Why do cox 2 inhibitors cause thrombosis

Retrospective cohorts were also carried out. Users of rofecoxib at high doses had 1. Additionally, the risk increased to 1. A Canadian cohort that followed and was published in 30 failed to demonstrate increases risk for AMI in recent users of rofecoxib when compared to the control group. After all the studies had rejected the possible association between rofecoxib and cardiovascular adverse events, other studies came to light to support the assumption initially suggested by the VIGOR 8 study.

A recent meta-analysis 31 sugested increased relative risk of those events among patients on many doses of rofecoxib when compared to those on naproxen, but not placebo. Juni and collaborators 32 developed, right afterwards, a new meta-analysis involving 18 randomized, controlled clinical trials, which compared rofecoxib with other NSAIDs or with placebo, and 11 observational studies on cardiovascular risk and naproxen. Observational studies also questioned the cardiovascular safety of rofecoxib.

In , Solomon and collaborators 33 conducted a case-control study where 54, patients over 65 years of age were observed in regard to hospital admission resulting from AMI. In addition, higher than 25 mg doses were associated to higher risk as compared to those under 25 mg. Finally, risk was shown to be high only in the first 90 days of use, but not in the period following it. The results reported by the study entitled Adenomatous Polyp Prevention on Vioxx APPROVe 13 were the ones to cause highest commotion among health authorities, scientific community, pharmaceutical companies and public opinion.

Results released in September, led to the immediate withdrawal of rofecoxib from the market by the company marketing it. A total of 1, patients with a history of colonic adenomas was included in a double-blind, randomized, controlled clinical trial, where one group would be administe red 25 mg of rofecoxib per day, and the other group, placebo.

The purpose was to investigate if the use of the agent for a 3-year period would change the risk of neoplastic polyp in the large intestine. After 18 months of treatment, the intervention was terminated after the agent was associated to significant increase of cardiovascular risk. Relative risk reported was 1. Even though, researchers found it to be uncertain whether re- sults obtained were due to rofecoxib alone or to COX-2 specific inhibitors class.

Concurrent to the publication of APPROVe, two other studies offered additional support to the hypothesis of class effect. Once again intervention had early termination, when data from 2.

When compared to placebo, the group being administered lower dose of the medication reported hazard ratio at 2. Celecoxib dose-dependent cardiovascular risk was determined at that point in time. APC study researchers discussed the results of other studies involving celecoxib. In their opinion, preliminary analyses of PreSAP trial did not report increase of cardiovascular risk.

They believe the fact that PreSAP uses a mg daily dose of celecoxib with no risk increase supports the assumption that sustained inhibition of prostacyclin to be accounted for such increase. In addition, researchers point out that such increase was also demonstrated in a celecoxib randomized, controlled clinical trial, with patients with Alzheimer disease, as reported to FDA Food and Drug Administration. While considering other publications on the cardiovascular adverse effects of other class agents, the APC study emphasizes the evidence of increased risks of those events with prolonged use of COX-2 inhibitors.

Another publication to support those evidences evaluated cardiovascular safety of another COX-2 inhibitor while administrating valdecoxib and its IV pro-drug parecoxib to 1, patients submitted to coronary artery bypass surgery 1. The agents were to be administered for 10 days to treat post-surgery pain as follows: IV parecoxib for at least 3 days, followed by valdecoxib orally up to day 10; IV placebo, followed by valdecoxib orally; or placebo for 10 days.

Patients also had access to opioids. Observational follow-up for adverse effects lasted 30 days. Higher rate of cardiovascular events was found among patients being administered parecoxib and valdecoxib as compared to those being administered placebo, with risk ratio at 3. The significant risk finding for thromboembolic events in patients at high risk for such events increased concerns related to the safety of medications in that class.

That concern had been raised in after the publication by Ott and collaborators 34 on similar results from a smaller study involving patients. Recent results from the last three studies involving rofecoxib, celecoxib and valdecoxib have provided further previous evidence 17 and convinced specialists of the higher risk of cardiovascular adverse events by the class, such as AMI, CVA, hypertension and heart failure 13, There are indications about cardiotoxicity being dose-dependent and proportional do COX-2 selectivity Such selectivity is ranked from highest to lowest: lumiracoxib; etoricoxib, rofecoxib and valdecoxib; celecoxib and diclofenac.

However, such result may have been due to study small range and short term It is also reasonable to add the role played by lumiracoxib half life in such difference in results It is also to be pointed out that although it is marketed as a non-specific NSAID, COX-2 selectivity reported by diclofenac is very similar to that by celecoxib 35 , which may have had some influence in CLASS study results and in others.

However, the delay in defining COX-2 inhibitors cardiotoxicity raised great concern in scientific community , Papers published on the subject have presented considerable failure in regard to that definition. The VIGOR 8 study, for instance, excluded individuals with recent cardiovascular events and aspirin users. By doing that, it eliminated a considerable risk group for those events.

High risk patients for cardiovascular disease are known to be responsible for a major share of COX-2 7 users; their exclusion stands for a serious bias in study screening. Another factor to hinder elucidation was the choice for primary outcome.

The study focused gastrointestinal effects by agent, which poses difficulty in measuring cardiovascular effects Those effects may have been mistakengly classified, or even gone unnoticed, which would have posed difficulty in formulating consistent associations.

In the first place, it was a short-term study, and APPROVe 13 demonstrated that a longer-term follow-up would be necessary to detect outcomes. Additionally, CLASS was similar to VIGOR, since it was not designed for formal and systematic detection of cardiovascular events, and included patients at relatively low risk for such events Finally, some suggest that CLASS does not deny evidence on cardiovascular risk increase when comparing the use of celecoxib with ibuprofen among non-users of aspirin The control group would count on users of naproxen associated to a proton-pump inhibitor, following FDA recomendation Such studies would follow to establish risk levels for each medication, treatment time associated to that risk, and possible populations to be benefitted from the use of the agents.

However, there is controversy on how practical those large scope studies could be, since patients at high risk for cardiovascular diseases would be reluctant to participate There is no scientific evidence of higher effectiveness by COX-2 specific inhibitors Therefore, their use should be restricted to patients for whom other treatment strategies had failed.

The clinical recommendations developed at scientific sessions of the American College of Cardiology ACC 35 for the use of anti-inflammatory are the following:. Should the use of COX-2 inhibitors be indispensable, therapy risks and benefits must be carefully analyzed Additionally, pacients must be aware of the risks involved; dose must be the lowest possible; and treatment time frame the shortest possible. It is important to point out, once again, that risk assessment still poses diffi- culties, since no proper evaluation has been made available by long-term studies involving high and low risk populations 16, Finally, regulatory agencies, such as the FDA in the United States, have received severe criticism regarding the control and the survaillance of such processes 5, Those agencies should require compatible knowledge update on the part of manufacturers, that including changes in package inserts and drug indications; education of patients and health professionals; advertising limitations; use restrictions for certain groups of patients; implementation of studies and assays related to the safety of agents; sales interruption and agent withdrawal from the market When drugs are so widely used 3,15,16, public health issues are involved; therefore, they should not have been marketed for such a long time without accurate definition on theis cardiovascular safety.

The concern in regard to public health must come before the commercial interests by the pharmaceutical companies. Based on article by Okie 45 , the FDA decided COX-2 selective inhibitors could be kept being marketed provided the following recommendations were complied with: packaging must contain warnings about recently proven effects, in addition to other measures to restrict use.

After FDA's positioning rofecoxib manufacturers declared the product will be reintroduced in the market in the near future The American agency also pointed out the need for a wide-range multicenter study to investigate the safety of the NSAIDs class as a whole - long and short-term.

The study would take years, and would include, according to the FDA, patients with osteoarthritis, rheumatoid arthritis, or chronic pain, stratified by cardiovascular disease risk at baseline. A proposal was made to compare groups with over 1, individuals being administered ibuprofen, naproxen, diclofenac and celecoxib, with a control group being administered therapeutic doses of aspirin and proton pump inhibitor or paracetamol with codeine.

Blood pressure should be monitored and patients at increased risk for conorany diseases should be administered aspirin at low doses. However, funds are not yet available for such study; and as discussed earlier, there is still controversy regarding its practicality. Complications of the COX-2 inhibitors parecoxib and valdecoxib after cardiac surgery.

N Engl J Med ; The coxibs, selective inhibitors of cycloxygenase Lessons from the withdrawal of rofecoxib. Six studies one cohort study and five case—control studies with 21 VTE events met our eligibility criteria and were included in the analysis [ 10—15 ]. The detailed characteristics and quality assessment of these six studies are described in Tables 1 and 2.

The two investigators independently performed this data extraction. Review Manager 5. We reported the pooled effect estimate of VTE using the combination of data from case—control and cohort analyses to increase the precision of our estimates.

We used the ORs of case—control studies as an estimate of the RRs to pool these data with the RR of the cohort study, as the outcome of interest was relatively uncommon [ 16 ]. Adjusted point estimates and standard errors were extracted from individual studies and were combined by the generic inverse variance method of DerSimonian and Laird [ 17 ].

Given the high likelihood of between-study variance with the different study designs, definitions of NSAID use and populations, we used a random effects model rather than a fixed effects model.

This statistic was complemented with the I 2 statistic, which quantifies the proportion of total variation across studies that is due to heterogeneity rather than chance.

Three studies reported a risk ratio for participants who used selective COX-2 inhibitors. With the concern over high heterogeneity, we performed jackknife sensitivity analysis by excluding one study at a time. The results of this sensitivity analysis suggested that our results were robust, as the pooled risk ratios remained significantly elevated, ranging from 1. Funnel plots to evaluate publication bias are shown in Fig.

The graph is asymmetric, suggesting that publication bias in favour of positive studies may be present. The VTE risk found in this study is slightly higher than the coronary artery thrombosis risk seen in another meta-analysis ranging from 0. With the widespread use of these medicines, this increased risk may have important public health implications.

Heterogeneity between studies was present in this meta-analysis. We suspect that differences in study design, definitions of NSAID exposure and population were the main source of heterogeneity. Three studies were done in hospitalized subjects [ 12 , 13 , 15 ] and three studies included both ambulatory and hospitalized subjects [ 10 , 11 , 14 ].

The definition and method of verification for NSAID exposure also varied from study to study, with some studies using a pharmacology-linked database [ 10 , 11 , 13 , 14 ], while other studies used a structured or phone interview [ 12 , 15 ]. Controlling for confounders might also contribute to this heterogeneity, as it was done differently between studies, from virtually no correction to control for a large number of confounders. It should be noted that, from our sensitivity analysis, exclusion of the study by Bergendal et al.

The pathophysiology of increased arterial thrombosis risk and thus coronary artery events is explained by a thromboxane—prostacyclin imbalance. Inhibition of the COX-2 enzyme has been shown to inhibit the synthesis of prostacyclins, potent platelet activation inhibitors, while stimulating the release of thromboxane, a potent platelet aggregation facilitator, from the activated platelets.

The activation and aggregation of platelets might, in turn, induce a coagulation cascade and clotting [ 4 , 19 , 20 ]. This mechanism might explain the increased risk of venous thrombosis we observed in this study. Even though the six studies included in this meta-analysis were of high quality, there are some limitations and thus our results should be interpreted with caution.

First, we cannot exclude the possibility of publication bias in favour of positive studies, as the funnel plot is asymmetric. Second, the statistical heterogeneity in this study is high and thus the data from individual studies might be too heterogeneous to combine.

Third, most of the included studies were conducted using a medical registry—based database, raising the possibility of coding inaccuracy. Fourth, this is a meta-analysis of observational studies and thus can only demonstrate an association, not establish cause and effect, so we cannot be certain that NSAIDs themselves or other potential confounders increase the risk of VTE.

For example, patients may have been prescribed NSAIDs for underlying illnesses causing pain and immobility or for chronic inflammatory disorders, which are linked to a higher VTE risk compared with the general population [ 23—25 ]. We could not perform a meta-regression to adjust for these potential confounders as the primary studies do not provide sufficient data to do so.

This is the first meta-analysis to investigate the association between NSAIDs and venous thromboembolism. This study demonstrated a statistically significant increased risk of venous thromboembolism among NSAID users. Funding : No specific funding was received from any funding bodies in the public, commercial or not-for-profit sectors to carry out the work described in this article.

Disclosure statement : The authors have declared no conflicts of interest. Google Scholar. Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide.

Sign In or Create an Account. Sign In. Advanced Search. Search Menu. Article Navigation. Close mobile search navigation Article Navigation. Volume Article Contents Abstract. Materials and methods. Non-steroidal anti-inflammatory drugs and risk of venous thromboembolism: a systematic review and meta-analysis.

Patompong Ungprasert , Patompong Ungprasert. Just before Celebrex and Vioxx were approved and launched, a group led by Garret FitzGerald, MD, chair of the department of Pharmacology, and director of the Institute for Translational Medicine and Therapeutics at Penn, observed that both drugs suppressed prostacyclin in humans, as reflected by its major metabolite in urine, PGI-M.

Based on the potentially cardioprotective properties of prostacyclin, which relaxes blood vessels and unglues platelets in test tube experiments, the team predicted that shutting down this protection with inhibitors would cause heart attacks and strokes. More than 10 years later, it is now clear what the COX inhibitors do in the body.

Eight placebo-controlled, randomized trials, performed to find new uses of these drugs, showed that they posed a cardiovascular hazard, similar in magnitude to that resulting from being a smoker or a diabetic, notes FitzGerald. Arguments against the proposed mechanism were threefold. First, it was proposed that COX-2 didn't exist under normal circumstances in the blood-vessel lining and PGI-M came from some other source.

The kidneys were suggested as the source by some researchers. Second, even if blood-vessel prostacyclin was blocked, other protective mechanisms, especially formation of nitric oxide NO would take over. And third, although NSAIDs elevate blood pressure, it was proposed that this observation was unrelated to COX-2 and treating high blood pressure would deal with the problem.

FitzGerald's group has now "closed the loop" with its earlier clinical studies and answered these questions in a paper just published in Science Translational Medicine. In it, they confirm that COX-2 is expressed in cells lining blood vessels and that selectively removing it predisposes mice to blood clotting and high blood pressure.

Indeed, the lost NO may not be the only step that magnifies the effects of losing prostacyclin. In a second paper, published in April , in the Proceedings of the National Academy of Sciences, FitzGerald's group shows that arachidonic acid, the fat broken down by COX-2 to make prostacyclin, can be shunted down another pathway to make a new series of dangerous fats called leukotrienes when COX-2 is disrupted.

Clinical studies have shown that those most at risk from COX-2 inhibitors are patients who already have heart disease.