GLP-1 drugs to rescue beta cells at diagnosis
Investigator-led (academic); drugs marketed by Novo Nordisk / Eli Lilly for other uses
GLP-1 receptor drugs — the semaglutide/tirzepatide family now famous for type 2 diabetes and weight loss — also protect and amplify beta cells in lab models. A striking but tiny, uncontrolled case series gave semaglutide to people within months of a T1D diagnosis and saw most come off mealtime insulin, with rising C-peptide. It is an intriguing, early signal that these drugs might preserve the body's own insulin production at diagnosis — not proof, and not without real ketoacidosis risk.
The scorecard
In a 10-person uncontrolled series begun within months of diagnosis, all stopped mealtime insulin and 7/10 stopped basal insulin, with C-peptide rising and HbA1c ~5.7% at one year — striking but unblinded and uncontrolled.[1]
Benefits held through 12 months of continued dosing in the case series, but there is no controlled or longer-term evidence that beta-cell function is durably preserved.[1]
Familiar GI side effects, but cutting insulin in T1D risks diabetic ketoacidosis; the small series reported none, yet this is the central safety question for the approach.[2]
Plausibly relevant only at or near diagnosis, when residual beta cells remain; the drugs are widely available but unapproved and unproven for this purpose.[1]
Only a single small, uncontrolled new-onset case series supports the beta-cell story; adjunct use in established T1D is better described but does not address preservation.[1]
The full picture
GLP-1 receptor drugs are best known for type 2 diabetes and weight loss, but in the laboratory they do something relevant to a cure for T1D: they help beta cells survive, function and even proliferate, and they blunt the stress that kills them. That raised an obvious question — if you start a GLP-1 drug at the moment of a T1D diagnosis, while some beta cells are still alive, can you protect the body's own insulin production?
The most eye-catching human hint comes from a small report in which ten people, all diagnosed within the previous few months, were started on semaglutide alongside insulin. As the dose rose, mealtime insulin was withdrawn in every patient, and background insulin in seven of ten, while their own C-peptide rose and HbA1c fell to near-normal levels over a year. It reads like a regeneration story — but it is uncontrolled, unblinded, and tiny, exactly the kind of result that can shrink or vanish in a proper randomized trial. There is also a real hazard: pulling insulin in T1D invites diabetic ketoacidosis. Broader off-label experience in established T1D shows glucose and weight benefits but does not speak to beta-cell preservation. This is a genuine, watch-this-space hypothesis, not a therapy you can rely on today.
As of our July 2026 review, that is still the entire human beta-cell-preservation case: those ten people. Nothing at the June 2026 ADA scientific sessions, and nothing published since, added a randomized C-peptide readout for a GLP-1 drug started at diagnosis. (GLP-1 drugs are being studied in T1D for glucose and weight — tirzepatide in particular — but those trials are not asking whether the body's own insulin production is saved.) So the honest position has not moved in a year: an interesting signal, still unreplicated, still uncontrolled.
The first registered attempt to test it properly has now appeared, and it is worth watching precisely because it is the thing that has been missing. SHIELD-T1D (NCT07614412) is a 240-person, four-arm, placebo-controlled Phase 2 sponsored by Saudi Arabia's Ministry of Health: adults aged 18–50, randomized within 100 days of their first insulin injection, to semaglutide alone, the recombinant zoster vaccine (Shingrix) alone, both together, or dual placebo. The primary endpoint is exactly the right one — change in stimulated C-peptide on a mixed-meal tolerance test at 12 months — with follow-up out to 24 months. Two large caveats: it has not opened yet (planned start January 2027, primary completion January 2028), and it is currently listed at a single site in Riyadh. A planned trial is not a result, and this one will not tell us anything before 2028.
Coming soon
ETA · Hypothesis-generating: one small uncontrolled case series in new-onset T1D; controlled trials needed before any beta-cell-preservation claim. No approval for this use. The first registered randomised beta-cell-preservation trial (SHIELD-T1D, NCT07614412) is planned but has not opened; it would not read out before 2028.
- →Randomized controlled trials of GLP-1 drugs started at T1D diagnosis with C-peptide as the endpoint
- →Careful evaluation of diabetic-ketoacidosis risk when insulin is reduced
- →SHIELD-T1D (NCT07614412) — a 240-person, four-arm Phase 2 at a single Saudi site testing semaglutide, the recombinant zoster vaccine, or both, against placebo, with 12-month stimulated C-peptide as the primary endpoint · Not yet recruiting; start planned January 2027, primary readout 2028
Sources
- [1]Semaglutide in Early Type 1 Diabetes (Dandona et al., correspondence) · peer-reviewed · 2023-01-01
- [2]Expectations and Outcomes From GLP-1 Receptor Agonists As Adjunct Treatment for Type 1 Diabetes — Case Presentations · peer-reviewed · 2024-01-01
- [3]SHIELD-T1D (NCT07614412): Recombinant Zoster Vaccine and GLP-1 Receptor Agonist for the Preservation of Beta-Cell Function in Adults With Recent-Onset Type 1 Diabetes · registry — Phase 2, four-arm, 240 participants, sponsored by the Ministry of Health, Saudi Arabia. Not yet recruiting as of July 2026; planned start January 2027, primary completion January 2028.