Inflammatory proteins associated with Alzheimer's disease reduced by a GLP1 receptor agonist: a post hoc analysis of the EXSCEL randomized placebo controlled trial
- Design
- Randomized trial · 3973 participants · Biomarker outcome
- Match to healthy normal-weight adults aged 55–75
- Different population[Auto] Participants had type 2 diabetes (age/BMI not reported in abstract).
- Could weight loss explain it?
- Possibly[Auto] Participants had obesity and/or type 2 diabetes and the abstract does not separate direct drug effects from weight loss or glycaemic improvement.
- Study tier
- Study tier 2[Auto] Randomized trial of moderate size; outcome type and follow-up limit certainty (auto-provisional).
- Assessment
- Version 1 · automatic, not yet reviewed by a person · Sep 13, 2026
Study facts come from the paper. Population match, weight-loss explanation and study tier are our judgments, made against the reference group of healthy normal-weight adults aged 55–75.
[Auto, unreviewed; quoted from abstract conclusions] EQW treatment was associated with significant change in inflammatory proteins associated with AD.
01Findings
What the study reported
- Drugs
- Exenatide
- Dose
- 2 mg
- Comparator
- placebo
- Primary outcome
- Not extracted
- Effect
- Not extracted
- 95% confidence interval
- Not extracted
- Follow-up
- Not stated
- Adverse events
- Not extracted
- Limitations
- Auto-classified from abstract only; effect estimates, adverse events and limitations not extracted. Requires manual review.
Who was studied
- Diabetes status
- type 2 diabetes present (all or most)
- Baseline condition
- Alzheimer's disease / MCI
- Sample size
- 3973
Study quality details
- Study design
- Randomized controlled trial
- Sample size
- 3973
- Randomization
- yes
- Blinding
- not stated
- Comparator
- placebo
- Follow up duration
- not stated
- Outcome type
- biomarker
- Replication
- not assessed (auto)
- Consistency with other evidence
- not assessed (auto)
- Population applicability
- INDIRECT
- Statistical precision
- not extracted
- Funding conflicts
- unclear
- Peer review status
- yes
01bRisk of bias
How much the result can be relied on
Risk of bias is judged per result, not per paper: the same study can report one marker at low risk and another at high. Two reviewers assess each result independently against a written guide, and every domain judgment below carries its reasoning and the sentences it rests on.
These words are not quality scores. On the RoB 2 scale, High risk of bias is the worst rating a result can receive and Low risk of bias is the best — the opposite of how the same words read on some other scales. Levels are always written out in full here for that reason.
Assessed, not settled (1)
Recorded rather than omitted. Leaving these out would make the appraisal look cleaner than it is, and a disagreement between two careful readers is itself worth knowing.
CRP (SomaScan aptamer measurement), treatment x timepoint interaction, EQW 2 mg vs placebo · baseline to 1 year
Assessed, not settled · RoB 2 · Two reviewers reached different judgments and nothing has settled it
This result has no overall judgment, because the two independent reviews did not reach one. Both readings are shown. The domains they did agree on are below.
| Where they differ | The two readings |
|---|---|
| D3 · Missing outcome data | One reviewer: Some concerns about risk of bias The other: High risk of bias |
| D4 · Measurement of the outcome | One reviewer: Some concerns about risk of bias The other: Low risk of bias |
| Domain | Judgment and reasoning |
|---|---|
| D1 · Randomisation process Low risk of bias | Parent trial identified by name and registration (EXSCEL, NCT01144338), so under guide item 6 the large-registered-trial presumption reaches the parent and this secondary analysis inherits it. Allocation concealment is PRESUMED from EXSCEL's character as a multinational 14,752- participant double-blind placebo-controlled registration trial, identified in this paper's Methods and abstract; it is NOT verified — no retrieved source describes the randomization sequence generation or concealment mechanism (the EXSCEL primary report is cited but not retrieved). The presumption is not rebutted: baseline characteristics of the biomarker cohort arms are concordant, including baseline CRP (-0.01 vs 0.00, p = 0.89). Baseline Module 2 score differs at p = 0.008 across many compared variables, which is within chance for a table of this width and does not bear on CRP.EXSCEL is registered on ClinicalTrials.gov: NCT01144338 on 10th of June 2010. EXSCEL was a multinational, double-blind, placebo-controlled, randomized trial evaluating the impact of the EQW on CV outcomes in people with T2D |
| D2 · Deviations from intended interventions Some concerns about risk of bias | Double blinding is stated at the parent-trial level, so participants and personnel were probably not aware of assignment (PN). However the paper reports a high rate of study drug discontinuation in EXSCEL and gives no per-arm discontinuation figures, no adherence data for the biomarker subset, and no statement that the biomarker analysis was intention-to-treat or that any sensitivity analysis addressed non-adherence. The mixed model is fitted on randomized group, which is the right analysis for the effect of assignment, but the guide directs D2 to be judged on discontinuation imbalance and ITT handling and both are NI here. Not High: nothing suggests the analysis departed from randomized groups, and biomarker readouts are not behaviourally mediated.EXSCEL was a multinational, double-blind, placebo-controlled, randomized trial Furthermore, there was a high rate of study drug discontinuation [ 21 ]. |
| D3 · Missing outcome data Some concerns about risk of bias | Against the randomized denominator (14,752 randomized in EXSCEL) only 3,973 contributed paired baseline and 1-year samples = 26.9% available, i.e. 73.1% of randomized participants lack the outcome. The >20% prong of the guide's D3 rule trips decisively, so "data available for nearly all" is N. The per-arm split IS available for the analysed set: 1,994 EQW vs 1,979 placebo (Table 1), i.e. essentially 50/50, and against approximately equal randomized arm sizes the between-arm difference in availability is well under 5 points, so the second prong does not trip. Per guide ruling 2, only one prong tripping means this is not automatic High. Availability was structural rather than outcome-driven — samples were taken only from consenting participants at a subset of sites, a condition set at baseline and prior to the outcome — and a supplementary comparison of the overall and biomarker cohorts is cited (Supplementary Table 1, not retrieved). But the 1-year sample requires survival in follow-up, exenatide discontinuation was high, and no sensitivity analysis addresses missingness that could depend on the true CRP value. Some concerns, not Low (the 73% shortfall is too large to call Low) and not High (no evidence of differential, value-dependent loss; arms balanced).It enrolled 14,752 participants (73.1% with and 26.9% without previous CV disease at 687 sites in 35 countries Of the 5668 EXSCEL participants who provided biomarker samples, we selected 3973 who had baseline and 1-year blood samples (Fig. 1 ). |
| D4 · Measurement of the outcome Some concerns about risk of bias | Measurement was identical in both arms, centrally performed on a single SomaScan v4 platform, with no indication the assay changed partway through, and outcome assessors were blinded by the parent trial's double-blind design — so the classic D4 risks are absent. The concern is the measure itself: this is not the central-lab hs-CRP the guide calls objective, but a relative aptamer-based measurement expressed in standard-deviation units after scaling, with no reported calibration to clinical hs-CRP concentrations and no accuracy or reproducibility data in the retrieved sources. For a claim about CRP concentration change, an unvalidated aptamer proxy is an inappropriate-measurement concern rather than a differential-assessment one. Not High: any aptamer measurement error is non-differential between randomized arms and biases toward the null rather than creating the observed contrast.Plasma proteins were measured using the SomaScan assay platform (SomaLogic Inc.), which uses slow off-rate modified DNA aptamers (SOMAmers) to bind to target proteins and quantify the relative concentrations of proteins. For this study, we used the v.4 assay comprising of 4,979 human proteins SOMAmer reagents mapped to 4776 unique proteins. All proteins were scaled to a mean 0 and standard deviation 1 distribution before analyses. |
| D5 · Selection of the reported result High risk of bias | Guide item 7 applies directly and affirmatively. The authors label the whole analysis post hoc in the title and again in the Introduction; that is affirmative evidence of non-prespecification and sets D5 to High, quoted, regardless of protocol retrievability (no protocol, SAP or registry entry was retrieved here). Two aggravating features rather than mitigating ones: (a) the eight/nine endpoints were selected after the fact from a 4,979-protein panel on the basis of a prior dementia-cohort paper, and the age and prior-CV subgroups are framed as things the authors "explored"; (b) multiplicity correction is partial and the corrected model is not the one carrying the paper's CRP conclusion — Benjamini-Hochberg FDR was applied only to the Wilcoxon tests and Model A, while Models B and C, including the Model C CRP interaction p < 0.001 reported in Table 3 and the subgroup CRP results, are nominal only. The result assessed here is the Model A FDR-corrected interaction, which IS multiplicity-corrected; but the post hoc statement alone is decisive under item 7, and the mixed correction scheme across models compounds it.Inflammatory proteins associated with Alzheimer's disease reduced by a GLP1 receptor agonist: a post hoc analysis of the EXSCEL randomized placebo controlled trial In this post-hoc analysis of the Exenatide Study of Cardiovascular Event Lowering (EXSCEL) [ 16 ], we sought to provide further evidence for the potential efficacy and mechanism of action in relation to AD of once-weekly exenatide (EQW) |
02Funding
Funding and conflicts
- Funding
- not reported in abstract
- Industry funded
- Unclear
- Manufacturer
- Novo Nordisk, Eli Lilly, Lilly, AstraZeneca, Sanofi, Boehringer Ingelheim, Amgen, Pfizer, Roche
- Sponsor role
- not reported in abstract
- Author conflicts
- IK received research support and honoraria from Novo Nordisk. He is a paid medical advisor for biotechnology (cfdx Ltd) and digital healthcare companies working in dementia (Five Lives SAS, Cognetivity, Mantrah Ltd). RJM received research support and honoraria from Abbott, American Regent, Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Boston Scientific, Cytokinetics, Fast BioMedical, Gilead, Innolife, Eli Lilly, Medtronic, Medable, Merck, Novartis, Novo Nordisk, Pfizer, Pharmacosmos, Relypsa, Respicardia, Roche, Rocket Pharmaceuticals, Sanofi, Verily, Vifor, Windtree Therapeutics, and Zoll. RRH reports personal fees from Anji Pharmaceuticals, AstraZeneca, and Novartis.
- Independent replication
- unknown
- Notes
- Authors declare relationships with the drug's manufacturer: AstraZeneca (originally Amylin/Eli Lilly)
Funding is shown on every study and never used to score it.
04Source
The source, as retrieved
Abstract
[BACKGROUND] Glucagon-like peptide-1 receptor agonists are a viable option for the prevention of Alzheimer's disease (AD) but the mechanisms of this potential disease modifying action are unclear. We investigated the effects of once-weekly exenatide (EQW) on AD associated proteomic clusters. [METHODS] The Exenatide Study of Cardiovascular Event Lowering study compared the cardiovascular effects of EQW 2 mg with placebo in 13,752 people with type 2 diabetes mellitus. 4,979 proteins were measured (Somascan V0.4) on baseline and 1-year plasma samples of 3,973 participants. C-reactive protein (CRP), ficolin-2 (FCN2), plasminogen activator inhibitor 1 (PAI-1), soluble vascular cell adhesion protein 1 (sVCAM1) and 4 protein clusters were tested in multivariable mixed models. [RESULTS] EQW affected FCN2 (Cohen's d -0.019), PAI-1 (Cohen's d -0.033), sVCAM-1 (Cohen's d 0.035) and a cytokine-cytokine cluster (Cohen's d 0.037) significantly compared with placebo. These effects were sustained in individuals over the age of 65 but not in those under 65. [CONCLUSIONS] EQW treatment was associated with significant change in inflammatory proteins associated with AD. [TRIAL REGISTRATION] EXSCEL is registered on ClinicalTrials.gov: NCT01144338 on 10th of June 2010.
Where this record came from
| Source | Retrieved | Identifier |
|---|---|---|
| pubmed | Sep 13, 2026 | 39358806 first ingestion |