NSAIDs and cardiac arrhythmias

I am beginning to understand Twitter a bit.

I was intending to go and write about the headlines earlier this week, about NSAIDs and the risk of cardiac arrhythmias, which made several front pages earlier this week.

The Metro put it like this ” popular painkillers such as ibuprofen can put patients at risk of heart disease and even death, it has been claimed. Taking the drugs daily can increase the chances of an irregular heart rhythm – known as atrial fibrillation – by up to 70 per cent, a new study found. ” Sky News said much the same: “Taking common painkillers such as ibuprofen regularly could make you up to 70% more likely to suffer an irregular heartbeat, new research has claimed”.

Thanks to Twitter, I have found that the research paper has been written about here, at Letting Off Steam. The big problem with the media reports is, as usual, the concentration on relative risks and not absolute risks. All drugs have side effects; the question is always what to do about the balance of benefit and harm. We need good information to make good decisions. That means absolute risks, and not relative risks.

I can understand the fact that Jay has complained to the Metro about the way the figures were represented: the study itself in the discussion section ‘fesses us: the better explained risk is equivalent to approximately four extra cases per year of atrial fibrillation per 1000 new users of non-selective NSAIDS and seven extra cases per year of atrial fibrillation per 1000 new users of COX 2 inhibitors.”

Shall we blame the press? No. Here is the press release.

Common painkillers linked to irregular heart rhythm

Research: Non-steroidal anti-inflammatory drug use and risk of atrial fibrillation or flutter: population-based case-control study
Editorial: NSAIDs and atrial fibrillation

Commonly used painkillers to treat inflammation are linked to an increased risk of irregular heart rhythm (known as atrial fibrillation or flutter), concludes a study published on bmj.com today.

The drugs include non-selective non-steroidal anti-inflammatory drugs (NSAIDS) as well as new generation anti-inflammatory drugs, known as selective COX-2 inhibitors.

These drugs have already been linked to an increased risk of heart attacks and strokes, but no study has examined whether they increase the risk of atrial fibrillation – a condition which is associated with an increased long term risk of stroke, heart failure, and death.

So a team of researchers, led by Professor Henrik Toft Sørensen at Aarhus University Hospital in Denmark, used the Danish National Registry of Patients to identify 32,602 patients with a first diagnosis of atrial fibrillation or flutter between 1999 and 2008.

Each case was compared with 10 age and sex-matched control patients randomly selected from the Danish population.

Patients were classified as current or recent NSAID users. Current users were further classified as new users (first ever prescription within 60 days of diagnosis date) or long-term users.

The researchers found that use of NSAIDs or COX-2 inhibitors was associated with an increased risk of atrial fibrillation or flutter.

Compared with non users, the association was strongest for new users, with around 40% increased risk for non-selective NSAIDS and around 70% increased risk for COX-2 inhibitors. This is equivalent to approximately four extra cases of atrial fibrillation per year per 1000 new users of non-selective NSAIDS and seven extra cases of atrial fibrillation per 1000 new users of COX-2 inhibitors.

The risk appeared highest in older people, and patients with chronic kidney disease or rheumatoid arthritis were at particular risk when starting treatment with COX-2 inhibitors.

The authors conclude: “Our study thus adds evidence that atrial fibrillation or flutter need to be added to the cardiovascular risks under consideration when prescribing NSAIDs.”

This view is supported by an accompanying editorial by Professor Jerry Gurwitz from the University of Massachusetts Medical School in the US. He believes that NSAIDS should continue to be used very cautiously in older patients with a history of hypertension or heart failure …. regardless of whether an association between NSAIDs and atrial fibrillation actually exists.”

Was this press release fair, balanced and helpful to patients? Sure, it included absolute risks, but maybe it should have included no relative risk information – it was going to be predictably picked up by the press as a headline with no guarantee that they would follow with the absolute risk as well.

I think the aim of press releases should sometimes be to attract very little press attention.

3 Responses to “NSAIDs and cardiac arrhythmias”

  1. Trish Groves July 7, 2011 at 8:30 am #

    Thanks for picking up the BMJ press release as well as the paper. Might you send a Rpaid response to the paper?

    We’re working on our reporting of absolute risks, and trying to ensure they’re always expressed in both paper and press release. Not always possible for all studies eg network meta-analyses, though.

    See this, and its rapid responses:
    Misleading communication of risk
    Gerd Gigerenzer, Odette Wegwarth, Markus Feufel
    BMJ 2010;341:doi:10.1136/bmj.c4830 (Published 12 October 2010)

    Interesting suggestion about leaving the relative risks out of press releases; one for us to ponder.

  2. Idris Harding July 7, 2011 at 2:07 pm #

    I’m not sure that anyone’s really made a mistake here, other than assuming prior knowledge on the part of their readers. In some caes this is a reasonable assumption, in others not.

    It is reasonable to assume that the readership of an academic journal will be able to integrate various different risk assessment statistics into an overall appreciation of the risk pattern; maybe it is not so safe to assume that non-specialist journalists reading a press release will be able to do the same. It is very optimistic to think that the public reading a third hand version of the research in their daily broadsheet will be able to read between the lines sufficiently to reach valid conclusions and interpret the research in a way that is relevant to their own decision making, and that, surely is the point?

  3. Margaret McCartney
    margaretmccartney July 7, 2011 at 9:04 pm #

    thank you – yes Trish I shall do that.
    I agree, it’s not so much about a mistake but about how to get fair information into the public domain. The purpose of a press release is usually to attract attention, I think, though there are honorable exceptions. And of course attracting attention is often very justifiable. But when you are letting statistics go off into the world, maybe we should focus on NNT and absolute risks and leave the RR for the journals.

    It would be interesting to see, if the RR was left in the journal article only, how often it was then cited.