Heart and lungs

Painkiller heart risk examined

“Painkillers commonly used to treat arthritis, post-surgery pain and frozen shoulder, can increase the risk of dying from a heart attack or stroke,” The Daily Telegraph has reported.

The news is based on an extensive and well-conducted review of the use of NSAID drugs, a group of medicines used to relieve both pain and inflammation. The study drew upon the results of 31 trials featuring over 110,000 patients to assess how the drugs affected the risk of problems such as heart attack and stroke. Notably, there was an increased risk of heart attack with the drugs rofecoxib and lumiracoxib, and an increased risk of stroke with ibuprofen and diclofenac. However, the overall risk of these problems was still low among NSAID users, who were generally taking doses much higher than are used for typical pain relief.

These findings should be viewed in context, as the side effects reported in this research were already known and are already considered when prescribing patients NSAIDs. For example, rofecoxib was withdrawn from the UK market in 2004 and and lumiracoxib is not licensed for use in the UK. Certain other NSAIDs are only considered when patients have a low cardiovascular risk and are unable take the preferred alternative medication.

If you have any concerns about taking these medicines, you can consult your GP or pharmacist for further advice.

Where did the story come from?

This review was carried out by researchers from University of Bern, Switzerland. The study was funded by the Swiss National Science Foundation. The study was published in the peer-reviewed British Medical Journal.

The press has correctly represented the findings of this review, though not all news sources have made it clear that prior to this study it was already known that NSAIDs and of cyclo-oxygenase-2 selective (COX-2) inhibitors carried a risk of adverse cardiovascular effects. The wide credibility intervals quoted for stroke risk following ibuprofen use also suggests that the extent of any risk increase is uncertain: it may, for example, be less than the trebled risk quoted by the Daily Mail.

What kind of research was this?

This was a systematic review and meta-analysis aiming to combine the findings from randomised controlled trials that had examined the association between use of non-steroidal anti-inflammatory drugs (NSAIDs) and cardiovascular safety. NSAIDs are a group of medicines that are used to both treat pain and to reduce inflammation and swelling; two properties that have led the drugs to be key in the management of arthritis. Ibruprofen is the most widely used NSAID drug, although there is a range of other NSAIDs that work via different biological mechanisms.

A well-conducted systematic review that searches all relevant literature databases to identify all studies relevant to the question is the best way of examining the effect of an intervention on a particular outcome. However, all reviews carry some limitation due to the variability in quality, methods, outcomes and follow-up of the individual trials that they include.

When the outcome of interest is an adverse effect, as it was in this review, it must also be noted that this may not have been the primary outcome that the individual trial had been designed to investigate. For example, a trial investigating the use of NSAIDs to ease arthritis will most likely be investigating its effect on pain rather than its effect on heart attack or strokes.

What did the research involve?

This review involved a search of numerous medical databases, in addition to a search of clinical trial registers, conference proceedings, the Food and Drug Administration (FDA) website and reference lists of obtained articles. The researchers were interested in large randomised controlled trials (with at least 100 patient years of follow-up [e.g. 50 patients followed for two years]) that had compared any NSAID against paracetamol, an inactive placebo or another NSAID for the treatment of any medical condition with the exception of cancer.

The primary outcome that the researchers were interested in was the effect upon fatal or non-fatal heart attack. Other secondary outcomes of interest were:

  • fatal or non-fatal stroke – both ischaemic stroke (due to a clot) and haemorrhagic stroke (due to a bleed)
  • death due to a cardiovascular cause
  • death due to any other cause
  • the combined risk of non-fatal heart attack, non-fatal stroke or heart attack

The researchers combined these trials in a network meta-analysis. In a standard meta-analysis, the researchers would combine all trials that had compared the same NSAID with the same comparator (e.g. all trials directly comparing ibuprofen to paracetamol). A network analysis differs as it combines the results of multiple studies assessing different drug combinations, allowing comparisons to be made across the trials. For example, if one trial has compared diclofenac with ibuprofen and another has compared ibuprofen with paracetamol, the effect of diclofenac against paracetamol may be inferred even though they have not been directly compared.

The precision of the results of a network meta-analysis are quoted as credibility intervals. These are different to the confidence intervals usually quoted in studies, but can be interpreted similarly.

What were the basic results?

The researchers identified 31 relevant trials assessing 116,429 patients covering over 115,000 patient years of follow-up.

The trials assessed the use of the NSAIDs ibuprofen, diclofenac, naproxen, celecoxib, etoricoxib, rofecoxib and lumiracoxib, as well as placebo (dummy) drugs. The findings of the review are extensive, with the main ones being that:

  • Rofecoxib was associated with an almost doubled risk of heart attack compared with placebo (rate ratio 2.12, 95% credibility interval [CrI] 1.26 to 3.56). This was the highest risk association of all of the drugs tested.
  • Lumiracoxib also gave an almost doubled risk of heart attack, but the credibility intervals are wider and non-significant (rate ratio 2.00, 95% CrI 0.71 to 6.21).
  • Ibuprofen was associated with the highest risk of stroke – almost three-and-a-half times the risk compared with placebo – though again this was of borderline significance and with wide credibility intervals (rate ratio 3.36, 95% CrI 1.00 to 11.6).
  • Diclofenac was also associated with an almost-trebled risk of stroke compared with placebo (rate ratio 2.86, 95% Cr I 1.09 to 8.36).
  • Etoricoxib and diclofenac were associated with the highest risk of cardiovascular death found with (Etoricoxib rate ratio 4.07, 95% CI 1.23 to 15.7) and (Diclofenac rate ratio 3.98, 95% CrI 1.48 to 12.7).
  • Naproxen did not appear to have any significant associations with any of the cardiovascular outcomes.

How did the researchers interpret the results?

The researchers conclude that “little evidence exists to suggest that any of the investigated drugs are safe in cardiovascular terms” They say that cardiovascular risk needs to be taken into consideration when prescribing any NSAID. Naproxen appeared to be the “least harmful” drug, they add.


This was an extensive and well-conducted review that has combined the results of 31 trials in order to further analyse the safety risks of NSAIDs.

The findings of the review should be interpreted in light of the known risks of certain NSAIDS and the restrictions currently placed upon them:

  • Rofecoxib is a particular type of NSAID known as a cyclo-oxygenase-2 selective (COX-2) inhibitor. The drug was withdrawn from the UK market in 2004 due to concerns over its cardiovascular safety.
  • Lumiracoxib is not licensed for use in the UK.
  • The COX-2 inhibitors that are currently licensed in this country – cyclocelecoxib and etoricoxib – are already recognised to increase the risk of heart attack and stroke. It is advised that these drugs are only used when there are contraindications to the use of standard ‘non-selective’ NSAIDs such as ibruprofen (e.g. for patients at a particularly high risk of developing gastroduodenal ulceration or bleeding). Even then regulatory advice is that they are only to be used in an individual considered to have a low cardiovascular risk.
  • The non-selective NSAIDs ibuprofen and diclofenac are also recognised togive a small increase in the risk of blood clots, even in a person with no cardiovascular risk factors. Low doses of ibuprofen and naproxen were already considered to have much lower risk.
  • Systematic reviews are inherently limited by the design and quality of the individual studies that they include. When considering the quality of the strengths and limitations of this particular review, the authors themselves highlight that not all marketed NSAIDs have been considered, and that only published safety data was available, while some relevant data may have been unpublished.
  • The review was also unable to assess fully the effects of dosing and prescribing regimen and of short-term compared with medium- and long-term use.

NSAID drugs are commonly used among the general public to treat pain and inflammation, both as a  prescribed treatment and when bought over-the-counter. These drugs are known to carry risk of gastrointestinal irritation and bleeding, particularly in the elderly. They can also cause other sensitivity reactions, including aggravating asthma.

While some news coverage might suggest that a new cardiovascular risk from NSAID use has been identified, the risks discussed in this study were known prior to this research. This review has helped to gather the evidence to better quantify the size of this risk and highlights the need for the potential cardiovascular risks from NSAIDs must be taken into account whenever the drugs are used.

Current UK prescribing advice already recommends that the lowest effective dose of NSAID or COX-2 inhibitors are prescribed for the shortest period needed to control symptoms and that the need for long-term treatment should be reviewed periodically. This review does not change this advice.

NHS Attribution