GLP-1 Medicines and Parkinson's Disease: What LIXIPARK, Exenatide-PD3 and the Cohort Data Show

Can GLP-1 medicines slow Parkinson's disease? Not on current evidence. Two small trials of GLP-1 drugs in Parkinson's disease found less motor decline than placebo, and two later trials, including the largest one, found none. Cohort studies keep reporting fewer Parkinson's diagnoses among people with diabetes who take these drugs, which is why the question has not gone away.

The idea is older than the weight-loss boom. Exenatide, a GLP-1 receptor agonist, showed neuroprotective effects in preclinical models of Parkinson's disease, and the drug class has receptors in the brain, which our explainer on what GLP-1 receptors do in the brain covers. Animal and cell findings are a reason to run trials, not evidence that a drug works in people.

This article goes through each placebo-controlled trial, the observational data, the side effects that matter in Parkinson's disease, and what is known about lower or microdose regimens, which in this condition is nothing at all.

How do Parkinson's trials measure whether a drug works?

The MDS-UPDRS part III

Parkinson's trials of GLP-1 drugs mostly use the motor section of the Movement Disorder Society Unified Parkinson's Disease Rating Scale, called MDS-UPDRS part III. A clinician scores tremor, rigidity, slowness and walking, and higher scores mean more disability. The LIXIPARK investigators give the scale's range as 0 to 132.

Two details shape how the results read. Some trials score people in the "off" state, after withholding their usual Parkinson's medicine, which shows more of the underlying disease. Others score in the "on" state. And the goal of these trials is disease modification, meaning slower decline over a year or more, rather than the quick symptom relief that levodopa provides. A gap of a few points between drug and placebo is the kind of result these trials are designed to detect.

What did the first exenatide trial find?

Athauda and colleagues, The Lancet, 2017

The 2017 exenatide trial found a modest motor benefit that outlasted the treatment. Athauda and colleagues at University College London randomized 62 people with moderate Parkinson's disease to exenatide 2 mg or placebo once weekly for 48 weeks, on top of their usual medicines, followed by a 12-week washout.

At 60 weeks, off-medication motor scores had improved by 1.0 point on exenatide and worsened by 2.1 points on placebo, an adjusted difference of 3.5 points (p=0.0318). Injection site reactions and gastrointestinal symptoms were common in both groups. There were six serious adverse events on exenatide and two on placebo, none judged related to the study drug.

The authors were careful about what this meant. They wrote that it was uncertain whether exenatide changed the disease itself or produced a lasting effect on symptoms. With about 30 people per group at a single centre, the trial was a signal to test further rather than an answer.

What did the LIXIPARK trial show?

Meissner and colleagues, New England Journal of Medicine, 2024

LIXIPARK found that lixisenatide held motor scores roughly flat for a year while placebo patients declined. The French phase 2 trial enrolled 156 people diagnosed with Parkinson's disease less than 3 years earlier, on stable symptom medicines and without motor complications, and gave 78 of them daily lixisenatide injections and 78 placebo for 12 months.

Baseline motor scores were about 15 in both groups. At 12 months the score had changed by -0.04 points on lixisenatide and worsened by 3.04 points on placebo, a difference of 3.08 (P=0.007). After a 2-month washout, at 14 months, off-medication scores were 17.7 on lixisenatide and 20.6 on placebo. Other secondary end points did not differ substantially.

The cost was gastrointestinal: nausea in 46% of the lixisenatide group and vomiting in 13%. The investigators concluded that longer and larger trials are needed to establish the effect and safety of lixisenatide in Parkinson's disease.

Why did the phase 3 exenatide trial find nothing?

Exenatide-PD3, Vijiaratnam and colleagues, The Lancet, 2025

Exenatide-PD3 was the test the 2017 result called for, and exenatide failed it. The trial ran at six UK research hospitals and randomized 194 people aged 25 to 80, 97 to extended-release exenatide 2 mg once weekly and 97 to placebo, for 96 weeks. 29% of participants were women and 71% were men.

At 96 weeks, off-medication motor scores had worsened by 5.7 points on exenatide and 4.5 points on placebo (p=0.47). Serious adverse events affected 9% of the exenatide group and 11% of the placebo group. The authors concluded that exenatide was safe and well tolerated but that they found no evidence to support it as a disease-modifying treatment.

The investigators did not fully close the door. They suggested that studies with agents showing better target engagement, or focused on specific subgroups of patients, are needed. Put plainly, exenatide may not reach the brain in the amounts needed, or only some patients might respond.

One more point for anyone citing this trial. In June 2026 The Lancet published an Expression of Concern on the Exenatide-PD3 paper. Such a notice tells readers that the editors have questions about a publication while they look into them. It is not a retraction, and as of this writing the published result stands.

Did a brain-penetrant GLP-1 drug do better?

NLY01, McGarry and colleagues, Lancet Neurology, 2024

No. NLY01 is a pegylated, longer-lasting version of exenatide built to cross into the brain and calm microglia, the brain's immune cells. The trial, at 58 movement disorder clinics in the USA, randomized 255 people with early untreated Parkinson's disease to 2.5 mg NLY01, 5.0 mg NLY01 or placebo for 36 weeks, 85 per group.

Neither dose differed from placebo on the combined MDS-UPDRS parts II and III: the difference was -0.39 points for 2.5 mg (p=0.77) and 0.36 for 5.0 mg (p=0.79). The authors found no improvement in motor or non-motor features. A subgroup analysis raised the possibility of a benefit in younger participants, which they flagged as exploratory.

NLY01 matters for the "better brain penetration" argument. It was designed for exactly that, and it still did not separate from placebo at 36 weeks.

What do the meta-analyses of these trials conclude?

Two 2026 reviews, two different answers

The pooled analyses disagree, depending on which trials and outcomes they include. A 2026 meta-analysis in Frontiers in Neurology pooled 8 randomized trials with 850 people and reported a small motor benefit, a standardized mean difference of -0.21 after treatment and -0.32 at follow-up.

A network meta-analysis in Neurological Sciences, also from 2026, pooled 5 trials with 708 people and found no significant overall improvement in on-state motor scores (mean difference -2.00, 95% CI -5.46 to 1.46) and no effect on off-state scores. It did find a benefit for lixisenatide 20 µg a day specifically, a mean difference of -3.08, which is the LIXIPARK result showing through. When two careful reviews of a small set of trials reach different conclusions, the honest reading is that the question is open.

Do people on GLP-1 drugs develop Parkinson's less often?

The observational evidence

In large cohorts of people with type 2 diabetes, GLP-1 users are diagnosed with Parkinson's disease less often than users of other diabetes drugs. The biggest recent example is a 2026 Scandinavian study by Engström and colleagues in Diabetes, Obesity and Metabolism, which used national registers in Denmark, Norway and Sweden to compare 158,961 new GLP-1 users with 188,065 new sulfonylurea users, all aged 45 or older.

The Parkinson's incidence rate was 5.2 per 10,000 person-years on GLP-1 drugs and 8.0 on sulfonylureas, an adjusted hazard ratio of 0.81. When the first 2 years of use were excluded, to reduce the chance that early, undiagnosed Parkinson's was driving drug choice, the hazard ratio was 0.84 with a confidence interval reaching 1.02, so that version of the analysis was not statistically clear.

A 2026 meta-analysis in Frontiers in Endocrinology pooled 6 real-world studies with 452,763 participants with type 2 diabetes and found a hazard ratio of 0.68. The authors themselves called the results preliminary and hypothesis-generating rather than conclusive. Cohort studies cannot fully account for differences between the people who are prescribed one drug over another.

Does semaglutide or tirzepatide help Parkinson's disease?

The evidence gap for newer drugs

Nobody can say yet. None of the placebo-controlled trials above tested semaglutide or tirzepatide, and even the large cohort studies were dominated by older drugs. In the Scandinavian cohort, liraglutide accounted for 72.9% of GLP-1 follow-up time and semaglutide for 13.4%.

There is a hint that biology may decide who responds. A 2026 study in Brain followed 7,039 adults with type 2 diabetes who started a GLP-1 drug and used a genetic score as a proxy for how much GLP-1 receptor each person expresses. Those with a higher score had a lower hazard of Parkinson's disease and related disorders, a hazard ratio of 0.78. That fits the Exenatide-PD3 authors' suggestion that subgroups may matter, though it remains a hypothesis. A similar story played out with dementia, where the large EVOKE trials of semaglutide tested the cohort signal directly; our article on GLP-1 medicines and dementia covers what they found.

What side effects matter for people with Parkinson's?

Nausea, gut symptoms and weight

Gastrointestinal effects dominated every trial. In NLY01, nausea affected 39% on 2.5 mg and 58% on 5.0 mg versus 19% on placebo, and any gastrointestinal disorder affected 61%, 75% and 36%. In LIXIPARK, 46% had nausea. The pattern matches what our guide to GLP-1 nausea and vomiting describes in other populations, and it climbs with dose.

Weight loss is the other issue. The Frontiers in Neurology meta-analysis found weight loss more frequent on GLP-1 drugs, with an odds ratio of 3.87, and gastrointestinal disorders with an odds ratio of 2.85. Many people with Parkinson's disease are not trying to lose weight, so an appetite-suppressing drug brings a trade-off that does not exist for someone with obesity. The same review reported fewer Parkinson's symptom exacerbations on GLP-1 drugs.

Has microdosing a GLP-1 been tested for Parkinson's disease?

What can and cannot be said

No. Every Parkinson's trial used a standard or higher dose: exenatide 2 mg weekly in both UCL trials, lixisenatide 20 µg a day in LIXIPARK, and NLY01 at 2.5 or 5.0 mg weekly. The cohort studies followed people with diabetes on ordinary prescriptions. There are no data on whether a smaller dose does anything in the Parkinson's brain.

The direction of the open question also runs against low doses. One of the two directions the Exenatide-PD3 authors proposed is agents with better target engagement, which, if it proves right, points toward more drug reaching the brain, not less. NLY01 tried that route and still came up empty at 36 weeks.

Lower doses do have one sourced advantage: gastrointestinal side effects in the NLY01 trial were less common at 2.5 mg than at 5.0 mg. For someone already taking a GLP-1 for weight or blood sugar, a slow dose climb, of the kind described in our article on the gentle ramp, is reasonable to discuss with a prescriber. Using any dose as a Parkinson's treatment is outside what the evidence supports.

Frequently Asked Questions