Cardiovascular safety of NSAIDs: epidemiology and pharmacology

Eight years after the rofecoxib withdrawn, the EMA has published the conclusion of the third review on cardiovascular safety for traditional non-steroidal anti-inflammatory drugs.1

“Traditional NSAIDs” refers to NSAIDs developed before it had been discovered that their efficacy was linked to the inhibition of isoform-2 of the cyclooxygenase enzyme (COX-2), whilst their toxicity was due to the inhibition of the other isoform, COX1. Studies on COX-2 pushed pharmaceutical companies to produce NSAIDs selective for this isoform which were named coxib, presuming a diminished gastrointestinal toxicity. However this reduced risk was accompanied by a higher risk for thromboembolic events, understandable for coxib but less clear for traditional NSAIDs.
The EMA concluded that diclofenac cardiovascular risk is higher than the other traditional NSAIDs and is comparable with the coxib risk, defined as “other painkillers”.
In Europe, around fifty traditional NSAIDs are commercialized - whilst only three coxib are available on the market (celecoxib, etoricoxib and parecoxib) - and moreover the cardiovascular risk seems to be very different among them. Besides, all NSAIDs are more selective on COX-2 than on COX-1 and all tend to lose their selectivity in a dose-dependent way. In other words, coxib are not another class of drugs, but are still NSAIDs made “different” only for marketing purposes. In addition, to call a coxib a “painkiller” is not particularly appropriated, since the key enzyme in pain transmission is COX-1, which is inhibited only at high doses by the most selective NSAIDs, cobix included.

Balance between gastrointestinal and cardiovascular toxicity

The majority of new data on NSAIDs cardiovascular toxicity comes from the SOS study - Safety on NSAIDs  (www.sos-nsaids-project.org). The SOS project is a translational study which evaluated NSAIDs cardiovascular and gastrointestinal safety by the systematic review of literature and combined use of databases containing the clinical and pharmacological data of 30 million of European citizens.
The systematic review has started with the evaluation of the results furnished by randomized and controlled studies and meta-analyses published until the end of 2008.2 The only high-quality data come from studies on coxib and traditional NSAIDs used as control (ibuprofen, naproxen and diclofenac). The randomized clinical trials results are then very limited and they are in line with the pharmacology and the selectivity grade of COX-2:3 naproxen is the least selective among the three traditional NSAIDs, then the most harmful at gastrointestinal level, but also the least cardiotoxic; diclofenac is the most selective of the three and then is the least harmful at gastrointestinal level, but is the most cardiotoxic. On the other hand, data on ibuprofen are less clear and it can be considered less dangerous at gastrointestinal level than naproxen and less cardiotoxic of diclofenac.
Data related to celecoxib do not differ too much from diclofenac ones, while it is not possible to draw any conclusion from the clinical trials data for etoricoxib.

What the observational studies say

The systematic review has then evaluated the results from observational studies. NSAIDs do not seem to increase stroke risk compared to the control group.4 On the other side, diclofenac, etoricoxib, etodolac and indomethacin report an increased risk of myocardial infarction, compared to the rofecoxib one. In the case of meloxicam, the risk increases but is less evident, whilst naproxen and celecoxib do not raise the risk of myocardical infarction compared to the control groups (non-NSAIDs users or old users).5

Not yet published data

The last part of the SOS study has analysed the data on 30 million of European citizens and on the most used NSAIDs in Europe. Definitive data are not yet public or published. Preliminary data confirm however an increased thromboembolic risk for diclofenac and etoricoxib, whilst celecoxib and other traditional NSAIDs have all a similar risk.
At present, the SOS project has not yet analysed the effect on cardiovascular risk of NSAIDs doses and therapy duration.
Waiting for these data, the EMA recommend the use of NSAIDs at the lowest possible dosages and for the shortest possible period. This advice is dictated by an empirical logic, but it does not consider the NSAIDs pharmacology or appreciate these two elements:

1. Cardiovascular risk of low-doses NSAIDs (of which some are not even subjected to prescription) has not been studied accurately;
2. COX-2 selectivity is reduced by increasing the doses, therefore low-doses of very selective drugs for COX-2, like diclofenac pr etoricoxib, could even be more cariotoxic then high-doses, in particular for high-risk patients.