Meta-analysisindependent cross-checkabstract checked★ key study I-ODD-117 in-depth review
Re-analysis of rosiglitazone trials: heart failure confirmed, heart attacks uncertain
Updating insights into rosiglitazone and cardiovascular risk through shared data: individual patient and summary level meta-analyses.
Overview
Wallach JD et al.. BMJ 2020 · PMID 32024657 · DOI 10.1136/bmj.l7078
HF OR 1.54; MI OR 1.17 (not significant).
Does the study support the specific claim? Partly · Methodological quality: moderate
Plain-language summary
Study explained simply
- What did the researchers want to find out?
Re-examine rosiglitazone cardiovascular risk with patient-level data.
- How was the study done?
IPD and summary meta-analyses.
- How many people, animals or samples were studied?
21,156 patients with IPD.
- What was actually found?
Higher composite risk driven by heart failure; MI uncertain.
- How large was the observed effect?
Composite OR 1.33; HF OR 1.54; MI OR 1.17.
- What do the results mean in everyday life?
The heart-failure risk is real; the heart-attack risk remains unproven.
- What are the study's limitations?
Analytic sensitivity; correction.
- What does it say about Elawady's specific claim?
Partly supports the claim; 43% MI figure is not robust.
- Do the results apply to Germany and Europe?
Limited (drug withdrawn).
- What can explicitly not be concluded from this study?
A precise MI risk.
Scientific in-depth review
1. Full reference
- Authors
- Wallach JD et al.
- Journal
- BMJ
- Year
- 2020
- Identifier
- PMID 32024657 DOI 10.1136/bmj.l7078
- Verification
- verified – PubMed efetch 2026-10-26; Crossref works API 2026-10-26
- Retracted
- no
- Corrections
Correction published (BMJ 2021;373:n1302) acknowledging discrepancies in fatal adverse event counts for four trials (ADOPT, AVA102670, AVA102672, AVD100521); main MI/heart-failure conclusions not stated to change (correction text read).
- Extent checked
- abstract only
- Conflicts of interest and funding
Not stated in abstract.
2. Research question
Does rosiglitazone increase cardiovascular risk when individual patient data are used?
3. Study design and population
- Study design
- Meta-analysis – Systematic review with individual-patient-level and summary-level meta-analyses of RCTs
- Population
Adults in rosiglitazone RCTs >=24 weeks
- Country
International
- Period
Searches to January 2019
- Sample size
33 trials with IPD (21,156 patients); 103-136 trials for summary analyses
- Inclusion and exclusion criteria
Randomised phase II-IV trials of rosiglitazone vs any control >=24 weeks
4. Methods and endpoints
- Exposure / intervention
Rosiglitazone
- Comparator
Any control
- Endpoints
Composite of MI, HF, CV death, non-CV death; components
- Statistical methods
Multiple meta-analyses with continuity corrections; random effects; OR and RR
5. Actual results
- Results
IPD composite OR 1.33 (1.09-1.61) (274/11,837 vs 219/9,319). MI OR 1.17 (0.92-1.51); HF OR 1.54 (1.14-2.09); CV death 1.15 (0.55-2.41). With summary-level data added: MI OR 1.09 (0.88-1.35), CV death 1.12 (0.72-1.74). IPD identified more MIs than summary data in 26 trials.
- Absolute risks
Crude event proportions 2.31% vs 2.35% (not trial-stratified; stats/C-ODD-09.py).
- Authors' conclusion
Rosiglitazone is associated with increased cardiovascular risk, especially heart failure; MI evidence varies and is attenuated with summary data.
6. Strengths
- Individual patient data
- Transparency of analytic choices
7. Weaknesses
- Sensitivity to continuity corrections and data source
- Correction for fatal event counts
8. Risk of bias
- Orienting tool
- AMSTAR 2
- Type of assessment
- simplified assessment (abstract only)
- Full formal assessment
- no
- Overall risk of bias
- some concerns
- Rationale
Results depend on analytic choices.
9. Assessment of the statistical interpretation
CIs log-symmetric. The crude pooled proportions are nearly equal (2.31% vs 2.35%), so the 33% composite increase emerges only in trial-stratified analysis with continuity corrections; this sensitivity supports 'MI not settled'.
10. Relevance to the specific claim
Confirms heart-failure harm and shows the MI increase is uncertain (OR 1.17, CI including 1).
11. Independent contrary findings
RECORD (I-ODD-13) MI HR 1.14 (0.80-1.63).
12. Transferability to Germany and Europe
Limited relevance to current EU practice (see existing review of C-ODD-09); market status not separately verified here.
13. Overall methodological judgement
Supports the current verdict: HF harm firm; MI harm uncertain.
| Multidimensional assessment | |
|---|---|
| Authenticity of the source | verified |
| Methodological quality | moderate |
| Risk of bias | some concerns |
| Precision of results | moderate |
| Directness | direct |
| Reproducibility | partly replicated |
| Consistency with the overall evidence | mixed |
| Regional transferability | high |
| Adequacy of the interpretation (by the archive) | not applicable |
14. Open questions
None practical (drug largely withdrawn).