Meta-analysisindependent cross-checkfull text checked★ key study I-ODD-107 in-depth review
816 trials: what each diabetes drug class does and does not do
Benefits and harms of drug treatment for type 2 diabetes: systematic review and network meta-analysis of randomised controlled trials.
Overview
Shi Q et al.. BMJ 2023 · PMID 37024129 · DOI 10.1136/bmj-2022-074068
SGLT2 inhibitors and GLP-1 agonists reduce death by about 12%; each class has specific side effects.
Does the study support the specific claim? Partly · Methodological quality: high
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Compare benefits and harms of all diabetes drug classes.
- How was the study done?
Network meta-analysis of 816 RCTs with GRADE.
- How many people, animals or samples were studied?
471,038 patients.
- What was actually found?
Mortality benefit for SGLT2i and GLP-1 RA; class-specific harms such as DKA and genital infections (SGLT2i), hypoglycaemia (sulfonylureas, insulin), heart failure (TZDs).
- How large was the observed effect?
Death OR 0.88; DKA OR 2.07; amputation OR 1.27.
- What do the results mean in everyday life?
No drug is free of side effects; for several classes the benefits clearly outweigh them in people at risk.
- What are the study's limitations?
Low certainty for older drugs.
- What does it say about Elawady's specific claim?
Supports 'every class has harms'; refutes 'net harmful'.
- Do the results apply to Germany and Europe?
High.
- What can explicitly not be concluded from this study?
Individual treatment choice.
Scientific in-depth review
1. Full reference
- Authors
- Shi Q et al.
- Journal
- BMJ
- Year
- 2023
- Identifier
- PMID 37024129 DOI 10.1136/bmj-2022-074068
- Verification
- verified – PubMed efetch 2026-10-26; Crossref works API 2026-10-26
- Retracted
- no
- Corrections
- none
- Extent checked
- full text
- Conflicts of interest and funding
Sichuan Science and Technology Programme and West China Hospital; funders had no role.
2. Research question
What are the benefits and harms of each drug class for adults with type 2 diabetes?
3. Study design and population
- Study design
- Meta-analysis – Systematic review and network meta-analysis of RCTs with GRADE (BMJ Rapid Recommendations evidence base)
- Population
Adults with type 2 diabetes in RCTs >=24 weeks
- Country
International
- Period
Searches to 14 October 2022
- Sample size
816 trials, 471,038 patients, 13 drug classes
- Inclusion and exclusion criteria
RCTs >=24 weeks comparing drugs of interest; non-English trials excluded
4. Methods and endpoints
- Exposure / intervention
SGLT2 inhibitors, GLP-1 RAs, DPP-4 inhibitors, thiazolidinediones, sulfonylureas, metformin, insulins, tirzepatide, finerenone, others
- Comparator
Standard treatment and each other
- Endpoints
All-cause and CV death, MI, stroke, HF hospitalisation, ESKD, amputation, severe hypoglycaemia, DKA, genital infection, GI events, weight, quality of life
- Statistical methods
Frequentist random-effects network meta-analysis; GRADE; absolute effects by baseline risk
5. Actual results
- Results
All-cause death: SGLT2i OR 0.88 (0.83-0.94) and GLP-1 RA 0.88 (0.82-0.93), high certainty; metformin possibly reduces (0.84, 0.67-1.04, low certainty); sulfonylureas possibly increase all-cause mortality (low certainty). HF hospitalisation: SGLT2i OR 0.66 (0.60-0.73); thiazolidinediones probably increase (1.54, 1.27-1.88). Severe hypoglycaemia: sulfonylureas OR 5.22 (3.88-7.01), basal insulin 2.38; SGLT2i and GLP-1 RA no increase. DKA: SGLT2i OR 2.07 (1.44-2.98), high certainty, about 2 more per 1,000 over 5 years. Amputation: SGLT2i OR 1.27 (1.01-1.61), moderate certainty, about 3 more per 1,000 over 5 years. Genital infection SGLT2i OR 3.30 (2.88-3.78).
- Absolute risks
SGLT2i: +2 DKA and +3 amputations per 1,000 over 5 years; deaths avoided depend on baseline risk (MAGIC decision aid).
- Authors' conclusion
Confirms substantial benefits of SGLT2 inhibitors and GLP-1 RAs in reducing cardiovascular and kidney outcomes and death; harms largely class-specific.
6. Strengths
- Largest network of RCTs
- GRADE certainty for each outcome
- Absolute effects
- Independent (non-industry) funding
7. Weaknesses
- Low certainty for older drugs (metformin, sulfonylureas)
- Inconsistency for amputation outcome
- Non-English trials excluded
8. Risk of bias
- Orienting tool
- AMSTAR 2
- Type of assessment
- simplified assessment (abstract only)
- Full formal assessment
- no
- Overall risk of bias
- low
- Rationale
Pre-registered (PROSPERO CRD42022325948); duplicate screening; risk of bias per trial.
9. Assessment of the statistical interpretation
Ratio CIs log-symmetric (stats/C-ODD-02.py, C-ODD-06.py). Updates and largely supersedes Palmer 2021 (BMJ), which had judged no DKA increase; the larger 2023 data set shows a high-certainty doubling of DKA. Overlaps with SMART-C (I-ODD-106) for SGLT2 trials; not independent confirmation.
10. Relevance to the specific claim
The most comprehensive benefit-harm comparison: confirms each class has specific harms (C-ODD-05) but that SGLT2i and GLP-1 RA reduce death (contradicting C-ODD-06).
11. Independent contrary findings
None of comparable scope.
12. Transferability to Germany and Europe
High.
13. Overall methodological judgement
Pivotal: harms exist and are drug-class specific; two classes reduce mortality with high certainty; sulfonylureas and thiazolidinediones have less favourable profiles.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | high |
| Risk of bias | low |
| Precision of results | high |
| Directness | direct |
| Reproducibility | replicated |
| Consistency with the overall evidence | consistent |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
Combination therapy effects; long-term data for tirzepatide.
15. Linked fact checks
- Every major class of glucose-lowering drug (metformin, sulfonylureas, thiazolidinediones, DPP-4 inhibitors, SGLT2 inhibitors, GLP-1 agonists) carries a documented signal of serious harm in the peer-reviewed literature. – Studies with mixed or unclear results
- Glucose-lowering drugs are net harmful and belong among the 'forbidden' items; dapagliflozin (Forxiga) is among the 'dirtiest' drugs and kills the brain; leg amputations happen only to people taking diabetes drugs and insulin. – Studies contradicting or limiting the claim
- SGLT2 inhibitors such as dapagliflozin (Forxiga) can cause ketoacidosis, sometimes with normal or low blood glucose. – Studies supporting the claim
- Sulfonylureas cause more myocardial infarctions, strokes and deaths than metformin (21% higher in 253,690 veterans). – Studies with mixed or unclear results