GLP-1 Receptor Agonist Therapy Is Associated With Lower Risk of Lower Extremity Amputation in Adults With Type 2 Diabetes and Foot Ulcer Without Coded Peripheral Arterial Disease: A Hypothesis-Generating Real-World Cohort Study
- Design
- Retrospective cohort · 42650 participants · Hard outcome
- Match to healthy normal-weight adults aged 55–75
- Different populationINDIRECT: all participants had type 2 diabetes and a diabetic foot ulcer.
- Could weight loss explain it?
- PossiblyPossibly: no weight or glycaemic mediation analysis is described, so weight and glycaemic change are not separated from a drug effect (the evidence model's default); matching on 22 covariates, background-therapy balancing and limb-severity adjustment address confounding, not mediation.
- Study tier
- Study tier 3LOW: a single retrospective database cohort, capped at LOW and described by the authors as hypothesis-generating.
- Assessment
- Version 4 · 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 database study matched adults with type 2 diabetes and a diabetic foot ulcer who started GLP-1 receptor agonists with similar people starting other glucose-lowering drugs. GLP-1 use was linked to a lower 5-year amputation risk (HR 0.60, 95% CI 0.54-0.67) and lower all-cause mortality (HR 0.48); the link was seen in people without coded peripheral arterial disease (HR 0.21) but not where that disease was coded (HR 0.97). The authors call the results hypothesis-generating; they show an association, not cause.
01Findings
What the study reported
- Drugs
- Glp-1 receptor agonists
- Comparator
- all non-GLP-1 RA glucose-lowering therapy (active comparator; insulin also examined)
- Dose
- various
- Primary outcome
- lower extremity amputation
- Effect
- 5-year amputation HR 0.60; HR 0.21 with no coded PAD and 0.97 with coded PAD (no association); all-cause mortality HR 0.48
- 95% confidence interval
- 95% CI 0.54-0.67 (overall amputation HR 0.60)
- P value
- p < 0.001 (ratio of hazard ratios, PAD effect modification)
- Follow-up
- 5-year outcome window
- Adverse events
- not reported in abstract
- Limitations
- Retrospective database study open to residual confounding; the authors describe the findings as hypothesis-generating, and the subgroup result relies on coded PAD diagnoses.
Who was studied
- Obesity status
- not reported in abstract
- Diabetes status
- type 2 diabetes required
- Cvd status
- mixed (coded peripheral arterial disease examined as an effect modifier)
- Ckd status
- not reported in abstract
- Baseline condition
- type 2 diabetes with diabetic foot ulcer
- Sample size
- 42650
- Inclusion exclusion
- new users; adults with type 2 diabetes and diabetic foot ulcer, propensity-matched 1:1 on 22 covariates
Study quality details
- Study design
- Retrospective cohort
- Sample size
- 42650
- Randomization
- no
- Blinding
- n/a
- Comparator
- all non-GLP-1 RA glucose-lowering therapy (active comparator; insulin also examined)
- Follow up duration
- 5-year outcome window
- Outcome type
- hard
- Replication
- not assessed
- Consistency with other evidence
- not assessed
- Population applicability
- INDIRECT
- Statistical precision
- CI reported: 95% CI 0.54-0.67 (overall amputation HR 0.60)
- Risk of bias
- not assessed
- Funding conflicts
- unclear
- Peer review status
- yes
02Funding
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.
04Source
The source, as retrieved
Abstract
[AIMS] To compare lower extremity amputation (LEA) risk in adults with Type 2 diabetes and diabetic foot ulcer (DFU) treated with glucagon-like peptide-1 receptor agonists (GLP-1 RAs) versus other glucose-lowering therapy, and to test whether coded peripheral arterial disease (PAD) modifies this association and replicates across comparator classes. [MATERIALS AND METHODS] This new-user, active-comparator retrospective cohort study used the TriNetX Global Collaborative Network (queried 11 July 2026). Adults with Type 2 diabetes and DFU were propensity-score matched 1:1 on 22 covariates, giving 21 325 pairs for the primary comparison. Effect modification by coded PAD (ICD-10 I70.2) was examined in independently matched subgroups. Six sensitivity analyses were pre-specified: 30- and 180-day landmarks, all-cause mortality, background-therapy balancing, limb-severity adjustment and a broadened PAD definition. [RESULTS] GLP-1 RA use was associated with lower 5-year LEA risk than all non-GLP-1 RA therapy (hazard ratio [HR]: 0.60, 95% CI: 0.54-0.67). The association was confined to the no-PAD stratum (HR: 0.21, 0.13-0.33) and absent where PAD was coded (HR: 0.97, 0.78-1.20); ratio of hazard ratios 0.21 (0.13-0.36; p < 0.001). All-cause mortality was also lower (HR: 0.48, 0.45-0.50). The insulin no-PAD estimate persisted under background-therapy balancing (0.14), limb-severity adjustment (0.16) and 30- and 180-day landmarks (0.28, 0.33). Broadening the PAD definition attenuated the no-PAD association without abolishing it (HR: 0.34, 0.17-0.65). [CONCLUSIONS] In Type 2 diabetes with DFU, GLP-1 RA use was associated with lower LEA risk, surviving six pre-specified sensitivity analyses. These observational findings are hypothesis-generating and require prospective confirmation.
Where this record came from
| Source | Retrieved | Identifier |
|---|---|---|
| pubmed | Oct 18, 2026 | 42875725 first ingestion |