Liraglutide Did Not Preserve Residual C-Peptide in Long-Standing Type 1 Diabetes: What a 2026 Pilot RCT Shows

A randomized, blinded, placebo-controlled pilot trial published online on August 13, 2026 tested whether liraglutide could preserve residual beta-cell function in adults with long-standing type 1 diabetes and detectable fasting C-peptide.[1] The result was negative on the prespecified primary endpoint: liraglutide did not improve stimulated C-peptide area under the curve compared with placebo after 52 weeks.[1]

The paper, published in Diabetes, Obesity and Metabolism (DOI: 10.1111/dom.71204; PMID: 42595732), adds a small but important piece of human evidence.[1] It directly tests a beta-cell preservation hypothesis in a population that differs from the newly diagnosed participants studied in earlier liraglutide research.[1][2]

It does not establish that liraglutide damages beta cells, settle every possible role for GLP-1 receptor agonism in type 1 diabetes, or change liraglutide’s regulatory status.[1][4] The study enrolled only 18 adults, and its more concerning longitudinal signal came from exploratory post hoc analysis rather than the prespecified primary comparison.[1]

What changed on August 13, 2026

The new development is the publication of a human pilot randomized controlled trial focused on residual endogenous insulin secretion, not weight loss or routine glucose lowering.[1] The investigators assigned 18 adults to liraglutide or placebo for 52 weeks.[1] Eligible participants had lived with type 1 diabetes for at least five years and still had detectable fasting C-peptide.[1]

Mixed-meal tolerance tests were performed at baseline and at weeks 6, 52, and 65.[1] The primary endpoint was the change from baseline to week 52 in stimulated C-peptide area under the curve.[1]

The main findings were:

• liraglutide did not improve stimulated C-peptide versus placebo at week 52[1]
• exploratory longitudinal analysis across weeks 0 through 65 suggested a steeper stimulated C-peptide decline in the liraglutide group[1]
• that exploratory pattern was accompanied by increasing insulin requirements over time[1]

The first finding is the trial’s primary result.[1] The second and third are signals generated by post hoc analysis.[1] They should not be presented as equivalent levels of evidence.[1]

Why residual C-peptide is the central endpoint

C-peptide is used in this research context as a marker of endogenous insulin secretion.[1] A mixed-meal challenge asks whether surviving beta cells can increase secretion in response to a standardized metabolic stimulus.[1] Integrating that response over time produces the stimulated C-peptide area under the curve used as the trial’s primary endpoint.[1]

This endpoint matters because detectable beta-cell function can persist even after many years of type 1 diabetes.[3] Observational human evidence has associated residual beta-cell function in long-term type 1 diabetes with a lower incidence of hypoglycemia, although an association does not by itself establish that experimentally increasing C-peptide would cause the same outcome.[3]

The 2026 trial therefore asked a narrower and more mechanistic question than whether participants lost weight or needed less insulin during treatment.[1] It asked whether a year of liraglutide changed the measurable capacity of the remaining beta cells relative to placebo.[1]

That distinction is valuable.[1] A therapy can change insulin requirements during treatment without demonstrating durable beta-cell preservation.[2] Conversely, a modest beta-cell signal should not automatically be treated as a clinically meaningful outcome.[3]

Trial design: stronger than an anecdote, smaller than a definitive RCT

The design has several strengths:

1. Randomized comparison. Liraglutide was compared with placebo rather than only with each participant’s pretreatment value.[1]
2. Blinding. The trial was blinded, which reduces some expectation and measurement biases.[1]
3. Defined population. Participants had type 1 diabetes for at least five years and detectable fasting C-peptide.[1]
4. Repeated challenge testing. Mixed-meal tests were conducted before treatment, during treatment, at the primary 52-week point, and after treatment at week 65.[1]
5. Mechanism-aligned endpoint. Stimulated C-peptide directly addressed the residual-function hypothesis.[1]

The central weakness is scale.[1] Eighteen randomized participants provide limited power to detect modest effects, evaluate subgroups, or characterize uncommon adverse outcomes.[1] Small samples are also more vulnerable to baseline imbalance and to a few influential observations.[1]

The abstract does not provide enough participant-level information to determine whether baseline C-peptide, disease duration, insulin delivery technology, glycemic control, or other features were evenly distributed.[1] Those variables should be examined in a larger confirmatory analysis because this pilot randomized only 18 adults.[1]

The trial should therefore be read as a negative pilot RCT with an exploratory warning signal, not as definitive evidence of benefit or harm.[1]

Primary result versus exploratory signal

The prespecified comparison is straightforward: stimulated C-peptide was not improved by liraglutide versus placebo at week 52.[1] This does not prove that the two strategies are identical under every condition.[1] It means the study did not demonstrate the preservation effect it was designed to detect in this population.[1]

The exploratory analysis deserves careful wording because it was post hoc.[1] When the investigators modeled measurements across the full 65-week period, including post-treatment follow-up, stimulated C-peptide declined more steeply in the liraglutide group and insulin requirements increased over time.[1]

Several cautions follow:

• Post hoc status limits the inference. The full-period trajectory was exploratory rather than the prespecified primary comparison.[1]
• The sample was very small. Only 18 adults were randomized.[1]
• The abstract does not establish a mechanism. It reports a C-peptide trajectory without identifying beta-cell death or another causal explanation.[1]
• Replication is essential. A larger study should prespecify the longitudinal model and the post-treatment assessment raised by this pilot.[1]

The defensible conclusion is not “liraglutide accelerates beta-cell failure.” The defensible conclusion is that the primary preservation hypothesis was not supported and that the exploratory trajectory should be tested prospectively.[1]

Why the NewLira comparison matters

The negative long-standing-disease result contrasts with the 2024 NewLira trial in adults with newly diagnosed type 1 diabetes.[2] NewLira randomized 68 participants to liraglutide or placebo for 52 weeks.[2] During treatment, stimulated C-peptide area under the curve was maintained with liraglutide and declined with placebo.[2]

That earlier finding also had an important boundary. Six weeks after liraglutide stopped, C-peptide area under the curve was similar between the groups.[2] The result suggested an on-treatment preservation or secretory effect during the first year after diagnosis, not clearly durable disease modification after withdrawal.[2]

The contrast makes disease stage an important research hypothesis.[1] The two trials show different on-treatment C-peptide results in newly diagnosed and long-standing populations.[2] They do not establish which biological feature caused the difference.[1]

That explanation is plausible, not proven.[1] The two studies were conducted by different teams in different populations and were not a single stratified head-to-head trial.[2] Differences in baseline C-peptide, participant characteristics, analysis, and chance could also contribute.[1]

Mechanism evidence is not disease-modifying evidence

Liraglutide is a GLP-1 receptor agonist.[2] The site’s incretin-axis overview explains how GLP-1 receptor signaling can affect insulin and glucagon biology. The two randomized studies show why an on-treatment metabolic effect must be distinguished from durable beta-cell preservation.[1][2]

Three evidence layers should remain separate:

Acute pharmacology

A GLP-1 receptor agonist can change insulin secretion and insulin requirements while the drug is present.[2] Those changes do not by themselves prove that beta-cell survival changed.[1][2]

Beta-cell functional endpoint

Stimulated C-peptide tests endogenous secretory capacity under defined assay conditions.[1] A during-treatment difference does not by itself identify whether cell survival, cell function, or both changed.[1] NewLira’s loss of the between-group difference after withdrawal illustrates why post-treatment measurement matters.[2]

Clinical outcomes

Severe hypoglycemia and insulin requirements are clinically distinct from a stimulated C-peptide biomarker.[3] A biomarker can be informative without guaranteeing improvement in those outcomes.[3]

The 2026 paper is strongest at the second layer: it reports no improvement on the primary beta-cell functional endpoint.[1] It is not an efficacy trial establishing clinical benefit, and its exploratory trend is not proof of clinical harm.[1]

Regulatory status did not change

The trial is investigational evidence in type 1 diabetes.[1] It is not an authorization or label expansion.[4]

The FDA prescribing information for Victoza lists type 2 diabetes indications and states that the product should not be used in patients with type 1 diabetes mellitus.[4] The 18-person pilot does not alter that language or create an approved type 1 diabetes indication.[1][4]

This distinction is especially important because the study question involved residual beta-cell function in a selected research population.[1] It should not be converted into treatment, dosing, or self-experimentation guidance.[1][4]

What the study does and does not show

Supported by the evidence

• The study was a blinded randomized comparison in 18 adults with long-standing type 1 diabetes and detectable fasting C-peptide.[1]
• Liraglutide did not improve stimulated C-peptide area under the curve versus placebo at 52 weeks.[1]
• A post hoc longitudinal analysis suggested a steeper decline through week 65 and increasing insulin requirements.[1]
• Earlier NewLira evidence in newly diagnosed adults produced a different on-treatment C-peptide result.[2]
• Liraglutide is not approved by the FDA for type 1 diabetes.[4]

Not established

• that liraglutide causes beta-cell destruction
• that every person with residual C-peptide would respond similarly
• that the finding applies to semaglutide, tirzepatide, or another incretin-based molecule
• that a C-peptide change necessarily predicts a clinical outcome
• that the difference from NewLira is caused only by disease duration
• that the pilot changes clinical practice or regulatory status

What researchers should watch next

A useful confirmatory study would be larger, prospectively registered, and explicitly powered for stimulated C-peptide change. It should include:

Prespecified disease-duration strata

New-onset and long-standing type 1 diabetes should not be pooled without a planned interaction analysis. Baseline stimulated and fasting C-peptide should be reported in enough detail to assess residual-function heterogeneity.

A locked longitudinal model

The statistical plan should prespecify the week 52 primary comparison and the slope through a post-treatment period. That would turn the 2026 post hoc observation into a testable hypothesis rather than a retrospective signal.[1]

Multiple endpoint layers

Stimulated C-peptide should be paired with insulin requirements, continuous glucose monitoring outcomes, severe hypoglycemia, glycated hemoglobin, body composition, adverse events, and patient-reported outcomes. Mechanistic biomarkers and clinical outcomes should be interpreted separately.

Post-withdrawal follow-up

NewLira’s between-group C-peptide difference was no longer present six weeks after treatment stopped.[2] Longer follow-up can help distinguish an on-drug secretory effect from durable preservation.

Transparent safety reporting

The new abstract emphasizes beta-cell outcomes and does not provide a complete safety dataset. A larger study should prespecify gastrointestinal, glycemic, pancreatic, gallbladder, cardiovascular, and treatment-discontinuation reporting rather than relying on a small trial to exclude uncommon events.

Molecule-specific conclusions

The result should not be generalized automatically across the incretin class. Liraglutide, semaglutide, tirzepatide, and retatrutide differ in receptor pharmacology, exposure, and evidence base.

Bottom line

The August 13, 2026 publication provides a credible negative result from a small human randomized trial.[1] In 18 adults with type 1 diabetes for at least five years and detectable fasting C-peptide, liraglutide did not preserve stimulated C-peptide versus placebo at the 52-week primary endpoint.[1]

The exploratory longitudinal analysis raises a question about a potentially steeper decline, but it does not prove beta-cell harm.[1] The contrast with NewLira suggests that disease stage may matter, while also showing why indirect comparisons must remain cautious.[1][2]

For researchers, the durable lesson is methodological: define the disease stage, separate on-treatment metabolic effects from post-treatment beta-cell preservation, and distinguish a prespecified endpoint from a post hoc trajectory. The evidence supports a larger confirmatory study, not a treatment recommendation.

This guide is for scientific education only. It does not provide medical advice, treatment, dosing, or sourcing guidance.

Sources

[1] https://doi.org/10.1111/dom.71204 — Liraglutide and Beta-Cell Function in Young Adults With Long-Standing Type 1 Diabetes and Residual C-Peptide [2] https://doi.org/10.1111/dom.15889 — NewLira: Liraglutide in Newly Diagnosed Type 1 Diabetes [3] https://doi.org/10.1172/JCI143011 — Residual beta cell function in long-term type 1 diabetes associates with reduced hypoglycemia [4] https://www.accessdata.fda.gov/drugsatfda_docs/label/2023/022341s039lbl.pdf — Victoza (liraglutide) FDA Prescribing Information

Related Research

• Incretin Axis - receptor-level context for GLP-1, GIP, insulin, and glucagon signaling
• Endocrine System - framework for interpreting peptide hormones, feedback, and downstream biomarkers
• Metabolic Regulation and Peptide Signaling - broader context for nutrient sensing and peptide-mediated metabolic control
• Peptide Pharmacokinetics and Bioavailability - why exposure and duration of action are different from durable biological change
• Semaglutide Research Guide - molecule-specific GLP-1 receptor agonist evidence and limits
• Tirzepatide Research Guide - dual GIP and GLP-1 receptor pharmacology without assuming class equivalence
• Retatrutide Research Guide - investigational triple-agonist context and evidence boundaries