NPC Archive Item: NSAIDs and cardiovascular risks – more evidence, but this shouldn’t be new news

NOTE – This is an archive post from the NPC and has not been updated since first publication. Therefore, some hyperlinks may no longer be working.
MeReC Rapid Review NPC Logo

1st February 2011

The conclusions of a network meta-analysis of randomised trials broadly confirm previous research showing an increased risk of cardiovascular (CV) events with many NSAIDs. Naproxen was associated with the lowest CV risk overall, and the roughly fourfold increased risk of death from CV causes seen with both diclofenac and etoricoxib are of concern. Interpretation of the results for ibuprofen is hampered by the absence of stratification by dose, which other work suggests is important.

Level of evidence:
Level 1 (good quality patient-oriented evidence) according to the SORT criteria.

Health professionals should assess both the CV and gastrointestinal (GI) risk of each patient individually and carefully consider the balance between benefit and risk before starting treatment with any NSAID. However, this study re-emphasises concerns around the CV safety of diclofenac and etoricoxib in particular. Despite encouraging trends in NSAID prescribing over recent years, diclofenac is still the most widely prescribed of all the NSAIDs (1.4million items, 35% of all NSAID prescriptions in England, September to November 2010). Some areas have been very effective in reducing diclofenac prescribing, whereas in others a change in prescribing in line with this important safety data has been slower and of a smaller magnitude. Etoricoxib is less widely used (115,000 items, 3% of all NSAID prescriptions in England, September to November 2010) but is associated with additional specific concerns regarding its effects on blood pressure particularly at high doses.

Prescribers should continue to follow our previous recommendations given in MeReC Extra 30 and discussed in MeReC rapid Review 1597. Prescribing managers should review local prescribing trends for NSAIDs as suggested in the document Key therapeutic topics 2010/11 – Medicines management options for local implementation, produced by the NPC as part of the QIPP programme.

What is the background to this?
Following the worldwide withdrawal of rofecoxib in 2004, and a European-wide review, the CV risk associated with coxibs was well recognised. In June 2006, a meta-analysis of randomised controlled trials (RCTs) by Kearney, et al raised concerns about the CV safety of traditional (non-COX-selective) NSAIDs. It suggested that high doses of diclofenac and ibuprofen, but not naproxen, were associated with an increased risk of CV events. In October 2006, a review of the CV safety of selective and non-selective NSAIDs, reported by the CHM, identified that diclofenac (especially 150mg per day) was associated with a small increased thrombotic risk similar to that of etoricoxib and possibly other coxibs. High-dose ibuprofen (e.g. 2400mg per day) also appeared to be associated with a small increased thrombotic risk, whereas low-dose ibuprofen (e.g. 1200mg per day or less) and naproxen (1000mg per day) were not.

Two epidemiological studies, reviewed in the February 2009 edition of Drug Safety Update, provided further evidence that a thrombotic CV risk may apply to all NSAID users, regardless of their baseline risk, with the greatest concern being for chronic users of high doses (in particular coxibs and diclofenac). In 2010 a large cohort study found that diclofenac and rofecoxib were associated with a similar increase in CV mortality and morbidity in healthy individuals. Naproxen was found to have a safer CV risk profile.

The authors of this study also examined the CV risks of NSAIDs via a meta-analysis. They conducted an exhaustive search for all RCTs, which compared any NSAID with other NSAID(s), paracetamol or placebo, with a follow up of at least 100 patient years. They then conducted a network meta-analysis. This is a potentially powerful tool which allows indirect comparisons of two treatments (A vs C on the basis of A vs B and B vs C), but which has some limitations, which are discussed briefly below.

What does this study claim?
The authors conclude “Although uncertainty remains, little evidence exists to suggest that any of the investigated drugs are safe in cardiovascular terms. Naproxen seemed least harmful. Cardiovascular risk needs to be taken into account when prescribing any non-steroidal anti-inflammatory drug”.

The primary outcome was myocardial infarction (MI). Secondary outcomes included stroke, death from CV disease, death from any cause and a composite of non-fatal MI, non-fatal stroke, or cardiovascular death. Patients were allocated to naproxen, ibuprofen, diclofenac, celecoxib, etoricoxib, rofecoxib, lumiracoxib, or placebo. The authors calculated the rate ratio (ratio of events per patient year) and the credibility ratio (analogous to the 95% confidence interval).

Among other findings, etoricoxib (rate ratio 4.07, credibility interval 1.23 to 15.7) and diclofenac (rate ratio 3.98, credibility interval 1.48 to 12.7) were associated with the highest risk of death from CV causes compared with placebo. Rofecoxib was associated with the highest risk of MI (rate ratio 2.12, 95% credibility interval 1.26 to 3.56). Ibuprofen was associated with the highest risk of stroke (rate ratio 3.36, credibility interval 1.00 to 11.6), followed by diclofenac (rate ratio 2.86, 1.09 to 8.36. Full results are given in the study details section below.

So what?
Despite some limitations, this study broadly confirms the conclusions of previous studies. Taken together with these, there is clear evidence of an increased CV risk associated with most NSAIDs, both traditional (non-COX-selective) and selective (the coxibs), but naproxen is associated with either no increased risk or a smaller increased risk than most other NSAIDs. Interpretation of the results of this study for ibuprofen are hampered by the absence of stratification by dose: other work (discussed above) suggests that ibuprofen at 1200mg per day or less is not associated with an increased CV risk, whereas at higher doses it carries a CV risk broadly similar to other NSAIDs.

As the related editorial states, these data should prompt practitioners to review their prescribing. However, the issues, findings and recommendations of this study should not be ’new’ news to practitioners. In the second half of 2007 the NPC initiated a number of activities to raise awareness of these safety issues and encourage review of NSAID prescribing, including a programme of NPC-funded local workshops supported by educational materials and MeReC Extra 30 (November 2007). Since then there has been a significant decline in the prescribing rate of diclofenac and increase in the relative use of naproxen, as discussed in MeReC stop Press 293.

Nevertheless, diclofenac remains the most widely prescribed NSAID, which may expose a substantial number of individuals to the risk of CV adverse events, as we have discussed previously. The absolute risk will vary according to an individual patient’s risk factors; however, across the whole population this is about three additional CV events per 1000 patients per year compared with placebo. To put that into perspective, the expected reduction in CV events produced by simvastatin 40mg in people at a baseline 20% 10-year risk of CV events is round about five per 1000 patients per year.

As we have pointed out in our document Key therapeutic topics 2010/11 – Medicines management options for local implementation, produced as part of the QIPP programme, additional concerns have also been highlighted specifically regarding etoricoxib’s effects on blood pressure. In 2005 the MHRA advised that etoricoxib may be associated with more frequent and severe effects on blood pressure than some other COX-2 inhibitors and NSAIDs, particularly at high doses. Hypertension should be controlled before treatment with etoricoxib is commenced and special attention should be paid to blood pressure monitoring during treatment with it. The manufacturers of etoricoxib advise that blood pressure should be monitored within two weeks after initiation of treatment and periodically thereafter. If blood pressure rises significantly, alternative treatment should be considered. The NICE osteoarthritis guideline recommends against using the 60mg dose first-line on the basis of cost-effectiveness. In view of the amount of etoricoxib prescribing nationally, local review of etoricoxib prescribing would seem to be justified.

Are there any limitations to the results?
Network meta-analysis is a potentially powerful tool, but also has some potential limitations and caveats, as discussed in the editorial related to this paper. It allows indirect comparisons to be made through a chain of direct comparisons (for example, A vs C or A vs D, even though no trials have directly compared A and C or D, through A vs B, B vs C and C vs D). However, certain assumptions about homogeneity of study populations and outcomes measured are necessary for these estimates to be valid. Those assumptions can be checked and uncertainty can be reflected in the results, but the limited number of comparisons for many NSAIDs limits the precision with which the estimates of comparative risk can be made. In addition, there were some limitations to the original data which the authors discuss. Among other things, large scale RCTs were not available at all for some NSAIDs (including meloxicam, which is still widely used); and Merck, the manufacturer of rofecoxib and etoricoxib, was not willing to provide unpublished safety data on these drugs, thus limiting the analysis. Nevertheless, the overall message from this study is in keeping with what we already know.

Study details
Trelle S et al. Cardiovascular safety of non-steroidal anti-inflammatory drugs: network meta-analysis BMJ 2011; 342:c7086

Design Network meta-analysis of published and unpublished large scale RCTs comparing any NSAID with other NSAIDs, paracetamol or placebo.

Patients 116,429 patients from 31 trials with 117,218 patient years of follow-up


Outcomes and results Rate ratio compared with placebo, with 95% credible interval (analogous to 95% confidence interval): credible intervals which do not include 1.00 are highlighted


Myocardial infarction
Naproxen 0.82 (0.37 to 1.67)
Ibuprofen 1.61 (0.50 to 5.77)
Diclofenac 0.82 (0.29 to 2.20)
Celecoxib 1.35 (0.71 to 2.72)
Etoricoxib 0.75 (0.23 to 2.39)
Rofecoxib 2.12 (1.26 to 3.56)

Lumiracoxib 2.00 (0.71 to 6.21)

Naproxen 1.76 (0.91 to 3.33)
Ibuprofen 3.36 (1.00 to 11.60)
Diclofenac 2.86 (1.09 to 8.36
Celecoxib 1.12 (0.60 to 2.06)
Etoricoxib 2.67 (0.82 to 8.72)
Rofecoxib 1.07 (0.60 to 1.82)
Lumiracoxib 2.81 (1.05 to 7.48)

Cardiovascular death
Naproxen 0.98 (0.41 to 2.37)
Ibuprofen 2.39 (0.69 to 8.62)
3.98 (1.48 to 12.70)
Celecoxib 2.07 (0.98 to 4.55)
Etoricoxib 4.07 (1.23 to 15.70)

Rofecoxib 1.58 (0.88 to 2.84)
Lumiracoxib 1.89 (0.64 to 7.09)

Death from any cause
Naproxen 1.23 (0.71 to 2.12)
Ibuprofen 1.77 (0.73 to 4.30)
Diclofenac 2.31 (1.00 to 4.95)
Celecoxib 1.50 (0.96 to 2.54)
Etoricoxib 2.29 (0.94 to 5.71)
Rofecoxib 1.56 (1.04 to 2.23)

Lumiracoxib 1.75 (0.78 to 4.17)

Composite of non-fatal myocardial infarction, non-fatal stroke, or cardiovascular death
Naproxen 1.22 (0.78 to 1.93)
Ibuprofen 2.26 (1.11 to 4.89)

Diclofenac 1.60 (0.85 to 2.99)
Celecoxib 1.43 (0.94 to 2.16)
Etoricoxib 1.53 (0.74 to 3.17)
Rofecoxib 1.44 (1.00 to 1.99)
Lumiracoxib 2.04 (1.13 to 4.24)

Some individual authors and their organisations have received grants or funding from the Swiss National Science Foundation, which had no role in the study design, data collection, data analysis, data interpretation, writing of the manuscript, or decision to submit the manuscript.

More information on NSAIDs can be found on the musculoskeletal pain floor of NPC

Include if appropriate

Please comment on this blog in the NPC discussion rooms, or using our feedback form.

Make sure you are signed up to NPC Email updates — the free email alerting system that keeps you up to date with the NPC news and outputs relevant to you.

Related PJ Online article: Cardiovascular risk of NSAIDs should not be disregarded, say researchers