Background
Parkinson's disease (PD) is a progressive disorder characterised by both motor and non-motor problems. Glucagon-like peptide-1 (GLP-1) receptor agonists, licensed fortreatment oftype 2 diabetes, work by stimulating GLP-1 receptors in the pancreas, which triggers the release of insulin. GLP-1 receptors have been found in the brain. Insulin signalling in the brain plays a key role in neuronal metabolism and repair and in synaptic eEicacy, but insulin signalling is desensitised in the brain of people with PD. Researchers are exploring the neuroprotective effects of GLP-1 receptor agonists in neurodegenerative disorders such as PD.

Objectives
To evaluate the eEectiveness and safety of GLP-1 receptor agonists for Parkinson's disease.

Search methods
We searched the Cochrane Movement Disorders Group trials register; the Cochrane Central Register of Controlled Trials (CENTRAL), in the Cochrane Library; and Ovid MEDLINE and Embase. We also searched clinical trials registries, and we handsearched conference abstracts. The most recent search was run on 25 June 2020.

Selection criteria
We included randomised controlled trials (RCTs) of adults with PD that compared GLP-1 receptor agonists with conventional PD treatment, placebo, or no treatment.

Data collection and analysis
Two reviewauthors independently assessed studies forinclusion, extracted data, and assessed risk of bias.We rated the quality of evidence using GRADE. We resolved discrepancies between the two data extractors by consultation with a third review author.

Main results
Through our searches, we retrieved 99 unique records, of which two met ourinclusion criteria. One double-blind study of exenatide versus placebo randomised 62 participants, who self-administered exenatide or placebo for 48 weeks and were followed up at 60 weeks after a 12-week washout. One single-blind study of exenatide versus no additional treatment randomised 45 participants; participants in the intervention group self-administered exenatide for 12 months, and all participants were followed up at 14 months and 24 months following absence of exenatide for 2 months and 12 months, respectively. These trials had low risk of bias, except risk of performance bias was high for Aviles-Olmos 2013.

Exenatide versus placebo

Primary outcomes

We found low-certainty evidence suggesting that exenatide improves motor impairment as assessed by the Movement Disorder Society Unified Parkinson's Disease Rating Scale (MDS-UPDRS) Part III in the oE-medication state (mean difference (MD) -3.10, 95% confidence interval (CI) -6.11 to -0.09). The diEerence in scores was slightly greater when scores were adjusted for baseline severity of the condition (as reported by study authors) (MD -3.5, 95% CI -6.7 to -0.3), exceeding the minimum clinically important difference (MCID).

We found low-certainty evidence suggesting that exenatide has little or no eEect on health-related quality of life (HRQoL) as assessed by the Parkinson's Disease Questionnaire (PDQ)-39 Summary Index (SI) (MD -1.80, 95% CI -6.95 to 3.35), the EuroQol scale measuring health status in five dimensions (EQ5D) (MD 0.07, 95% CI -0.03 to 0.16), or the EQ5D visual analogue scale (VAS) (MD 5.00, 95% CI -3.42 to 13.42).
Eight serious adverse events (SAEs) were recorded, but all were considered unrelated to the intervention. Low-certainty evidence suggests that exenatide has little or no effect on weight loss (risk ratio (RR) 1.25, 95% CI 0.89 to 1.76).

Exenatide versus no treatment

Primary outcomes at 14 months
We found very low-certainty evidence suggesting that exenatide improves motor impairment as assessed by MDS-UPDRS Part III off medication (MD -4.50, 95% CI -8.64 to -0.36), exceeding the MCID. We are uncertain whether exenatide improves HRQoL as assessed by the PDQ-39 SI (MD 3.50, 95% CI -2.75 to 9.75; very low-quality evidence). We found very low-certainty evidence suggesting that exenatide has little or no eEect on the number of SAEs (RR 1.60, 95% 0.40 to 6.32). We found very low-certainty evidence suggesting that exenatide may
lead to weight loss (MD -2.40 kg, 95% CI -4.56 to -0.24).

Primary outcomes at 24 months
We found evidence as reported by study authors to suggest that exenatide improves motor impairment as measured by MDS-UPDRS Part III off medication (MD 5.6 points, 95% CI 2.2 to 9.0). Exenatide may not improve HRQoL as assessed by the PDQ-39 SI (P = 0.682) and may not result in weight loss (MD 0.1 kg, 95% CI 3.0 to 2.8).

Authors' conclusions
Low- or very low-certainty evidence suggests that exenatide may improve motor impairment for people with PD. The difference in motor impairment observed between groups may persistfor some time following cessation of exenatide. This raises the possibility that exenatide may have a disease-modifying eEect. SAEs were unlikely to be related to treatment. The eEectiveness of exenatide for improving HRQoL, non-motor outcomes, ADLs, and psychological outcomes is unclear. Ongoing studies are assessing other GLP-1 receptor agonists.