Aspirin and COX-2 selective NSAIDs:
what the evidence tells us:
It was hoped that the introduction of the coxib class of COX-2 selective NSAIDs
(rofecoxib, celecoxib, etoricoxib, valdecoxib) had at last overcome the problem of
poor gastrointestinal tolerability with NSAIDs. These new agents, by selectively
inhibiting inducible COX-2 not constitutive COX-1, should have offered a safer
treatment for arthritis. Instead, they became yet another example of misguided
optimism, as long-established and evidence-based treatments were discarded in
favour of relatively untested new drugs.
What were the concerns about non-COX selective NSAIDs?
There was no evidence of differences in efficacy among older NSAIDs so attention
naturally focused on their relative safety. Before the advent of the coxibs, the major
concern about non-aspirin NSAIDs was dose-related gastrointestinal (GI) toxicity. It
was estimated that 10 - 20 percent of patients had dyspepsia while taking an NSAID
and 5 - 15 percent discontinued treatment within 6 months primarily due to adverse
GI effects1. In 1994 the Committee on Safety of Medicines ranked the most widely
prescribed non-aspirin NSAIDs in order of the risk of upper GI toxicity: ibuprofen was
associated with least risk; diclofenac, naproxen, ketoprofen and indomethacin with
intermediate risk; and piroxicam with highest risk2. Compared with ibuprofen, these
agents were associated with a 1.6 - 4.2-fold increased risk of gastrointestinal
bleeding and peptic ulcer perforation.
Non-aspirin NSAIDs were also associated with other serious adverse effects such as
renal, hepatic, allergic and haematological reactions but these events were less
common than GI toxicity and the differences between NSAIDs were less marked.
What did COX-2 selective NSAIDs offer?
COX-2 selective NSAIDs reduced the risk of GI toxicity compared with non-selective
NSAIDs. Two large trials appeared to support this. The CLASS study compared
celecoxib, diclofenac and ibuprofen in 8059 patients with rheumatoid arthritis or
osteoarthritis. It found that celecoxib was associated with a significantly lower
incidence of upper GI bleeding, perforation or ulceration and symptomatic ulcers
(1.40 vs 2.91 percent with the other NSAIDs) after 6 months; there were no
differences in the risk of cardiovascular events. The VIGOR study compared
rofecoxib and naproxen in 8076 patients with rheumatoid arthritis. After 9 months,
rofecoxib was associated with less than half the risk of upper GI bleeding, perforation
or ulceration and symptomatic ulcers (2.1 vs. 4.5 percent).
The fall of the coxibs:
Rofecoxib was withdrawn in September 2004 when, in the APPROVe trial, it was
associated with a 2-fold increased risk of myocardial infarction and stroke (6 events
per 400 pt.yrs vs. 3 per 400 pt.yrs with placebo). This was subsequently confirmed
by a meta-analysis of clinical trials. Next, another unpublished trial linked celecoxib
with an increased risk of cardiovascular events (2.3-fold increased risk at 400 mg/day
and 3.4-fold increased risk at 800 mg/day). The European Medicines Evaluation
Agency (EMEA), Europe's drug regulatory body, then advised that valdecoxib and
parecoxib (licensed only for the treatment of pain in the UK) were contraindicated in
patients undergoing coronary artery bypass surgery because of an increased risk of
serious cardiovascular thromboembolic events. Following a review of the safety of
the coxibs by the EMEA, the UK Medicines and Healthcare Products Agency issued
advice that the coxibs should not be prescribed for patients with ischaemic heart
disease or cerebrovascular disease (formerly they could be prescribed with caution
for patients with ischaemic heart disease), or in patients with moderate to severe
heart failure; an alternative treatment (an NSAID with gastroprotection if indicated)
should be considered for all patients.
Some studies have found that the coxibs do not increase cardiovascular risk and
the most recent analyses suggest that rofecoxib is associated with a higher risk than
celecoxib. But not all safety concerns focused on cardiovascular events. The
EMEA has warned that valdecoxib and parecoxib are associated with serious skin
reactions; and a Canadian analysis showed that the advantage of the lower GI risk
associated with rofecoxib and celecoxib was outweighed by their increased use, so
that total hospital admissions for GI haemorrhage actually increased after these
agents were introduced.
What is the mechanism underlying these effects?
There seems little doubt that the increased cardiovascular risk associated with the
coxibs is a class effect, with varying degrees of expression in different drugs within
the class. The underlying mechanism is still unclear but it seems likely that selective
COX-2 inhibition destabilises the balanced cardiovascular effects of thromboxane A2
and prostacyclin I2.
Prostacyclin I2 is produced by endothelial cells by COX-2; it inhibits platelet
aggregation, causes vasodilatation and (in vitro) prevents vascular cell proliferation.
Thromboxane A2 is produced by platelets by COX1 has opposing effects, promoting
platelet aggregation, vasoconstriction and vascular proliferation. COX-2 selective
NSAIDs therefore inhibit production of prostacyclin I2 without affecting thromboxane
A2 and this results in increased blood pressure and an exaggerated thrombotic
response; animal studies also suggest it enhances atherogenesis.
What lessons can we learn?
The rise and fall of the coxibs provides a salutary reminder that drug safety cannot be
taken for granted. Premarketing clinical trials involve too few patients to provide a
reliable estimate of the possible risk of uncommon but serious adverse events -
particularly when those events are relatively frequent, as is the case with myocardial
infarction. Only long-term use can provide the necessary clinical experience in large
numbers of patients, including groups with comorbidities who are more vulnerable.
Now is the right time to reappraise the safety and efficacy of the long-established
NSAIDs for which there is reliable evidence.
0 comments:
Post a Comment