24 / 28
Celecoxib noninferior to ibuprofen and
naproxen for cardiovascular safety
The New England Journal of Medicine
Take-home message
•
The authors assessed the noninferiority of celecoxib compared with ibuprofen
and naproxen in 24,081 patients who required NSAIDs due to osteo- or rheuma-
toid arthritis and so were at higher cardiovascular risk. Gastrointestinal and renal
adverse events were also evaluated. Median follow-up was 34 months. During the
study, 68.8% of the patients stopped taking their medication and 27.4% discon-
tinued follow-up. In the intention-to-treat analysis, the primary composite outcome
(cardiovascular death [including hemorrhagic death], nonfatal myocardial infarction,
and nonfatal stroke) was met by 2.3% of the celecoxib group compared with 2.5%
of the naproxen group and 2.7% of the ibuprofen group (celecoxib noninferior to
naproxen and ibuprofen; P < 0.001 for both comparisons). Additionally, celecoxib
was associated with a lower risk of gastrointestinal side effects than naproxen (P =
0.01) and ibuprofen (P = 0.002). The risk of kidney events was lower with celecoxib
than ibuprofen (P = 0.004) but not naproxen.
•
Relative to cardiac safety, celecoxib at moderate doses was found to be noninferior
to ibuprofen and naproxen.
Abstract
BACKGROUND
The cardiovascular safety of
celecoxib, as compared with nonselective
nonsteroidal antiinflammatory drugs (NSAIDs),
remains uncertain.
METHODS
Patients who required NSAIDs for
osteoarthritis or rheumatoid arthritis and were
at increased cardiovascular risk were randomly
assigned to receive celecoxib, ibuprofen, or
naproxen. The goal of the trial was to assess
the noninferiority of celecoxib with regard to the
primary composite outcome of cardiovascular
death (including hemorrhagic death), nonfatal
myocardial infarction, or nonfatal stroke. Nonin-
feriority required a hazard ratio of 1.12 or lower,
as well as an upper 97.5% confidence limit of
1.33 or lower in the intention-to-treat population
and of 1.40 or lower in the on-treatment popula-
tion. Gastrointestinal and renal outcomes were
also adjudicated.
COMMENT
COX-2 inhibitors: safe or not?
By Peter Lin
MD, CCFP
T
he basis of this question comes from the cyclooxygenase
(COX) enzyme system, which creates prostaglandins that
have many different physiologic effects. For example, the
COX-1 enzyme in the stomach provides protection from acidity.
On the other hand, the COX-2 enzyme in joints creates prosta-
glandins that cause inflammation and pain. The original NSAIDs
inhibited both COX-1 and COX-2 and hence, they were good
for joint pain but bad on the stomach lining, resulting in ulcer
development.
This led to the development of selective COX-2 inhibitors, which
did not affect the stomach but still had good anti-inflammatory
effects. The only problem was that the COX enzymes are also
found elsewhere in the body, not only in the stomach and joints.
In the endothelium coating of blood vessels, the COX-2 enzymes
produce prostaglandins that prevent platelets from sticking and
clotting. So, a COX-2 inhibitor would block that process and
platelets would be more likely to stick and form clots. Perhaps
this is the explanation of why rofecoxib had increased cardiac
events compared with naproxen.
Interestingly, in the platelets, COX-1 makes prostaglandins that
promote clot formation, so the theory is that an NSAID that blocks
both COX-1 and COX-2 equally would still maintain balance –
COX-2 is blocked in the endothelium but COX-1 is blocked in the
platelet. Hence, balance is maintained and there is no increase
in clot formation. That is the simple explanation of why naproxen
does not seem to increase CV risk. However, because naproxen
blocks COX-1 in the stomach, GI complications are still there.
The PRECISION trial was designed to see if celecoxib blocks
enough COX-2 to have the same CV risk issues as rofecoxib. This
was funded by Pfizer but was done with the advice of the FDA.
This study enrolled over 24,000 patients and randomly assigned
them to receive celecoxib, naproxen, or ibuprofen. Over 90%
were osteoarthritis patients, with 10% having rheumatoid arthri-
tis. The mean follow up was 34 months.
Unfortunately, 68% of the patients stopped their therapy and
27% of the patients dropped out of the study. This is terrible for
the study, but it reflects what our patients do in real life. They
take their anti-inflammatories when they need to and they do
not take them on a regular basis.
Despite these drop outs there were still large amounts of use
data. From the remaining patients, there was no difference in
the primary cardiovascular outcomes among any of the ther-
apies. Interestingly, despite the use of PPIs in the study, there
were still less gastrointestinal events in the celecoxib group.
Below are some of the significant findings of the on-treatment
patients in this study.
Outcome for on treatment patients
Celecoxib vs
naproxen
Celecoxib vs
ibuprofen
Primary APTC outcome
0.90 (0.71–1.15)
0.81 (0.65–1.02)
Serious gastrointestinal events
0.45 (0.33–0.63)
0.44 (0.32–0.61)
Renal events
0.66 (0.44–0.97)
0.54 (0.37–0.80)
Death from any cause
0.65 (0.46–0.92)
0.68 (0.48–0.97)
It would seem that celecoxib is as safe as naproxen or ibuprofen
for CV events, but definitely celecoxib had fewer gastrointestinal
complications. And thankfully, most of our patients do not take
these agents continuously. They tend to stop and start in real
life, which, in fact, may reduce their exposure and
hence reduce their potential risks.
Dr Lin is Director of Primary Care Initiatives,
Canadian Heart Research Centre, and Medical
Director at LinCorp Medical Inc, Canada.
RHEUMATOLOGY
24
PRACTICEUPDATE DERMATOLOGY