Intended for healthcare professionals

Editor's Choice

Rosiglitazone: a cautionary tale

BMJ 2010; 341 doi: (Published 07 September 2010) Cite this as: BMJ 2010;341:c4896

This article has a correction. Please see:

  1. Fiona Godlee, editor, BMJ
  1. fgodlee{at}

    This week a BMJ investigation suggests that the diabetes drug rosiglitazone should never have been licensed and should now be withdrawn (doi:10.1136/bmj.c4848).

    The tale of rosiglitazone is a cautionary one from which we must hope the main parties will learn for the future. Hailed as a much needed new approach for patients with type 2 diabetes, the drug was licensed 10 years ago with only limited evidence of its effectiveness and concerns over its safety. While allowing the drug onto the market, the regulators asked the manufacturer GlaxoSmithKline to do additional trials. This the company did while marketing the drug around the world. Millions of prescriptions later, the results of the open label RECORD trial are hotly disputed. GlaxoSmithKline says it shows the drug is safe and the European Medicines Agency seems to have taken this conclusion on trust. Officers at the Food and Drug Administration, however, prompted by concerns about GlaxoSmithKline’s conduct (BMJ 2010; 340:c1848, doi:10.1136/bmj.c1848) and aided by the FDA’s requirement for individual patient data, decided earlier this year to take a closer look. They uncovered errors in the way the trial was done that systematically favoured the company’s drug.

    Meanwhile the BMJ has discovered that the UK’s Committee on Human Medicines advised the UK regulator, the Medicines and Healthcare products Regulatory Agency, in August that the risks of rosiglitazone outweighed the benefits and that it had no place on the UK market. But doctors received only a muted version of this guidance in a letter in August suggesting merely that they seek alternatives to rosiglitazone. The committee’s much starker conclusion would not have been known until later this month when the EMA meets to decide rosiglitazone’s fate, and might not have been known at all if the EMA were to decide to keep the drug on the market.

    So what should happen? Our commentators (doi:10.1136/bmj.c4805, doi:10.1136/bmj.c4812) and others quoted in the investigation are clear on a number of things, none of which are being said for the first time (BMJ 2009;338:b1025, doi:10.1136/bmj.b1025). Europe’s regulators should be much more transparent. They should require a higher quality of evidence, including proof that new drugs are better than existing drugs before being licensed. And if they do ask the manufacturer to undertake post-marketing trials, they must do a better job of overseeing the way these trials are designed and done.

    Will this put extra strain on what many see as an already overregulated pharmaceutical industry? Certainly it will. So the patent for new drugs should be extended, from the current 20 years to perhaps 25 or 30. We all need the pharmaceutical sector to flourish and innovate. We should also seek to modify the increasingly destructive relationship between industry and the public. This would require concessions on both sides: far greater transparency from industry and the regulators, including access to raw data and funding for independent trials; and greater understanding from the public that there is no such thing as a completely safe drug.


    Cite this as: BMJ 2010;341:c4896


    View Abstract