GLP-1 Clinical Trials Are Reshaping Half of Medicine: What to Watch in 2026

Why This Stopped Being a Diabetes Story

When exenatide cleared FDA review in 2005, the argument for GLP-1 agonism was narrow. You got modest glycemic improvement, modest weight loss, and a curious appetite-suppressing effect that nobody was quite sure how to monetize. For almost a decade, the drugs sat inside diabetes clinics. Endocrinologists used them. Cardiologists ignored them. Nephrologists had never heard of them.

The pivot happened in stages. Liraglutide's LEADER trial in 2016 showed a cardiovascular benefit in patients with type 2 diabetes. Semaglutide's SUSTAIN-6 pointed in the same direction the same year. These were interesting signals, but they were restricted to a diabetic population and were not large enough to change how people outside endocrinology thought about the class. The real shift arrived in late 2023, when the SELECT trial demonstrated a 20 percent reduction in major adverse cardiovascular events with semaglutide 2.4 mg in patients with established cardiovascular disease and obesity but without diabetes. That was the moment the clinical world started paying attention as a single system.

Since then the trial machinery has run almost continuously. Novo Nordisk and Eli Lilly, with more limited entries from Amgen and AstraZeneca, have moved a portfolio of incretin-based drugs into what is now probably the most expensive and ambitious clinical program in modern pharmaceutical history. The scale matters. When one class of drug has concurrent Phase 3 readouts in cardiovascular, renal, hepatic, and neurodegenerative disease, a lot of adjacent research questions stop being optional.

The Cardiovascular Readouts: SELECT, SURPASS-CVOT, and SOUL

SELECT, published in NEJM in November 2023, enrolled 17,604 patients with established cardiovascular disease, overweight or obesity, and no diabetes. Subcutaneous semaglutide 2.4 mg weekly reduced MACE by 20 percent over a median follow-up of roughly 40 months. The absolute risk reduction was small, because baseline event rates were moderate, but the relative effect sat in the same neighborhood as PCSK9 inhibitors. For a weight-loss drug, this was an extraordinary result.

SURPASS-CVOT, testing tirzepatide head-to-head against dulaglutide in roughly 13,300 patients with type 2 diabetes and atherosclerotic cardiovascular disease, is the larger question that the community has been waiting on. A superiority result would begin to differentiate the dual GIP/GLP-1 agonist from the pure GLP-1 class. A noninferiority result still matters, because it would let cardiologists treat tirzepatide's label as a cardiovascular-safe drug and remove a barrier to prescribing for its metabolic effects.

SOUL, the first large cardiovascular outcomes trial of oral semaglutide in patients with type 2 diabetes and established disease, reported in 2025 a significant reduction in MACE. That result quietly did something important. It decoupled the cardiovascular benefit from the injection. Any serious argument that the effect of GLP-1 agonism was a secondary consequence of weight loss from a self-administered injection must now contend with the fact that oral dosing produces the same signal.

The research question that has not been answered well by any of these trials is the mechanism. Is the cardiovascular benefit mediated by weight loss, by direct vascular anti-inflammatory action, by improvements in blood pressure and lipid handling, by reduced alcohol and food reward, or by something that we still do not have a clean readout for? The trial program is not designed to resolve this. If you are a translational researcher and you are choosing a mechanistic question that matters, this is one of them.

FLOW and the Kidney Disease Pivot

FLOW was stopped early for efficacy in late 2023 and published in NEJM in May 2024. It tested semaglutide 1.0 mg in patients with type 2 diabetes and chronic kidney disease. The composite primary endpoint, which included kidney failure, a sustained 50 percent or greater decline in eGFR, cardiovascular death, and kidney death, was reduced by 24 percent. Kidney-specific endpoints moved in the same direction.

This matters for the field in a way that is easy to underappreciate. For most of the last decade, SGLT2 inhibitors have owned the narrative on slowing diabetic kidney disease progression. FLOW does not displace that. What it does is open a serious question about combination therapy, because the mechanism of action is different enough that additive benefit is plausible. Trials comparing and combining the two classes in patients with CKD, with and without diabetes, are underway, and the readouts over the next two to three years will shape how nephrology practices prescribe.

The other thing FLOW did was make the case for extending GLP-1 trials into non-diabetic CKD. Several early-stage programs are now exploring exactly that question. If a non-diabetic CKD trial shows benefit, the implication for population-level kidney care, and for the basic science of how GLP-1 signaling interacts with the kidney, becomes harder to ignore.

STEP-HFpEF, SUMMIT, and Heart Failure With Preserved Ejection Fraction

Heart failure with preserved ejection fraction has been one of the hardest targets in cardiology. For years, trials of everything from neprilysin inhibitors to mineralocorticoid receptor antagonists produced ambiguous or small results in this population. Then, inside of about eighteen months, two incretin-based drugs produced some of the clearest benefit anyone has seen.

STEP-HFpEF, published in NEJM in August 2023, tested semaglutide 2.4 mg in patients with HFpEF and obesity. It showed large improvements in the Kansas City Cardiomyopathy Questionnaire summary score and in six-minute walk distance. STEP-HFpEF DM extended the finding to patients with HFpEF, obesity, and type 2 diabetes. SUMMIT, published in NEJM in November 2024, showed that tirzepatide reduced a composite of cardiovascular death and worsening heart failure events in HFpEF with obesity by 38 percent, alongside the expected symptom and functional improvements.

There is a subtlety here worth flagging. The enrollment criteria for these trials specifically required obesity. How much of the benefit comes from weight loss per se, how much from reduced volume overload, and how much from a direct cardiac effect is not clear. A design decision that asks whether the benefit persists in leaner HFpEF patients would teach the community something that these trials cannot. If you run a cardiac physiology lab, the mechanistic angle is almost certainly a fundable question for the next cycle.

Alzheimer's Disease: evoke, evoke+, and the Hardest Question in the Pipeline

The Alzheimer's trials are the ones I watch most carefully, because the translational story is the least well-established and the implications, if they come in positive, are enormous. Novo Nordisk's evoke and evoke+ trials test oral semaglutide in patients with early Alzheimer's disease. The design is large, with roughly 3,500 participants combined, and the program is powered for a disease-modifying effect over roughly two years.

The scientific rationale has multiple threads. GLP-1 receptors are expressed in the brain. Observational data in patients with type 2 diabetes on GLP-1 therapy has pointed toward reduced dementia incidence, though this is the kind of signal that is easy to misread from insurance claims data. Preclinical work has shown effects on neuroinflammation, on insulin signaling in the CNS, and on tau pathology in some models. None of this establishes that a trial will work.

If evoke reads out positive, it will have implications far beyond Alzheimer's. Every field that has been gesturing at a metabolic-inflammation link in a chronic disease — Parkinson's, ALS, multiple sclerosis — will immediately have a reason to run confirmatory work. The exenatide-PD3 trial for Parkinson's, which did not meet its primary endpoint in 2024, complicates the narrative, but does not close it. A different molecule, a different trial population, and a different endpoint combination could still produce a different answer.

If evoke reads out negative or neutral, the collapse will also be informative. It will force the field to be much more careful about where the boundaries of GLP-1 benefit actually sit, and it will remove some of the narrative pressure that has been pushing these drugs toward indications where the mechanistic evidence is thin.

Addiction and Reward: The NIAAA and NIDA Trials

One of the more unusual signals coming out of the GLP-1 literature is a reduction in addictive behavior. Patients on semaglutide for obesity reported lower alcohol consumption. Patients on tirzepatide reported less craving for high-reward foods and, anecdotally, reduced use of nicotine and other substances. The Kurt Fraser group and others have shown behavioral effects in rodent models that make biological sense: GLP-1 agonism modulates dopaminergic signaling in the ventral tegmental area and nucleus accumbens.

The NIAAA has funded a growing set of randomized trials testing semaglutide in alcohol use disorder. Early readouts, most notably a placebo-controlled trial reported in 2025, showed modest reductions in drinking outcomes that were statistically significant but smaller than some of the observational data had suggested. This is a pattern worth remembering. Retrospective signals from electronic health records almost always overstate treatment effects relative to a properly blinded trial.

NIDA-funded work on nicotine, opioid use disorder, and cocaine is earlier in the pipeline. The behavioral neuroscience community is in an unusually strong position here because the receptor pharmacology is increasingly well-characterized and because the current drugs are available off-label for studies that simply would not have been possible with a novel molecular entity. If you are on the NIDA or NIAAA side of the research world, the funding climate for well-designed mechanistic work is unusually favorable. The relevant question is not whether GLP-1 agonism reduces addictive behavior in some patients. It is which patients, through which circuits, and at what cost in other domains.

Retatrutide, Orforglipron, and the Next Wave

The molecular landscape is shifting faster than most people outside the field realize. Semaglutide is a pure GLP-1 agonist. Tirzepatide is a dual GIP/GLP-1 agonist. Retatrutide, in Phase 3 with Eli Lilly under the TRIUMPH program, is a triple agonist that adds glucagon receptor activity. Its Phase 2 weight-loss data was striking, with mean placebo-adjusted weight loss approaching 24 percent at 48 weeks at the highest dose.

Orforglipron, also from Eli Lilly, is a small-molecule oral GLP-1 agonist that does not require the complicated manufacturing that has constrained peptide supply. If Phase 3 readouts match Phase 2, the world will suddenly have an oral GLP-1 therapy at scale, which changes the access equation in ways that the current injectable market has made hard to appreciate.

CagriSema, the combination of semaglutide with the amylin analog cagrilintide, has had a more mixed story. The REDEFINE-1 trial readout in late 2024 was described by the sponsor as disappointing relative to internal expectations, though it still produced substantial weight loss. The lesson for trialists is familiar. Mechanistic combinations that make biological sense do not always add cleanly in vivo, and the variance in human response is large enough that subgroup analyses will often tell you more than headline numbers.

What This Means for Researchers Writing Grants Right Now

For anyone planning an NIH submission in the next two cycles, the GLP-1 program has reshaped the funding landscape in four specific ways. First, adjacent mechanistic work is fundable. Questions about central nervous system GLP-1 signaling, about vascular biology in incretin-treated patients, and about immunomodulation by these drugs are attracting RFAs and are well-represented in recently awarded R01 portfolios. You can see this pattern clearly in the NIH Reporter data; the count of new awards mentioning GLP-1 mechanisms in aims has risen sharply over the past three fiscal years.

Second, disease-area teams that previously had no reason to talk to each other are now co-applying. Endocrinology, cardiology, nephrology, and neurology groups are running joint human-phenotyping studies, and R01 applications with cross-institute relevance have a natural home. The multi-institute review path is sometimes slower, but the scientific justification almost writes itself when the drug in question has six active indications.

Third, the translational-science case for animal-model work is being rewritten. The classic criticism of a rodent-only aim in metabolic research is that it does not translate. That criticism is weaker when a clinical trial program of this scale is producing human validation in the same timeframe as your preclinical data. Strong mechanistic work in mice that asks a question the trials cannot answer is unusually well-positioned.

Fourth, and more cautiously, if your project is downstream of a specific clinical indication that is still in Phase 3, your timeline needs a contingency. A negative evoke readout would reduce enthusiasm for neurodegeneration-adjacent GLP-1 work. A strongly positive readout would accelerate it. Grants should be written so that the scientific question survives either outcome. This is just good practice, but it is especially relevant when the underlying drug is in active trials.

Frequently Asked Questions