tips.

PeptIQ Team
Peptide Research & Education
Stopping GLP-1 Therapy in Type 2 Diabetes: What a Large Italian Cohort Study Reported

PeptIQ app

Track your peptide protocols in PeptIQ

Log injections, vial supply, and reminders. Use peptiq.io calculators when you need math on the web.

Get the appOpen web app

iOS, Android, and web

Research supplier

Semaglutide listings are for laboratory / research use only — not for human consumption.

Also mentioned

Research use only. Partner link. PeptIQ may earn a referral fee. Not medical advice.

Ask PeptIQ

Two taps on our guided quiz, then keep going in iMessage with sources — not Instagram screenshots.

Start the quizOr try Peptide X Quiz →

Under 25? Talk to a licensed clinician and get baseline labs before starting anything. Why we say this · Labs checklist

> Note: PeptIQ is an education and tracking app. It does not prescribe, sell, or recommend medicines. This article summarizes published research. It is not medical advice.

TL;DR

In November 2026, Ceriello and colleagues published a retrospective cohort study in The Lancet Regional Health – Europe. Among 80,302 adults with type 2 diabetes (T2D) and no prior cardiovascular disease who started a GLP-1 receptor agonist (GLP-1 RA), 60.5% had discontinued by their last clinic visit. After propensity-score matching, people with time off therapy had about twice the incidence of a composite of non-fatal myocardial infarction, non-fatal stroke, or all-cause death, and a larger gap for all-cause mortality alone.

That pattern is association, not proof that stopping the drug caused those events. The study population is T2D treated in Italian diabetes clinics, not everyone using GLP-1 medicines for weight management alone. Talk with a licensed clinician before changing or stopping treatment.

Why persistence matters in the real world

Randomized trials established that GLP-1 RAs can improve cardiometabolic risk factors and, in some populations, hard cardiovascular outcomes when treatment continues. In routine care, many people stop within the first year because of cost, gastrointestinal side effects, access, or other reasons cited in the paper and in prior observational work.

Ceriello’s team asked a practical question: in people who already started a GLP-1 RA for T2D, does stopping line up with more heart attacks, strokes, or deaths than staying on therapy? They used the nationwide AMD (Associazione Medici Diabetologi) quality registry: 301 diabetes clinics, data from 2004 through 2024.

What the study did

Inclusion: Adults with T2D actively followed in clinic who received at least one GLP-1 RA prescription and had no major cardiovascular event at first prescription.

Groups: “Persistent users” still had GLP-1 RA prescribed at the last visit. The discontinuation group had no GLP-1 RA prescription at two consecutive visits (with sensitivity analyses using a stricter one-visit rule).

Endpoints:

  • Primary: Composite of non-fatal MI, non-fatal stroke, or death from any cause.
  • Secondary: Death from any cause.

Methods worth noting: To reduce immortal time bias, person-time on drug counted toward the persistent-exposure window even for people who later stopped; time after stopping counted toward discontinuation exposure. The main analysis used propensity score matching (age, BMI, HbA1c, diabetes duration, sex, sulfonylurea use) to balance baseline traits.

Key numbers from the paper

Among 80,302 initiators without baseline CVD:

FindingPersistent usersDiscontinuation group
Still on GLP-1 RA at last visit31,687 (39.5%)48,615 (60.5%) had stopped
Mean GLP-1 RA exposure2.97 ± 2.93 years1.21 ± 2.13 years
Mean follow-up3.65 ± 3.03 years4.25 ± 3.41 years

After 1:1 matching (28,713 per group):

  • Primary composite: incidence 1.96 vs 3.94 events per 100 person-years (persistent vs discontinuation). IRR 2.01 (95% CI 1.77–2.28); HR 2.03 (95% CI 1.92–2.14).
  • All-cause death: incidence 0.48 vs 2.23 per 100 person-years. IRR 4.67 (95% CI 4.11–5.30); HR 4.81 (95% CI 4.39–5.26).

Obesity subgroup: Among people with BMI ≥30 kg/m², the HR for the primary endpoint was 2.27 (95% CI 2.12–2.43) vs 1.61 (95% CI 1.47–1.78) when BMI was below 30.

Authors emphasized that changes in HbA1c, weight, blood pressure, and lipids after discontinuation did not clearly separate those who later had an event from those who did not, suggesting the signal may not be explained by classic risk-factor drift alone—while still warning about residual confounding and reverse causation (for example, stopping because of new illness).

How to read this without over-interpreting

  • Observational design. People who stop treatment are not randomly assigned to stop. They may differ in ways the registry cannot measure (frailty, cancer, socioeconomic stress, intolerance).
  • T2D clinic population. Most participants used GLP-1 RAs for diabetes; semaglutide was less common than in recent obesity-only cohorts, and tirzepatide was not available for most of the observation window. Findings may not transfer to Wegovy-only users without diabetes.
  • No causation claim. The authors state discontinuation may mark higher underlying risk rather than directly causing events. The especially large mortality association may partly reflect unmeasured factors.
  • Not a tapering guide. The study compares continued use with time off therapy. It does not test gradual dose reduction or structured maintenance plans.

For weight-focused withdrawal evidence in non-diabetic obesity trials, see our separate guide on weight regain after stopping GLP-1 medicines.

How this fits next to trial evidence

TopicStudy typePopulationTakeaway for readers
SUSTAIN 6Randomized CV outcomes trialT2D at high CV riskSemaglutide reduced major adverse cardiovascular events while on treatment in trial conditions.
LEADERRandomized CV outcomes trialT2DLiraglutide reduced CV events vs placebo over trial follow-up.
STEP 4Randomized withdrawalOverweight/obesity, no diabetesContinued semaglutide vs switch to placebo with lifestyle support: different average weight trajectories.
STEP 1 extensionFollow-up after trial endSame STEP 1 subsetAverage weight regain after stopping drug and structured lifestyle.
SURMOUNT-4Randomized withdrawalObesity without diabetesContinued tirzepatide vs placebo after initial treatment period.
Ceriello 2026Retrospective cohortT2D, no baseline CVDHigher incidence of MI/stroke/death associated with GLP-1 RA discontinuation in matched analysis.

Wiki pages for semaglutide, tirzepatide, and the GLP-1 discontinuation study entry link these sources for deeper reading.

Practical, non-prescriptive takeaways for trackers

PeptIQ does not tell you to start, stop, or dose any medicine. If you and your clinician are discussing persistence, cost, or side effects, structured notes can make that conversation clearer:

  • Log each injection or dose with date, product name as prescribed, and site if relevant.
  • Track weight, glucose, blood pressure, and symptoms your care team already monitors—not to self-diagnose, but to show trends over time.
  • Record why a dose was missed or treatment paused (access, nausea, travel, clinician instruction) in your own words.
  • Bring questions explicitly: “What are we trying to maintain if therapy changes?” and “How will we follow up if I stop or switch?”

Our calculator hub includes dosing math tools for education; clinical decisions stay with licensed prescribers. For access and safety framing, see when to talk to a clinician about GLP-1 access.

How PeptIQ.io helps

The PeptIQ.io download page offers injection logging, reminders, and export-friendly history. Use it to keep an accurate timeline—not to replace medical follow-up. PeptIQ AI and articles cannot determine whether stopping a GLP-1 RA is appropriate for you.

Frequently asked questions

Q: Does this study prove stopping Ozempic or Wegovy causes heart attacks?

A: No. It reports associations in people with type 2 diabetes in Italian clinics. Causation was not established, and the paper lists confounding and reverse causation as major limits.

Q: I use semaglutide only for weight loss without diabetes. Does this apply to me?

A: Not directly. The cohort was defined by T2D care. Weight-loss-only populations were not the focus, and tirzepatide was largely unavailable during the study period.

Q: Should I never stop a GLP-1 medicine?

A: Only a licensed clinician can answer that for your situation. This article does not recommend staying on or stopping treatment.

Q: How does this relate to weight regain studies?

A: Withdrawal trials (STEP 4, SURMOUNT-4, STEP 1 extension) mostly report weight outcomes in specific trial populations. Ceriello 2026 reports cardiovascular and mortality endpoints in real-world T2D practice. They address related but distinct questions.

Sources

  • Ceriello A, Prattichizzo F, Lucisano G, et al. GLP-1 receptor agonist discontinuation and cardiovascular events and all-cause mortality in people with type 2 diabetes. Lancet Reg Health Eur. 2026. DOI: 10.1016/j.lanepe.2026.101881
  • Marso SP, Bain SC, Consoli A, et al. Semaglutide and Cardiovascular Outcomes in Patients with Type 2 Diabetes (SUSTAIN 6). N Engl J Med. 2016. PubMed
  • Marso SP, et al. Liraglutide and Cardiovascular Outcomes in Type 2 Diabetes (LEADER). N Engl J Med. 2016. PubMed
  • Wadden TA, et al. STEP 4. JAMA. 2021. PubMed
  • Wilding JPH, et al. STEP 1 extension withdrawal analysis. Diabetes Obes Metab. 2022. PubMed
  • Aronne LJ, et al. SURMOUNT-4. JAMA. 2024. PubMed

Related guides

This article is educational content from PeptIQ.io. It is not medical advice. Talk with a licensed clinician before starting, stopping, or changing any peptide, medication, or injection practice.

For clinics: running peptide or GLP-1 programs for patients? See PeptIQ.io for providers: protocol templates, a patient app, and adherence tracking in one console.

Share this article

Track Your Peptide Protocols

Use PeptIQ to log injections and vial supply, use peptiq.io calculators for math on the web, and keep protocols organized.

Stay ahead of the peptide conversation

Weekly-ish notes on new research, wiki trends, and practical tracking tips. No spam — unsubscribe anytime.

Educational updates only — not medical advice. Privacy