To say that arthritis sufferers were eager to get their hands on these drugs is an understatement. COX-2s were so popular that doctors wrote more than 100 million prescriptions for celecoxib (Celebrex) and rofecoxib (Vioxx) in the year 2000 alone. Vioxx had sales of $2.6bn in 2001, and Celebrex has raked in more than $4bn for Pfizer since 1999 (Inflammopharmacology, 2005; 13: 419-25).

But now, researchers at Nottingham University have revealed that the COX-2s are just as bad as the NSAIDs they are meant to replace. In a study of more than 9400 patients diagnosed with a stomach ulcer or GI bleeding, the researchers found an increased risk of GI problems that was associated not only with the conventional NSAIDs such as ibuprofen, but with the COX-2s as well (BMJ, 2005; 331: 1310-6).

In other words, years of COX-2 commercial success have been based on claims that turn out not to be true.

This astonishing discovery, which has taken around 15 years to make, was based on patients’ records from 367 general practices in the UK. The data showed that rofecoxib was the most likely to cause a serious GI problem - hardly a surprise, given that this drug was withdrawn in 2004, after a new clinical trial showed an increased risk of serious thrombotic events (including heart attack and stroke) with long-term use.

The safety risk was deemed slightly less for celecoxib, the only COX-2 selective NSAID left on the market after valdecoxib (Bextra), like rofecoxib, was recalled due to concerns over cardiovascular events. However, a recent study found that the patients taking this drug developed even more dyspepsia than those taking the conventional NSAID naproxen (Am J Med, 2005; 118: 1271-8).

In fact, the large number of COX-2 prescriptions - due, no doubt, to the avid promotion of their better tolerability - has been linked to the growing number of cases of GI haemorrhage (Prescrire Int, 2005; 14: 177-8).

Celebrex: the sole survivor
Celebrex was the first COX-2 to be approved, in 1998, by the US Food and Drug Administration (FDA). The results of the CLASS (Celecoxib Long-Term Arthritis Safety Study) claimed that it caused fewer GI side-effects compared with the traditional NSAIDs. However, among patients given aspirin as a just-in-case medicine to prevent heart disease, those patients given celecoxib had similar GI problems (such as ulcers) as those given diclofenac and ibuprofen (JAMA, 2000; 284: 1247-55).

Then heart problems came to light. The Adenoma Prevention with Cele-coxib (APC) study, co-sponsored by the US National Cancer Institute and Pfizer, which makes the drug, revealed that patients taking 200 mg twice daily of the drug were nearly two and a half times more likely to suffer a heart attack or stroke. This risk rose to 3.4 times with 400 mg twice daily. Needless to say, the authors recommended that the cancer study be discontinued before it had run its intended course (N Engl J Med, 2005; 352: 1071-80); also, for more evidence about celecoxib, see www.cancer.gov/clinicaltrials/results/cox inhibitor-trials0205).

Even more damning for Pfizer, an unpublished study from 1999 had shown that elderly (Alzheimer) patients had four times the risk of a heart attack or stroke with the drug vs a placebo - a finding which only recently came to light.

A recent study of 2035 patients showed that Celebrex was associated with a dose-related, two- to threefold increase in death from cardiovascular causes, including stroke, heart attack and heart failure (N Engl J Med, 2005; 352: 1071-80). In terms of thromboembolic events (blood clots leading to stroke), Celebrex was no better than its banned sister Vioxx (Drug Saf, 2005; 28: 803-16).

Given these findings, a reconsideration of the cardiovascular safety profile of all NSAIDs - selective and non-selective - would seem warranted. Indeed, the FDA now recommends that all such drugs carry a ‘black-box’ warning for GI and cardiovascular risks (Curr Pain Headache Rep, 2005; 9: 377-89).

Kim Wallace