When I lecture medical students in our Graduate Entry Medical College in Swansea, one of the few things they know about the thiazolidinedione class of antidiabetic agents is that ‘rosiglitazone causes death through heart attacks’. They will, therefore, be puzzled to have read recent reports from the United States that “certain restrictions on the diabetes drug rosiglitazone (Avandia) are being removed because a new analysis…has found no increase in myocardial infarction risk”(1). What is the story behind this apparent U-turn?
It dates back to the publication in the New England Journal of Medicine (NEJM) in May 2007 of a meta-analysis that reported a significantly increased risk of myocardial infarction in patients receiving rosiglitazone versus alternative hypoglycaemic agents (2). None of the amalgamated studies set out to investigate the cardiovascular (CV) risk of the drug (hence the small number of events in the majority of the 42 studies included) and both the authors and the journal acknowledged methodological issues with the analysis. Indeed, looking at the raw data in the manuscript, there were actually more heart attacks in those patients receiving comparator drugs (72/11,633 [0.618%] versus 86/14,376 [0.598%] for rosiglitazone! However, there were other forces at play…
The NEJM was at the time strident in its criticism of the US regulatory authorities for lack of post-marketing safety follow-up of new drugs, arguing that the pharmaceutical industry should be mandated to deliver these analyses. One of the authors of the meta-analysis also had a track record in the area of CV drug safety. He had co-authored a publication, which led to the withdrawal of muraglitazar, a dual PPAR alpha and gamma anti-diabetic agent, shortly after the Federal Drug Agency (FDA) had given it approval (based on 35 vs. 9 CV events in the early phase studies) (3). He had also published a meta-analysis, which ultimately resulted in the withdrawal of Vioxx, a COX-2 inhibitor used for the treatment of arthritic pain, on the basis of increased CV risk (4). Interestingly, his original meta-analysis of rosiglitazone did not show a significant increase in CV or all-cause mortality and subsequent meta-analyses (including another one by the same author ) did not report increased death rates…
Clearly a study specifically designed to assess CV outcomes was needed and fortunately the RECORD study was already in progress, having been requested by the European regulators following the licensing of rosiglitazone in 2000. An interim analysis after a mean follow-up of 3.75 years was performed in response to the controversy and published in the NEJM in July 2007 (6). With the caveat that the study was under-powered due to the early analysis, the authors reported “no evidence of any increase in death from either cardiovascular causes or all causes”. RECORD continued to its pre-specified conclusion and the final analysis was published in the Lancet in 2009 (7). On this occasion, the authors concluded that “Although the data are inconclusive about any possible effect on myocardial infarction, rosiglitazone does not increase the risk of overall cardiovascular morbidity or mortality”.
Given this background, it might seem odd that in 2010 the FDA decided to impose severe restrictions upon the use of rosiglitazone, ruling that it should be ‘limited to existing users and new patients whose diabetes is not controlled by other treatments (8). In addition, the agency required a risk evaluation and mitigation strategy for those prescribing and using the drug.’ The British Medical Journal argued that the FDA had not gone far enough and in October 2010 the European Committee on Medicinal Products for Human Use agreed and recommended suspension of the marketing authorisations of rosiglitazone (Avandia, Avandamet) across the European Union (9,10).
The reason for these outcomes was that the regulators had been convinced that the data reported in the RECORD study were flawed i.e. that data published in a leading medical journal could not be believed. It is likely that the NEJM decision to publish in 2007 was also influenced by a belief that the pharmaceutical sponsor of rosiglitazone had not made public the CV concerns related to its drug. The reason for the about-turn in 2013 is that an extensive and independent adjudication of the RECORD data has shown “similar treatment effects of rosiglitazone compared with the original RECORD results” (11). In the USA, it is expected that rosiglitazone will once again have a similar licence to other diabetes drugs (i.e., for use along with diet and exercise to improve glucose control).
What then are the bottom lines following this six-year saga? In the UK (and the rest of Europe) it is highly unlikely that rosiglitazone will ever reappear on the market. The thiazolidinedione class as whole has been massively affected with pioglitazone new starts now limited (despite it being off-licence and therefore cheap) and development of new agents in this class has been largely suspended. More than 120,000 patients have been enrolled in placebo-controlled trials of new anti-diabetic drugs to prove CV safety at a cost of billions of dollars.
A good result all round?
2. Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes.
Nissen SE, Wolski K. N Engl J Med. 2007 Jun 14;356(24):2457-71. Epub 2007 May 21.
3. Effect of muraglitazar on death and major adverse cardiovascular events in patients with type 2 diabetes mellitus. Nissen SE, Wolski K, Topol EJ. JAMA. 2005 Nov 23;294(20):2581-6.
4. Risk of cardiovascular events associated with selective COX-2 inhibitors. Mukherjee D, Nissen SE, Topol EJ. JAMA. 2001 Aug 22-29;286(8):954-9.
5. An Updated Meta-analysis of Risk for Myocardial Infarction and Cardiovascular Mortality Steven E. Nissen, MD; Kathy Wolski. Arch Intern Med. 2010;170(14):1191-1201.
6. Rosiglitazone evaluated for cardiovascular outcomes–an interim analysis. Home PD, Pocock SJ, Beck-Nielsen H, Gomis R, Hanefeld M, Jones NP, Komajda M, McMurray JJ; RECORD Study Group. N Engl J Med. 2007 Jul 5;357(1):28-38.
7. Home PD, Pocock SJ, Beck-Nielsen H, Curtis PS, Gomis R, Hanefeld M, Jones NP, Komajda M, McMurray JJ; RECORD Study Team. Lancet. 2009 Jun 20;373(9681):2125-35. doi: 10.1016/S0140-6736(09)60953-3.
Professor Steve Bain