Exenatide exerts a potent antiinflammatory effect
- Design
- Randomized trial · 24 participants · Biomarker outcome
- Match to healthy normal-weight adults aged 55–75
- Different populationINDIRECT: participants had mean BMI 39.8 (exenatide) and 39.1 (placebo) and mean HbA1c 8.6% and 8.5%; the abstract does not name the condition they were enrolled for.
- Could weight loss explain it?
- PossiblyThere was no weight loss and acute effects were measured up to 6 h after a single dose, which separates the effect from weight loss; but fasting glucose, HbA1c and free fatty acids fell and were not separated from the inflammatory markers, so improved glycaemia could still contribute. No formal adjustment was reported.
- Study tier
- Study tier 3LOW: randomised placebo-controlled trial of 24 participants over 12 weeks with biomarker outcomes only; abstract only.
- Assessment
- Version 2 · Two reviewers differed; settled by a recorded ruling · Oct 19, 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.
This small randomised trial gave 24 patients (mean BMI about 39) exenatide or placebo injections for 12 weeks. With exenatide, several inflammation markers in blood cells and plasma fell, including plasma IL-6 by 22 ± 12%, HbA1c fell from 8.6 to 7.4%, and there was no weight loss. The abstract reports these changes within the exenatide group and does not give a direct comparison with placebo.
What the study reported
- Drugs
- Exenatide
- Dose
- 10 μg twice daily for 12 wk; single 5 μg dose for acute testing
- Comparator
- placebo twice daily
- Route
- subcutaneous
- Treatment duration
- 12 weeks
- Primary outcome
- inflammatory markers in mononuclear cells and plasma (ROS, NF-κB, cytokine mRNA, plasma MCP-1, MMP-9, SAA, IL-6)
- Effect
- After 12 wk of exenatide: ROS generation fell 22 ± 9% and NF-κB binding 26 ± 7%, TNFα mRNA 31 ± 12%, plasma IL-6 22 ± 12% (P < 0.05 for all); no weight loss; changes are within the exenatide group, comparison with placebo not reported in the abstract
- 95% confidence interval
- Not extracted
- P value
- P < 0.05 for all reported changes
- Follow-up
- 12 wk (samples at 0, 3, 6 and 12 wk)
- Adverse events
- not reported in abstract
- Limitations
- Very small trial (12 per arm) over 12 weeks with biomarker outcomes only; the abstract reports changes within the exenatide group rather than versus placebo, and states no blinding or diabetes diagnosis.
Who was studied
- Mean age
- 56 ± 3 yr (exenatide); 54 ± 4 yr (placebo)
- Bmi mean
- 39.8 ± 2 kg/m2 (exenatide); 39.1 ± 1.6 kg/m2 (placebo)
- Obesity status
- not reported in abstract as a criterion (mean BMI 39.8 and 39.1)
- Diabetes status
- not reported in abstract
- Cvd status
- not reported in abstract
- Ckd status
- not reported in abstract
- Baseline condition
- not reported in abstract
- Sample size
- 24
Study quality details
- Study design
- Randomized controlled trial
- Sample size
- 24
- Randomization
- yes
- Blinding
- not stated
- Comparator
- placebo twice daily
- Follow up duration
- 12 wk (samples at 0, 3, 6 and 12 wk)
- Outcome type
- biomarker
- Replication
- not assessed
- Consistency with other evidence
- not assessed
- Population applicability
- INDIRECT
- Statistical precision
- no CI reported; mean ± SE percentages with P < 0.05; 12 per arm
- Funding conflicts
- unclear
- Peer review status
- yes
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.
Risk of bias is not the same as whether the numbers work. A trial can be well conducted and still print a standard deviation that is arithmetically impossible, or a table whose sign contradicts its own text. Risk of bias has no question for that, so where it was found it is recorded separately below, under Source integrity.
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.
Plasma IL-6 concentration (the only graded systemic marker measured here). CRP was never measured, so this record yields no O1. TNF-alpha appears only as mononuclear-cell mRNA expression, a different construct from circulating TNF-alpha protein, and is therefore NOT assessed as O3. · 12 weeks
High risk of bias · RoB 2 · Two reviewers agreed on every domain
Why the overall is not simply the worst domain: The RoB 2 algorithm over D1 Some concerns / D2 Some concerns / D3 Some concerns / D4 Low / D5 Some concerns returns Some concerns. This is an override to High, with the written reason required by general rule 5, and the reason is not the accumulation of NIs. RoB 2 rates the risk that a reported effect estimate is biased. The estimate C-001 requires for O4 is a between-arm contrast in IL-6 change, exenatide versus placebo. That estimate is not reported anywhere in this source. What is reported is a within-arm change from baseline in the exenatide arm only, tested against zero. A within-arm change confounds the treatment effect with regression to the mean, seasonal and assay drift, and any secular change in the cohort — which is precisely what the randomized placebo arm existed to remove, and its data are not used in the reported comparison. A within-arm pre-post change is a biased estimator of the randomized contrast by construction, and randomization provides no protection to it. That is a defect in the reported result itself, not an information gap, and it is the dominant one. The weight-independence design, which is what makes this record interesting for O4, is also unverified: "There was no weight loss" is an unquantified assertion with no per-arm weight values, no SD and no equivalence test at n = 12 per arm, where the confidence interval around a null weight change would be wide enough to admit a clinically meaningful loss. Taken as a design claim it is plausible; taken as evidence that the IL-6 change is weight-independent it is unsupported. The unverified domains compound this rather than drive it: with no allocation-concealment method, no blinding statement, no analysis population, and no completer count at n = 12 per arm, there is no countervailing evidence that would justify a lower judgment. Per the task instruction, none of these NIs is read as evidence of poor conduct; the High rests on the defect in the reported estimate.
| Domain | Judgment and reasoning |
|---|---|
| D1 · Randomisation process Some concerns about risk of bias | Randomization is asserted and arms are numerically balanced on age, BMI and HbA1c as reported. But this is a small single-centre academic trial with no named parent trial and no registration number, so the v1.2 inheritance presumption does not apply (its first mandatory limit fails). The abstract describes no sequence-generation method and no allocation concealment mechanism, so under the D1 domain rule for small academic trials the answer is NI. Baseline balance on three reported variables is not evidence of concealment and is not treated as such.Twenty-four patients were prospectively randomized to be injected sc with either exenatide 10 μg twice daily mean age = 56 ± 3 yr; mean body mass index = 39.8 ± 2 kg/m(2); mean glycosylated hemoglobin (HbA1c) = 8.6 ± 0.4% |
| D2 · Deviations from intended interventions Some concerns about risk of bias | A placebo injection arm is stated, which is consistent with participant and carer blinding, but blinding is nowhere claimed and no blinding procedure is described. Nothing is reported on discontinuations, crossovers, co-interventions, or the analysis population, and the guide directs that D2 be judged on discontinuation imbalance and ITT handling, both of which are absent here. GLP-1 GI adverse events are not reported at all, so the unblinding prong cannot be evaluated either way. All substantive signalling questions are NI.or placebo twice daily (n = 12 Blood samples were also collected for up to 6 h after a single dose of exenatide (5 μg) or placebo. |
| D3 · Missing outcome data Some concerns about risk of bias | No information. The abstract reports 12 randomized per arm and gives a sampling schedule, but never states how many participants contributed an IL-6 measurement at 12 weeks, and reports no withdrawals. The v1.1 ruling 2 thresholds (>20% missing, >5-point arm difference) cannot be evaluated because the completer denominator is unknown; that ruling makes D3 Low only when the thresholds are shown not to trip, which cannot be shown here. At n = 12 per arm a single dropout is 8 percentage points, so the thresholds are not remote. Absence of a reported dropout is not evidence that there was none.Fasting blood samples were obtained at 0, 3, 6, and 12 wk. Twenty-four patients were prospectively randomized |
| D4 · Measurement of the outcome Low risk of bias | Plasma IL-6 is a laboratory-measured, objective biomarker, and the D4 domain rule in the guide treats IL-6 as objective with measurement bias Low unless the assay changed partway through or measurement differed between arms. Neither is reported or suggested. Note the limits of that judgment from an abstract: the assay platform, the laboratory, and whether outcome assessors were blinded are all unreported, and this is a rule-driven Low rather than a verified one. The residual concern that the *analysis* of the measurement was investigator-chosen is carried at D5 and in the overall judgment, not here.The plasma concentrations of monocyte chemoattractant protein-1, matrix metalloproteinase-9, serum amyloid A, and IL-6 were suppressed after 12 wk exenatide treatment by 15 ± 7, 20 ± 11, 16 ± 7, and 22 ± 12%, respectively (P < 0.05 for all). |
| D5 · Selection of the reported result Some concerns about risk of bias | v1.3 ruling 7, second limb. No protocol, no SAP and no registry entry are retrievable, and the full text is paywalled; the abstract contains no author statement that the IL-6 analysis was exploratory, post hoc, or unadjusted for multiplicity, and equally no claim that it was prespecified. Prespecification is therefore unverified rather than affirmatively absent, and D5 sits at the "at least Some concerns" floor — not High under ruling 3 (which requires a searchable protocol shown not to contain the analysis), and not Low (which requires a retrieved source showing the outcome and its analysis). Separately notable but carried in limitations rather than escalating D5: the abstract reports roughly eighteen marker results across two experiments, all significant at P < 0.05, with no multiplicity handling mentioned.Fasting blood samples were obtained at 0, 3, 6, and 12 wk. (P < 0.05 for all) |
Funding and conflicts
- Funding
- not reported in abstract
- Industry funded
- Unclear
- Manufacturer
- not stated in source
- Sponsor role
- not reported in abstract
- Author conflicts
- not reported in abstract
- Independent replication
- unknown
Funding is shown on every study and never used to score it.
Claims this study bears on
- GLP-1 receptor agonist treatment reduces systemic inflammatory markers beyond what can be explained by treatment-associated weight loss.Supports
In 24 adults with obesity and type 2 diabetes, 12 weeks of exenatide without weight loss lowered plasma IL-6 by 22 ± 12%, MCP-1 by 15 ± 7% and serum amyloid A by 16 ± 7% from baseline. The abstract reports within-arm changes only, with no between-arm contrast against placebo.
The source, as retrieved
Abstract
[OBJECTIVE] Our objective was to determine whether exenatide exerts an antiinflammatory effect. [RESEARCH DESIGN AND METHODS] Twenty-four patients were prospectively randomized to be injected sc with either exenatide 10 μg twice daily [n = 12; mean age = 56 ± 3 yr; mean body mass index = 39.8 ± 2 kg/m(2); mean glycosylated hemoglobin (HbA1c) = 8.6 ± 0.4%] or placebo twice daily (n = 12; mean age = 54 ± 4 yr; mean body mass index = 39.1 ± 1.6 kg/m(2); mean HbA1c = 8.5 ± 0.3%) for 12 wk. Fasting blood samples were obtained at 0, 3, 6, and 12 wk. Blood samples were also collected for up to 6 h after a single dose of exenatide (5 μg) or placebo. [RESULTS] Fasting blood glucose fell from 139 ± 17 to 110 ± 9 mg/dl, HbA1c from 8.6 ± 0.4 to 7.4 ± 0.5% (P < 0.05), and free fatty acids by 21 ± 5% from baseline (P < 0.05) with exenatide. There was no weight loss. There was a significant reduction in reactive oxygen species generation and nuclear factor-κB binding by 22 ± 9 and 26 ± 7%, respectively, and the mRNA expression of TNFα, IL-1β, JNK-1, TLR-2, TLR-4, and SOCS-3 in mononuclear cells by 31 ± 12, 22 ± 10, 20 ± 11, 22 ± 9, 16 ± 7, and 31 ± 10%, respectively (P < 0.05 for all) after 12 wk of exenatide. After a single injection of exenatide, there was a reduction by 20 ± 7% in free fatty acids, 19 ± 7% in reactive oxygen species generation, 39 ± 11% in nuclear factor-κB binding, 18 ± 9% in TNFα expression, 26 ± 7% in IL-1β expression, 18 ± 7% in JNK-1 expression, 24 ± 12% in TLR-4 expression, and 23 ± 11% in SOCS-3 expression (P < 0.05 for all). The plasma concentrations of monocyte chemoattractant protein-1, matrix metalloproteinase-9, serum amyloid A, and IL-6 were suppressed after 12 wk exenatide treatment by 15 ± 7, 20 ± 11, 16 ± 7, and 22 ± 12%, respectively (P < 0.05 for all). [CONCLUSIONS] Exenatide exerts a rapid antiinflammatory effect at the cellular and molecular level. This may contribute to a potentially beneficial antiatherogenic effect. This effect was independent of weight loss.