Rosiglitazone and heart failure (RECORD)

Overview

Komajda M et al.. Eur Heart J 2010 · PMID 20118174 · DOI 10.1093/eurheartj/ehp604

Heart failure doubled with rosiglitazone (2.6 extra cases per 1,000 patient-years), but overall heart deaths did not increase.

Does the study support the specific claim? Yes · Methodological quality: high

Original publication

Plain-language summary

Scientific in-depth review

1. Full reference

Authors

Komajda M, McMurray JJ, Beck-Nielsen H, Gomis R, Hanefeld M, Pocock SJ, Curtis PS, Jones NP, Home PD

Journal

Eur Heart J

Year

2010

Identifier

PMID 20118174 DOI 10.1093/eurheartj/ehp604 · PMC2848325

Verification

verified – PubMed E-utilities efetch, 2026-10-25

Retracted

no

Corrections

none

Extent checked

full text

Conflicts of interest and funding

RECORD sponsored by GlaxoSmithKline (rosiglitazone manufacturer); analysis performed by the sponsor's statistician; GSK paid open-access charges; two authors are GSK employees and stockholders; all other authors receive funding from pharmaceutical companies including GSK.

2. Research question

Heart failure risk with rosiglitazone

3. Study design and population

Study design

Randomised controlled trial – Secondary analysis of heart failure (HF) events in RECORD (NCT00379769), a multicentre, open-label, randomised non-inferiority trial with blinded endpoint adjudication; ITT plus per-protocol sensitivity analysis; stepwise multivariable model of HF predictors.

Population

Type 2 diabetes aged 40-75, BMI >25 kg/m2, HbA1c >7.0-<=9.0% on maximum-dose metformin or sulfonylurea monotherapy; excluded prior HF, recent major CV event, renal impairment, uncontrolled hypertension.

Country

364 centres in 25 countries in Europe and Australasia

Period

Recruitment April 2001 - April 2003; mean follow-up 5.5 years (12,338 vs 12,272 person-years).

Sample size

4,447 randomised and treated: rosiglitazone add-on 2,220 (1,117 on metformin, 1,103 on sulfonylurea); metformin+sulfonylurea 2,227.

Inclusion and exclusion criteria

As described in the abstract; details not available in abstract.

4. Methods and endpoints

Exposure / intervention

Add-on rosiglitazone 4-8 mg/day to metformin or sulfonylurea.

Comparator

Metformin plus sulfonylurea (glimepiride, gliclazide or glibenclamide).

Endpoints

Adjudicated HF hospitalisation or death (ESC-based predefined criteria, CEC blind to allocation); post-HF outcomes; baseline predictors of HF.

Statistical methods

Cox regression stratified by background medication (ITT); per-protocol sensitivity (on dual therapy + 30 days); stepwise multivariable regression of 30 pre-specified baseline variables; analysis by sponsor's statistician per steering-committee agreement.

5. Actual results

Results

HF events 61 (2.7%) vs 29 (1.3%), HR 2.10 (95% CI 1.35-3.27; P=0.0010); excess 2.6 (1.1-4.1) per 1,000 person-years. Per-protocol HR 1.91 (1.15-3.19). HF deaths 10 vs 2 (4 first events fatal). CV death or CV hospitalisation HR 0.99 (0.85-1.16); CV deaths 60 vs 71. Prior IHD: RR 2.16 (0.94-4.94); no prior IHD: 2.10 (1.25-3.51). Independent predictors: rosiglitazone HR 2.33 (1.46-3.70), age 1.10/year, micro/macroalbuminuria 3.09, BMI 1.11 per kg/m2, SBP 1.02 per mmHg. Loop diuretic use at 5 years 11.3% vs 7.2%.

Absolute risks

HF events 2.7% vs 1.3% over 5.5 years; excess 2.6 (1.1-4.1) HF events per 1,000 person-years.

Authors' conclusion

Confirms increased risk of HF events with rosiglitazone; it should not be continued in people developing symptomatic HF; close HF follow-up for elderly, obese, albuminuric or hypertensive patients; HF events with rosiglitazone are not all benign.

6. Strengths

• Randomised comparison with ~25,000 person-years
• HF adjudicated by blinded committee using predefined objective criteria
• Consistent ITT and per-protocol results

7. Weaknesses

• Open-label design (investigator-identified events may be influenced by knowledge of TZD fluid retention, though adjudication was blinded)
• Sponsor (GSK) analysed data; authors with extensive industry ties
• Few HF events (90) limit predictor and subgroup analyses; stepwise selection
• Only ~61%/51% completed on randomised treatment (88%/83% of person-time)
• Population excluded prior HF and renal impairment - absolute risk in higher-risk patients likely greater

8. Risk of bias

Orienting tool

RoB 2

Type of assessment

structured assessment based on the full text by a single reviewer (no formal double review)

Full formal assessment

no

Overall risk of bias

some concerns

Rationale

Single-reviewer RoB 2-oriented judgement from the full text; a formal assessment needs two independent reviewers and the protocol. Overall unchanged (some_concerns). Sponsor-analysed open-label trial, but bias would be expected to favour the sponsor's drug, so the HF signal is unlikely to be overestimated by sponsor influence.

Full-text review (25 October 2026): Inclusion/exclusion, arms by background therapy, event counts (61 vs 29), per-protocol HR 1.91, IHD subgroup, predictor model and loop-diuretic use added; COI/funding added (GSK sponsor, sponsor statistician, GSK employee authors); RoB 2 domains added, overall unchanged (some_concerns).

9. Assessment of the statistical interpretation

Clear doubling of relative HF risk with a modest absolute excess (2.6/1,000 person-years) and an excess of HF deaths (10 vs 2), while overall CV mortality/hospitalisation was not increased.

10. Relevance to the specific claim

Supports C-ODD-09.

11. Independent contrary findings

Same trial: no CV mortality increase.

12. Transferability to Germany and Europe

European trial.

13. Overall methodological judgement

Robust randomised evidence that rosiglitazone roughly doubles HF events (small absolute excess); no increase in overall CV death or hospitalisation in RECORD.

14. Open questions

None beyond those stated.

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 supporting the claim
• Rosiglitazone (Avandia) raised myocardial infarction risk by 43% (Nissen meta-analysis), leading to an FDA black-box warning, and RECORD confirmed a doubling of heart failure. – Studies supporting the claim

Use by the Tayebat archive

Note: according to the archive, the studies were compiled by the archive operators. Whether Dr. Elawady himself cited this study is not established.

What the archive says about the study

Narrative: RECORD confirmed a doubling of heart-failure events (HR 2.10).

A. What was actually studied?

Heart failure risk with rosiglitazone

B. What did the authors find?

HF HR 2.10 (1.35-3.27), 61 vs 29 events; excess 2.6 per 1,000 person-years; HF deaths 10 vs 2; CV death/hospitalisation HR 0.99; CV deaths 60 vs 71.

C. What conclusions do the authors themselves draw?

Confirms increased HF risk; stop rosiglitazone in symptomatic HF.

D. What does the archive derive from it?

Used as proof of thiazolidinedione harm.

E. Is this derivation scientifically justified?

Justified for the HF claim (doubling confirmed); the archive omits the small absolute excess and that overall CV mortality/hospitalisation was not increased.

Adequacy of the interpretation (by the archive)

essentially correct but incomplete