Gastrointestinal bleeding after the introduction of COX 2 inhibitors: ecological study
BMJ 2004; 328 doi: https://doi.org/10.1136/bmj.38068.716262.F7 (Published 10 June 2004) Cite this as: BMJ 2004;328:1415
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
The publication by Mamdani et al. seems to be a confirmation that COX
-2 inhibitors (celecoxib and rofecoxib) have gastrointestinal detrimental
effects which could lead to serious public health consequences such as an
increase in the rate of admission for gastrointestinal bleeding in elderly
patients.
However, if we admitted that this increase was “due” to the increase
in the percentage of elderly who have taken NSAIDs or COX-2 inhibitors, we
should be surprised that no change in the rate of admission for heart
failure was observed.
Indeed, it has been shown that COX-2 inhibitors as well as non-
selective NSAIDs favour salt and water retention, increasing the risk of
hypertension and of oedema formation (1). Moreover, non-selective NSAIDs
is known to increase the risk of heart failure (2). So why did Mamdani et
al. not observe any increase in the rate of admission for heart failure,
especially in such an elderly population? Does it mean that COX-2
inhibitors do NOT increase the risk of heart failure?
The authors admitted that they “cannot prove causation” but that they
“believe that the striking temporal correlation, biological plausibility”
as well as “the lack of any other trends that would explain the
association strongly suggest that the two events are directly related”. If
we used the same arguments for the lack of increase in the rate of
admission for heart failure, this would lead to the conclusion that COX-2
inhibitors does NOT have any effect on the risk for heart failure.
However, this conclusion would be in contradiction with a recent
publication in the Lancet by… Mamdani et al. themselves (3)! In a
population-based retrospective cohort study of elderly patients, they
showed that the risk of admission for heart failure was increased in users
of rofecoxib and non-selective NSAIDs (3)!
Interestingly, in the BMJ publication, the rate of admission for
heart failure and for myocardial infarct was assessed. The authors did not
mention why they measured these rates. Can we suppose that the authors
expected some increase at least in the rate of admission for heart
failure? If it is the case, why did they not discuss that observation?
In fact, the increase in the prescription of COX-2 inhibitors may be
the causal agent for this increase in the rate of admission for
gastrointestinal bleeding but the demonstration here is not convincing.
Ecologic studies are “easily and inexpensively conducted” but the results
are often “difficult to interpret” (4). In fact, I am not convinced that
they have controlled for all possible confounders: thus, hospital
admission for any conditions in elderly patients has many determinants
which might not have been considered into that study.
Finally, quoting A.B. Hill (5), to “pass from […] (an) association to
a verdict of causation”, one important step is to consider the “coherence”
of the observations. Here, we have an absence of coherence in the
observations by Mamdani et al., raising some doubts about their
conclusion.
I would be glad to see how the authors can explain the contradiction
between their two publications.
1. Chiolero A, Maillard M, Burnier M. Cardiovascular hazard of
selective COX-2 inhibitors: myth or reality? Expert Opin Drug Safety 2002;
1: 45-52
2. Page J, Henri D. Consumption of NSAIDs and the development of
congestive heart failure I elderly patients: an unrecognised public health
problem. Arch Int Med 2000; 160: 77-784
3. Mamdani M, Juurlink DN, Lee DS, Rochon PA, Kopp A, Naglie G, Austin PC,
Laupacis A, Stukel TA. Cyclo-oxygenase-2 inhibitors versus non-selective
non-steroidal anti-inflammatory drugs and congestive heart failure
outcomes in elderly patients: a population-based cohort study. Lancet
2004; 363(9423): 1751-1756
4. Morgenstern H. Ecologic studies. In: Modern epidemiology by K.J.
Rothman and S. Greenland. Lippincott, Williams and Wilkins, 1998, pp 459
5. Hill AB. The environment and the disease: association or causation?
Proc R Soc Med 1965; 58: 295-300
Competing interests:
None declared
Competing interests: No competing interests
This apparent paradox is not surprising since COX-2 inhibitors are
often reserved for patients at high risk of gastrointestinal (GI) adverse
events. Moreover, this study was in an elderly population, who are likely
to be at higher risk of GI events, even in the absence of other risk
factors. A similar survey in the US (Ofman et al 2004) showed that users
of COX-2 inhibitors were more likely to be at higher risk of GI bleeding
than users of conventional NSAIDs, but gastroprotective prophylaxis with
PPIs was only used in 11% of patients at high risk of GI bleeding. These
studies do confirm that gastroprotection should be considered both in
patients treated with COX-2 inhibitors as well as those treated with
conventional NSAIDs.
Reference
Ofman JJ, Badamgarav E, Henning JM, Knight K, Laine L. Utilization of
nonsteroidal anti-inflammatory drugs and antisecretory agents: a managed
care claims analysis. Am J Med 2004;116:835-42
Competing interests:
None declared
Competing interests: No competing interests
Comment on Mamdani et. al. BMJ 2004
Editor – Mamdani et al (BMJ 2004) state that in Ontario the annual
absolute increase of more than 90,000 additional individuals annually
using NSAIDs – and the absolute increase of more than 650 upper
gastrointestinal haemorrhage hospitalizations annually - is entirely
attributable to the use of COX-2 specific inhibitors.
These findings reflected aggregate data only, and as such, did not
provide the authors with the capability to look comparatively at the
individual medications studied (non-specific NSAIDS or individual COX-2
specific inhibitors).
In the analysis three new drugs were introduced in Ontario that are
unique and fall into 2 separate classes: Celebrex and Vioxx are coxibs
whereas, Meloxicam is an oxicam. A large cohort study using the UK
General Practice Research Database, concluded that coxib exposure, but not
meloxicam was associated with a significantly lower risk of
gastrointestinal hemorrhage than older non-specific NSAID exposure (1).
According to data from the Brogran Inc. Ontario Drug Benefit database (2),
meloxicam was considered a COX-2 specific inhibitor which accounted for an
average of 51% of all new COX-2 specific inhibitor users as defined by
Dr.Mamdani from September 2001 to February 2002 in the Ontario Drug
Benefit Plan.
Additionally, there may be channelling bias, which means the
introduction of safer NSAIDs like COX-2 specific inhibitors celecoxib and
rofecoxib may have attracted higher risk patients.
Furthermore we consider the rate of hospitalisation for GI
haemorrhage per 10,000 NSAID prescriptions, rather than per 10,000
population would be more relevant and which we would expect to have
fallen, given a 40% increase in prescription rates compared to only a 10%
rise in hospitalisation. Singh et al recently demonstrated that in a
representative sample of US hospitals, rates for hospitalisation for GI
haemorrhage has indeed fallen over a similar time interval studied by
Mamdani et al. (3)
Previously Mamdani et al. demonstrated the risk of hospitalisation
for GI haemorrhage at the individual level with the COX 2 specific
inhibitors rofecoxib and celecoxib was significantly lower than that of
conventional NSAIDs (4) and celecoxib was no different from a community
control group.
We believe that these additional factors must be considered when
discussing the therapeutic impact of the COX-2 specific inhibitors.
Mitch Gandelman, MD, PhD
Vice President, Worldwide Medical - Oncology, Pain and Inflammation
Pfizer, Inc.
1. Macdonald T M, Morant S V, et al. Channeling bias and the
incidence of gastrointestinal haemorrage in users of meloxicam, coxibs,
and older non-specific non-steroidal anti-inflammatory drugs. Gut Sept
2003. 52. 1265
2. Brogan Inc. ODB Database
3. Singh S, Mithal A, Triadafilopoulos G. Presentation: Age-adjusted
hospitalization rates for complicated gastric and duodenal ulcers in the
US: have COX-2 specific inhibitors and PPIs made any difference? New
Orleans, LA: American Association for the Study of Liver Diseases,
American Gastroenterological Association, American Society for
Gastrointestinal Surgery, and Society for Surgery of the Alimentary Tract:
Digestive Disease Week 2004; May 15-20, 2004: Abstract 106923.
4. Mamdani M, Rochon PA, et al. Observational study of upper
gastrointestinal haemorrhage in elderly patients given selective.
cyclooxygenase2 inhibitors or conventional nonsteroidal anti-
inflammatory drugs. BMJ Sept 2003, 325. 624
Competing interests:
Employee Pfizer, Inc.
Competing interests: No competing interests