State of diabetes research — 2026
A dated evidence review of disease modification, cell replacement, immune protection, incretin therapies, monitoring, and medical AI.
By The Diabetes Guide editorial project · Updated 30 Sept 2026
On this page
A 2026 review of the newest research: immune medicines, lab-grown cells, gene editing, new Type 2 drugs, devices and AI.
Step 1: Slow the loss
Teplizumab can delay or slow Type 1 in selected people. It does not remove the need for insulin.
For example: Slowing a snowball, not melting it.
Step 2: Replace cells
In a small 2025 study, most people in the full-dose group needed no insulin for a year, with medicines to dampen the immune system.
For example: A successful first flight of a new plane; not a national airline yet.
Step 3: What is next
Longer follow-up, safety, cost and access. Nobody can honestly say “a cure in X years”.
For example: A bridge is only useful when it reaches the other side.
Remember: Publication, trial registration and approval are three different events.
The full story
Want more? Below is the detailed version with the real science words. It is fine to skip it.
Immune modulation: a changed US label
In April 2026, the US teplizumab label expanded Stage 2 eligibility to children aged one year and older. In June, the FDA granted accelerated approval for slowing decline of endogenous insulin production in selected recently diagnosed Stage 3 patients aged 8–17. The PROTECT trial included 328 participants; the accelerated indication relied on C-peptide preservation, a surrogate rather than proof of permanent disease elimination. Insulin treatment remains essential when required. 12
Cell replacement: a meaningful but limited human signal
The 2025 zimislecel study reported insulin independence at one year in 10 of 12 full-dose participants. This was a small, selected, immunosuppressed population. The finding supports function after cell replacement, not broad long-term safety or access. 3
Gene editing: protection is not yet a routine solution
A 2025 report described a single recipient of gene-edited donor islets with secretion and no detected immune response at 12 weeks without immunosuppression. It is a proof of concept with very short follow-up; it must not be presented as an established stem-cell cure or generalized insulin independence. 4
Encapsulation and regeneration
Encapsulation seeks a barrier that admits oxygen, nutrients and insulin while limiting immune injury. Device fibrosis and nutrient diffusion create difficult tradeoffs. A registered trial such as VX-264 establishes that an approach is studied, not that it worked. We do not infer efficacy from recruitment or assign an approval date. Beta-cell regeneration remains a separate experimental strategy requiring functional, safely controlled growth. 9
Incretins and amylin-based combinations
Tirzepatide's 176-week prevention analysis concerned 1,032 people with obesity and prediabetes. It reduced diagnoses while treatment was ongoing; follow-up after withdrawal was much shorter. REDEFINE 2 tested cagrilintide–semaglutide in 1,206 adults with Type 2 and overweight or obesity for 68 weeks. Weight and glucose benefits do not establish medication-independent remission or a permanent cure. This review reports trial findings, not an unverified approval claim. 65
Devices and AI
Automated insulin delivery is established clinical technology for appropriate users, with 2026 guidance reflecting expanding applications. It addresses glucose management rather than the immune disease. Prediction models using health records, wearables or molecular data still need task-specific external validation and prospective clinical evaluation. 78
What has to happen next?
| Research direction | Next evidence needed | Routine-use forecast |
|---|---|---|
| Immune modulation | Durable clinical outcomes, safety, broader applicability | Existing indications are specific; no blanket prevention claim |
| Manufactured islets | Long-term benefit, safety, immune protection, manufacturing consistency | No universal timetable |
| Edited cells | Replication, adequate therapeutic dose, long-term surveillance | Unknown |
| Encapsulation | Reliable survival, oxygenation and protection | Unknown |
| Regeneration | Safe functional growth in humans | Unknown |
| Incretin/amylin combinations | Long-term outcomes, safety, access and maintenance | Depends on product and jurisdiction |
| AI prediction | External calibration, impact trials, safe implementation | Depends on the specific clinical task |
The next breakthrough may be a combination of partial solutions. A forecast of “X years to cure” would not be scientifically justified.
Trace the evidence
Sources and further reading
1.Tzield prescribing information, revised June 2026 (opens in a new tab)
US Food and Drug Administration · 2026 · Regulatory
Who was studied, limits and source check
Limitations: US label. Age and staging criteria differ between indications; specialist evaluation and infection precautions are required.
Source checked 2026-09-30. See the original publication for full methods.
2.FDA accelerated approval: teplizumab in selected recently diagnosed pediatric Stage 3 T1D (opens in a new tab)
US Food and Drug Administration · 2026 · Regulatory
Who was studied, limits and source check
Population: Children aged 8–17 recently diagnosed with Stage 3 T1D
Sample: PROTECT: 328 participants
Limitations: June 12 accelerated approval based on C-peptide preservation, a surrogate. Confirmatory clinical benefit and label conditions matter.
Source checked 2026-09-30. See the original publication for full methods.
3.Stem Cell–Derived, Fully Differentiated Islets for Type 1 Diabetes (opens in a new tab)
Reichman et al.; New England Journal of Medicine · 2025 · Early human study
Who was studied, limits and source check
Population: Selected adults with T1D, severe hypoglycemia and impaired awareness
Sample: 14 with ≥12-month follow-up; 12 at full dose
Limitations: Small, uncontrolled interim analysis; required immunosuppression; serious adverse events and deaths occurred. Not a routine cure.
Source checked 2026-09-30. See the original publication for full methods.
4.Survival of Transplanted Allogeneic Beta Cells with No Immunosuppression (opens in a new tab)
Carlsson et al.; New England Journal of Medicine · 2025 · Early human study
Who was studied, limits and source check
Population: One adult with long-standing T1D
Sample: 1 participant; 12-week reported follow-up
Limitations: Single participant, low cell dose and short follow-up. Does not demonstrate generalizable insulin independence.
Source checked 2026-09-30. See the original publication for full methods.
5.Cagrilintide–Semaglutide in Adults with Overweight or Obesity and T2D (opens in a new tab)
REDEFINE 2 Study Group; New England Journal of Medicine · 2025 · Randomized trial
Who was studied, limits and source check
Population: Adults with BMI ≥27, T2D and HbA1c 7–10%
Sample: 1,206 randomized participants
Limitations: 68-week trial; weight and glycemia outcomes are not evidence of a permanent cure. Publication does not establish regulatory approval.
Source checked 2026-09-30. See the original publication for full methods.
6.Tirzepatide for Obesity Treatment and Diabetes Prevention (opens in a new tab)
Jastreboff et al.; New England Journal of Medicine · 2025 · Randomized trial
Who was studied, limits and source check
Population: Adults with obesity and prediabetes
Sample: 1,032 in the prediabetes analysis
Limitations: 176-week treatment and limited off-treatment follow-up; manufacturer-funded; prevention during therapy is not a permanent cure.
Source checked 2026-09-30. See the original publication for full methods.
7.Diabetes Technology: Standards of Care 2026 (opens in a new tab)
American Diabetes Association · 2026 · Guideline
Who was studied, limits and source check
Limitations: Access, training, device labeling, and safe-use capacity affect applicability.
Source checked 2026-09-30. See the original publication for full methods.
8.Machine learning-based T2D risk prediction in primary care: a scoping review (opens in a new tab)
PubMed-indexed primary-care evidence review · 2026 · Review
Who was studied, limits and source check
Limitations: Evidence through December 2025; few studies, limited prospective deployment, external validation and calibration.
Source checked 2026-09-30. See the original publication for full methods.
9.VX-264 Phase 1/2 study registration (opens in a new tab)
ClinicalTrials.gov; Vertex Pharmaceuticals · 2026 · Early human study
Who was studied, limits and source check
Limitations: Registry accessed in 2026. Enrollment and registration are not proof of efficacy; consult the live record for recruitment and results.
Source checked 2026-09-30. See the original publication for full methods.
Source checking is an editorial literature check, not independent medical review. This page is for learning. It cannot diagnose you or make a treatment plan. Evidence labels describe the cited claims, not the whole topic.
What does “Preliminary” mean?
Early results. They need to be repeated. Like a first test drive of a new car.
Keep reading
The body
The one-page mental model
Diabetes means there is too much sugar in the blood for too long, because the body does not have enough insulin or cannot use it well.
Type 1
Type 1 diabetes: the complete journey
Type 1 diabetes is usually an autoimmune condition. The body’s defence system harms the beta cells, so less and less insulin is made.
Type 2
Type 2 diabetes: how the system changes
Type 2 diabetes happens when the body’s need for insulin grows and the pancreas cannot keep up, so sugar slowly rises.