 |
||
Volume 2 Number 5
|
||
EDITORIAL
Nonsteroidal anti-inflammatory drugs in Juvenile Idiopathic
Arthritis: What’s the real risk?
Introduction
Given the known pharmacology
of nonsteroidal anti-inflammatory drugs (NSAIDs), it should come as
no surprise to us that such drugs may be associated with cardiovascular
effects (both “good” and “bad” ); after all, these effects are why acetylsalicylic
acid in low doses is given to individuals at increased risk of thrombosis. As Harry Gewanter
recently pointed out on the Pediatric Rheumatology Listserver, for any
intervention we have to balance the potential benefits of the
intervention with the potential side effects of the intervention,
and the potential effects of not intervening. Therefore the issue
for us prescribing these drugs, and our patients taking them, revolves
around the risk/benefit ratio of using these drugs, and the risk/benefit
ratio of not using these drugs. As Harry noted, people tend to forget
that not taking a drug is also an active therapeutic decision Pharmacology of
NSAIDs
NSAIDs have the potential
for cardiovascular effects because a significant part of their antiinflammatory
mechanism of action is to inhibit cyclooxygenase thereby decreasing
the production of prostaglandins which have potent vascular activities
(1). It has been postulated that whether or not an NSAID is associated
with an increased or decreased cardiovascular risk may depend in large
part on the resulting balance between the inhibition of cyclooxygenase
2 COX-2 and the inhibition of cyclooxygenase 1 (COX-1)
caused by any particular NSAID (2). In this hypothesis the more a drug
is specific for COX-2, the greater will be the inhibition of prostacylin
with its vasodilating/antithrombotic effects. Similarly, the more a
drug is specific for COX-1, the greater will be the induced inhibition
of thromboxane with its platelet aggregating effects. Consequently the
more specifically the drug inhibits COX-2, the greater the likelihood
of a vascular occlusive event occurring. This hypothesis is not the
whole story however as other factors affecting vascular wall integrity
such as systemic inflammation and atherosclerosis are undoubtedly important.
There is also good evidence that there are differences between individual
COX-2 inhibitors in where they locate in the cell membrane lipid bilayer;
these differences may help explain differences in the vascular toxicity
of celecoxib compared to rofecoxib (3). For more details about this
controversial area the reader is referred to a recent excellent Medscape
review (4). Cardiovascular
Toxicity of NSAIDs in adults
It has been known
for many years that elderly individuals on NSAIDs or individuals with
known renal disease are more likely to develop hypertension and cardiac
failure (5,6). Given this knowledge, although it is not clear at the
time of writing this editorial what were the cardiovascular events that
lead to the discontinuation of naproxen in the prevention of Alzheimer’s
trial, it does not seem greatly surprising that an increase in cardiovascular
events was found given the presumed
advanced age of individuals at risk for Alzheimer’s disease,. Fries and colleagues in 1991 evaluated
the relative toxicity of NSAIDs (COX non-specific) using a sophisticated
toxicity index (7). Using the ARAMIS data bank they studied 2747 patients
with rheumatoid arthritis and 5,642 courses of the 11 most frequently
used NSAIDs with the study period encompassing 8,481 patient-years.
They found substantial differences between the
individual NSAIDs, but cardiovascular toxicity was not a finding in
this study. In 2002 Ray et al.
(8) performed a retrospective cohort study of individuals on the expanded
Tennessee Medicaid program involving 202916 non-users, 24132 users of
rofecoxib, and 151728 users of other NSAIDs. The study provided fairly
convincing evidence that rofecoxib at doses in excess of 25 mg daily
had an increased risk of serious coronary heart disease (RR 1.7 95%
CI 0.98-2.95, p=0.058) compared to non-use). Among new users the RR
increased to 1.93 (95% CI 1.09-3.42, p-0.024). There was no such increased
risk for rofecoxib at doses of 25 mg or less, or
for other NSAIDs. In the Medscape review
(4) Ray looked at data from 5 recent trials. The evidence from these
trials is that naproxen is not associated with an increased RR for coronary
heart disease compared to non NSAID use. Celecoxib is associated with
a slightly decreased risk while the RR for rofecoxib is >1 with the
RR increasing with doses greater than 25 mg. He also concludes in this
review that: “- one of the fortunate by-products perhaps of the rofecoxib
debacle is that naproxen has become very well studied, and there is
little evidence of its cardiotoxicity”. Further data concerning
the cardiovascular risks of rofecoxib is well presented in reference
4, and there seems little doubt that it is associated with significantly
increased risk of cardiovascular disease compared to non-specific NSAIDs
or to celecoxib, particularly when given in doses greater than 25mgs.
The data on celecoxib released by Pfizer
on December 17th 2004 that led to the halting of a clinical
trial investigating the role of celecoxib in preventing colon polyps,
indicates that patients taking larger doses than is normally prescribed
for rheumatoid arthritis (400 mg twice a day) had a 3.4 times greater
risk of cardiovascular events compared to patients taking placebo (9).
For patients taking 200 mg twice a day the RR was 2.5. The implications
of these findings for adults with RA or OA taking celecoxib at lower
doses are unknown. There is now good
epidemiological evidence that both RA and OA are associated with increased
cardiovascular events, with relative risks of about 1.5 to 2 for various
events (10,11). What role, if any, NSAIDs play
in this increased risk is unclear. It may be that joint inflammation
per se is the main factor associated with the increased risk. Cardiovascular
Toxicity of NSAIDs in Children
Although it is well
recognized that cardiovascular disease can occur in some forms of juvenile
idiopathic arthritis (JIA) (notably myocarditis and pericarditis in
systemic arthritis, and aortitis in rheumatoid factor positive arthritis
and enthesitis-related arthritis), there is little if any convincing
evidence that NSAIDs are associated with any increased risk of cardiovascular
disease. The first major report evaluating the toxicity of NSAIDs in
juvenile rheumatoid arthritis was by Barron and colleagues in 1982 (12).
They found a high frequency of toxicity necessitating discontinuation
of the NSAID in 79 of 101 (78%) children who had received more than
one NSAID. Acetylsalicylic acid appeared to be the major
culprit; none of the side effects appeared to be cardiovascular in origin.
A study by Flatø et al. in 1998 (13)
used a toxicity index similar to that used by Fries et al. (7) in adults.
They did not demonstrate any definite cardiovascular toxicity in 117
children (72 with JIA)
followed for a mean of 8.6 years. There is a known increase
in mortality in children with JIA, and an increased mortality among
adults with a history of JIA. There is very little evidence that any
of this increased mortality is due to cardiovascular disease. In a recent
national study from Finally, although
the Evidence-Based Medicine gurus would probably sneer, I believe that
the huge combined wealth of experience of pediatric rheumatologists
is critical to this discussion. There seems to be no concern from pediatric
rheumatologists around the world that cardiovascular disease is a significant
problem despite over 30 years of experience with NSAIDs therapies. Re-evaluating the
decision to intervene with NSAIDs
So what is the potential cardiovascular risk of treating
a child with an NSAID? My assessment of the data presented above is
that the risk of cardiovascular events for children treated with the
classical NSAIDs is vanishing small-certainly during their childhood
years. It may be that cardiovascular mortality is increased for adults
who had, or who continue to have active JIA, but whether or not this
is so is unclear, and it is even less clear if NSAIDs contribute to
this possible increased mortality. Even if we assume
that there is an increased risk for cardiovascular events in children
with JIA on NSAIDs, and even assuming a preposterously high relative
risk of 2.0 (about the level of risk for high dose rofecoxib in adults),
given childhood cardiac mortality rates of about one in 100,000 (15),
only about one child with JIA per year would have an
NSAID-related cardiovascular death for every 100,000 children treated
with long term NSAIDs. Therefore the potential
cardiovascular risk of using NSAIDs in our population is tiny. What is the potential benefit of treating a child with
an NSAID? There is no doubt that NSAIDs are moderately effective at
minimizing inflammation in JIA (16,17). However
I suspect only a relatively small number of patients-perhaps 20-25%
has complete resolution of their arthritis with NSAIDs alone. In other
words NSAIDs monotherapy is relatively ineffective at inducing remission,
and the majority of children with JIA require the addition of a second-line
agent (usually methotrexate). So to truly assess the potential
benefits of treating a child with an NSAID we really need to know
what is the marginal benefit of adding an NSAID to
methotrexate monotherapy. An RTC comparing methotrexate alone
as initial therapy to methotrexate and an NSAID is needed.
This might be a useful trial for CARRA or PRINTO to undertake.
However, how easy it would be to recruit patients in this present climate
is unclear. Given our present knowledge it is probably fair to say that
the potential benefit of NSAID therapy in JIA is moderately high. What is the potential risk of not treating a child with an NSAID? Younger pediatric rheumatologists will not know the awful damage wreaked by chronic arthritis in the days before the modern NSAIDs, intra-articular Triamcinolone, and methotrexate became available (see figure). Juvenile Idiopathic Arthritis can still occasionally cause severe joint damage, but much, much less frequently then 20 or so years ago. The risk of not effectively suppressing joint inflammation is huge! As mentioned above, it is unclear what is the relative importance of NSAID therapy, compared to particularly methotrexate, in decreasing this risk; however until we have evidence to the contrary, NSAIDs must remain central to our treatment regimens. |
||
 |
||
What should we
say to our patients?
Based on the present evidence there
is no doubt that, for the vast majority of children with JIA, anti-inflammatory
therapies (including NSAIDs) should be instituted early in the disease
course, and used aggressively to try and fully suppress joint inflammation.
Our job as pediatric rheumatologists is to explain as clearly as possible
to patients and parents that this approach has the most favourable risk/benefit
ratio. Unfortunately, all the evidence is that most human beings are
not good at risk assessment. The use of examples such as the approximately
one in three life time risk of cancer, or the fact that the child has
a hugely higher risk of being killed in a motor vehicle accident may
help. If the family wish to discontinue NSAIDs
that the child is already taking, or in the case of a newly diagnosed
patient, it might be best to use conventional therapies without NSAIDs,
the use of a second-line agent (methotrexate, or for enthesitis-related
arthritis, sulfasalazine) alone is probably an acceptable compromise.
Only for a tiny minority of children is no treatment an acceptable option.
Conclusions
The cardiovascular risk of NSAIDs in
childhood is almost non-existent. The risk of continued use into adulthood
is also extremely small. The benefits of such treatment for children
with JIA far outweigh the risks. |
||
| References 1. Moncada S, Vane JR. Arachidonic acid metabolites and the interactions between platelets and blood-vessel walls.N Engl J Med. 1979;300:1142-7. 2. FitzGerald GA. Nonsteroidal anti-inflammatory drugs, coxibs, and cardio-renal physiology: a mechanism-based approach. Am J Cardiol. 2002;89:1D-2D. 3. Walter MF, Jacob RF, Day
CA, Dahlborg R, Weng Y, Mason RP. 4. Lüscher TF et al. The pulse of current cardiovascular concern: anti-inflammatory medications. www.medscape.com/viewprogram/3690 Dec 22nd 2004. 5. Johnson AG, Day Ro. The problems and pitfalls of NSAID therapy in the elderly (part 1). Drugs Aging 1991;1:130-43. 6. Dieppe P, Bartlett C, Davey P, Doyal L, Ebrahim S. Balancing benefits and harms: the example of non-steroidal anti-inflammatory drugs. BMJ. 2004;329:31-4. 7. Fries JF, Williams CA, Bloch DA. The relative toxicity of nonsteroidal antiinflammatory drugs. Arthritis Rheum. 1991;34:1353-60. 8. Ray WA, Stein CM, Daugherty
JR, Hall K, Arbogast PG, Griffin MR. 9. Pfizer statement on new information regarding cardiovascular safety of Celebrex. www.pfizer.com/are/investors_releases/2004pr/mn_2004_1217.cfm 10. Wolfe F, Freundlich B, Straus WL. Increase in cardiovascular and cerebrovascular disease prevalence in rheumatoid arthritis. J Rheumatol. 2003;30:36-40. 11. Haara MM, Manninen P, Kroger H, Arokoski JP, Karkkainen A, Knekt P, Aromaa A, Heliovaara M. Osteoarthritis of finger joints in Finns aged 30 or over: prevalence, determinants, and association with mortality. Ann Rheum Dis. 2003;62:151-8. 12. Barron KS, Person DA, Brewer EJ. The toxicity of nonsteroidal antiinflammatory drugs in juvenile rheumatoid arthritis. J Rheumatol. 1982;9:149-55. 13. Flatø B, Vinje O, Førre Ø. Toxicity of Antirheumatic and anti-inflammatory drugs in children. Clin Rheumatol. 1998;17:505-510. 14. Thomas E, Symmons DP, Brewster DH, Black RJ, Macfarlane GJ. National study of cause-specific mortality in rheumatoid arthritis, juvenile chronic arthritis, and other rheumatic conditions: a 20 year followup study. J Rheumatol. 2003;30:958-965. 15. Health, United States 2004. www.cdc.gov/nchs
17. Leak
AM, Richter MR, Clemens LE, Hall MA, Ansell BM. A crossover study of
naproxen, diclofenac and tolmetin in seronegative juvenile chronic arthritis.
Clin Exp Rheumatol. 1988;6:157-160. |