GLP-1 Drugs and Parkinson's Disease: What LIXIPARK and Exenatide-PD3 Showed

No GLP-1 drug is approved for Parkinson’s disease, and the largest trial so far found that one of them did not slow it. In the 2025 phase 3 Exenatide-PD3 trial, motor scores worsened slightly more on weekly exenatide than on placebo over two years.^[2] A smaller 2024 trial of daily lixisenatide, LIXIPARK, found less motor decline at one year, the result behind most of the headlines.^[1] Large patient-record studies link these drugs to fewer Parkinson’s diagnoses in people with diabetes, a finding that has not been tested in a trial.^[6]

Why Parkinson’s researchers became interested

In laboratory and animal models of Parkinson’s, GLP-1 drugs showed protective effects on nerve cells, and the first candidates tested in people were exenatide and lixisenatide, both older diabetes drugs.^[1][2] The trials asked whether that protection carries over to people, which means whether the drug slows the disease rather than only easing symptoms. All four trials below used the same scale, the motor section of the MDS-UPDRS, where a higher score means more disability.

The randomized trials

LIXIPARK is the trial most often cited as evidence that GLP-1 drugs work in Parkinson’s. Over 12 months, people on lixisenatide held their motor scores almost exactly where they started, while those on placebo worsened by about three points.^[1] Three points on a 132-point scale is a small change for any one person, but it was the whole of the decline the placebo group showed over the year, which is why the result drew attention. The other measures the trial tracked did not differ much between groups, and 46% of people on lixisenatide had nausea.^[1] The authors called for longer and larger trials.

Exenatide-PD3 was that kind of trial, for exenatide. It followed the 2017 single-center result with a phase 3 study more than three times the size and twice the length. Motor scores off medication worsened in both groups, slightly more on exenatide, and the difference was not significant.^[2] The authors wrote that they found no evidence to support exenatide as a disease-modifying treatment, while noting that the drug was safe and well tolerated.^[2] Our article on exenatide covers the drug itself.

The 2017 exenatide trial is a reminder of how small early results can mislead. Its own authors wrote that it was uncertain whether the drug changed the disease or simply had a lasting effect on symptoms.^[3] When four placebo-controlled trials with 667 patients were pooled in a 2026 meta-analysis, GLP-1 drugs showed no difference from placebo on motor scores, other symptoms, thinking or quality of life.^[5]

Why the trials point in different directions

The positive and negative trials differ in ways that make a single verdict hard. They used different drugs: lixisenatide daily, exenatide weekly, and a long-acting pegylated exenatide.^[1][2][4] They enrolled people at different stages. LIXIPARK took people diagnosed within the previous three years and without motor complications, NLY01 took people not yet on Parkinson’s treatment, while the 2017 exenatide trial took people with moderate disease who were already having wearing-off effects.^[1][4][3]

They also measured at different moments. LIXIPARK’s main result was scored while people were on their usual Parkinson’s medicines; the exenatide trials scored people after their medicines had worn off, which is closer to a measure of the underlying disease.^[1][2][3] After LIXIPARK’s two-month washout, off-medication scores still favored lixisenatide (17.7 against 20.6), a hint that the effect was more than symptomatic, but that was a secondary finding in a phase 2 trial.^[1] The NLY01 authors saw a possible benefit in younger participants, again only in a subgroup.^[4] Exenatide-PD3’s authors concluded that future studies need drugs with better target engagement or specific groups of patients.^[2]

The observational data on Parkinson’s risk

A separate question is whether people taking a GLP-1 are less likely to be diagnosed with Parkinson’s in the first place. A 2026 study of national registers in Denmark, Norway and Sweden compared 158,961 new users of GLP-1 drugs with 188,065 new users of sulfonylureas, an older diabetes drug, in people aged 45 and over. Parkinson’s was diagnosed at 5.2 per 10,000 person-years on GLP-1 drugs against 8.0 on sulfonylureas, an adjusted hazard ratio of 0.81.^[6]

That design is better than most, because it compares two groups of people with diabetes starting a new drug. It still cannot rule out differences between the people who were prescribed each one. When the analysis ignored diagnoses in the first two years, to allow for Parkinson’s that was already developing, the result was no longer statistically clear (hazard ratio 0.84, with a confidence interval crossing 1).^[6] Most of the GLP-1 use was liraglutide; semaglutide made up 13.4% of follow-up time and tirzepatide was not studied. The authors described the finding as needing replication.^[6]

What is shown, and what is not

• Shown: in one phase 2 trial, daily lixisenatide slowed motor decline over a year in early Parkinson’s, at the cost of frequent nausea.^[1]
• Shown: in a phase 3 trial, weekly exenatide did not slow Parkinson’s over two years.^[2] A long-acting exenatide also failed.^[4]
• Not shown: that semaglutide or tirzepatide has any effect on Parkinson’s. Neither has a completed randomized trial in the disease.
• Not shown: that taking a GLP-1 prevents Parkinson’s. The risk data come from patient records in people with diabetes, and they have not been tested in a trial.^[6]

For someone with Parkinson’s who also has type 2 diabetes or obesity, a GLP-1 can be prescribed for those conditions, and in the pooled trials motor scores on the drug did not differ significantly from placebo in either direction.^[5] Weight loss deserves watching: the pooled analysis found it was more common on GLP-1 drugs than on placebo in people with Parkinson’s, and nausea was common in the lixisenatide and NLY01 trials.^[5][1][4] Whether to start one is a question for the neurologist and the prescriber together.

Alzheimer’s went the same way for semaglutide, with two large trials finding no slowing of cognitive decline; that story is in the EVOKE result. For other effects on the nervous system, see GLP-1 drugs and brain fog and daytime sleepiness. The cardiovascular outcome trials of the two drugs tested here are summarized on the EXSCEL and ELIXA trial pages.