COX-2 inhibitors (coxibs)

General information

The cyclo-oxygenase (COX) that is responsible for prostaglandin synthesis exists in two isoforms, COX-1 and COX-2, which differ in their structure, regulation, expression, and function. COX-1 is expressed normally in a constant amount in almost all body tissues and produces prostaglandins important for the maintenance of normal homeostasis. In particular, among other important functions, they help to protect the gastrointestinal mucosa against ulceration and regulate renal function and platelet activity. COX-1 appears to be largely unaffected by inflammatory stimuli. In contrast, the COX-2 isoform is constitutively expressed only in the brain, in bone (associated with osteoblast activity), in the female reproductive tract (associated with both the ovulatory cycle and implantation of fertilized ova), and in the kidney, where it mayn play an important role in regulating renal function. In many other cells COX-2 is expressed at very low levels or is undetectable, but it is readily induced by inflammatory cytokines, mitogens, and endotoxins. Therefore, prostaglandins generated by COX-2 mediate pain and inflammation in many tissues and probably also have a role in renal, brain, and reproductive physiology, and in tissue repair [ ].

Traditional non-steroidal anti-inflammatory drugs (NSAIDs) inhibit both COX-1 and COX-2, providing benefit at sites of inflammation, but at the cost of potential adverse effects related to COX-1 inhibition, particularly in areas such as the gastrointestinal mucosa, platelets, and the kidney. The development of a drug that inhibits only COX-2 would offer the promise of relieving pain and inflammation without some, if not all, of these adverse effects. However, the concept that selective inhibition of COX-2 is only a positive event and inhibition of COX-1 a bad one may be simplistic, as it is flawed by many experimental data that have also implicated COX-2 as an integral component in the maintenance of physiological homeostasis [ ]. In particular, COX-2 may reduce inflammation by generating anti-inflammatory prostanoids [ ]; COX-2, like COX-1, may be present in normal gastric mucosa [ , ] and can play a physiological role there in defense mechanisms [ , ], and there is evidence that prostaglandins derived from COX-2 may be important in the healing of gastric ulcers [ ]. Furthermore, COX-2 inhibition alone may be insufficient to resolve inflammation and pain [ ], suggesting that COX-1 inhibition may play a role in reducing inflammation as well [ ]. Although the relevance of these experimental observations to the possible beneficial role of COX-2 activity is uncertain, they cannot be simply dismissed.

Nearly all of the traditional NSAIDs are predominantly COX-1-selective. In fact, only two of them, meloxicam and nimesulide [ , ], have been shown to have some COX-2 selectivity in humans. Two other NSAIDs, etodolac and nabumetone, may be COX-2 preferential inhibitors, but the evidence is less convincing than for meloxicam or nimesulide. Unfortunately, despite their wide use, clinical and epidemiological evidence that supports the claim of better tolerability of these preferential inhibitors with respect to other NSAIDs, is scanty and even controversial [ , ]. Furthermore, for both meloxicam and nimesulide there are data consistent with reduced COX-2 selectivity at high doses, those usually used in clinical practice. These considerations have prompted research for more selective COX-2 inhibitors, and three compounds, celecoxib, parecoxib, and rofecoxib, have become available. These compounds are much more selective than previous preferential inhibitors and they have no effect on COX-1 (COX-1 sparing drugs) over the whole range of doses used and concentrations achieved in clinical use. In particular, they have been shown in humans to spare gastric COX-1 to a much greater extent than traditional NSAIDs.

Is there convincing evidence that COX-2-selective NSAIDs offer a distinct therapeutic advantage over non-selective ones? To prove the usefulness of these compounds, clinical studies should show:

whether the COX-2 selective inhibitors are as effective as older NSAIDs in the relief of pain and/or inflammation;
whether in the effective dosage range there is evidence of less damage to the mucosa of the upper gastrointestinal tract than with conventional NSAIDs;
whether there are any effects on platelet in effective doses;
whether unexpected adverse drug reactions compromise the safety of these compounds.

As far as the last of these criteria is concerned, the finding that rofecoxib increases the risk of cardiovascular disease (heart attacks and strokes) during long-term therapy has led to a reappraisal of the role of the COX-2 inhibitors. Furthermore, both rofecoxib and valdecoxib have been withdrawn by their manufacturers, the former because of its adverse cardiovascular effects and the latter because of an increased risk of cardiovascular events after coronary artery bypass surgery and reports of potentially fatal skin reactions.

Celecoxib and rofecoxib have been evaluated in large randomized trials lasting from weeks to 1 year, for the relief of the signs and symptoms of osteoarthritis or rheumatoid arthritis in adults [ , ]. In all of these studies both compounds were more effective than placebo and at least as effective as traditional NSAIDs (that is ibuprofen, diclofenac, nabumetone, naproxen). However, there is less information about their efficacy as analgesics. A few randomized, double-blind trials have shown that both compounds are more effective than placebo; rofecoxib was as effective as non-selective NSAIDs (ibuprofen, naproxen) in relieving the pain of osteoarthritis or after dental surgery, but celecoxib was less effective [ ]. Rofecoxib was also analgesic in patients with primary dysmenorrhea [ ]. However, there are no published data on the usefulness of COX-2-selective NSAIDs in other types of acute pain, such as that related to acute gout, migraine, cancer, or biliary or ureteric colic.

General adverse effects and adverse reactions

Although the scientific literature has been flooded by articles and reviews on the use of selective COX-2 inhibitors, the safe prescribing and use of these compounds is controversial [ ].

COX-2 selectivity may be a double-edged sword. Because the cardiovascular risk outweighs the gastrointestinal risk in adults with rheumatoid arthritis or osteoarthritis, the harm would outweigh the benefits in most clinical settings. This means that the total number of serious adverse events would be increased by COX-2-selective NSAIDs compared with non-selective NSAIDs.

Much information on celecoxib and rofecoxib has come from two large pivotal randomized clinical trials, the Celecoxib Long-term Arthritis Safety Study (CLASS) and Vioxx Gastrointestinal Outcomes Research (VIGOR) studies, in which the efficacy and safety of the two coxibs and various non-selective NSAIDs were compared [ , ].

The COX-2-selective inhibitors are as efficacious as traditional NSAIDs in treating inflammation and pain due to osteoarticular diseases. Among unresolved questions, the most important relate to the possibility that COX-2 inhibitors may cause complicated ulceration in some subgroup of patients and may retard ulcer healing in others. Therefore, whether the COX-2 inhibitors in place of non-selective NSAIDs will prove to be cost-effective by reducing ulcer-related morbidity remains to be demonstrated. Moreover, it is likely that such drugs can cause the same renal problems as traditional NSAIDs do. Finally, the effect of COX-2 inhibitors on the incidence of vascular disease needs scrutiny, since vascular disease is a commoner cause of death than ulcer perforation or bleeding and since some of them clearly increase the long-term risk of heart attacks and strokes.

Criticisms of outcome studies that support the safety of COX-2 inhibitors continue to appear. A number of flaws in both the design and analysis of these studies have been reviewed [ ].

Organs and systems

Cardiovascular

Major cardiovascular events

Selective inhibitors of COX-2 were developed with the expectation that their use would be accompanied by a reduction in the severe gastrointestinal and renal adverse effects associated with conventional NSAIDs, adverse effects that were thought to be largely a result of inhibition of the COX-1 isoenzyme [ ]. However, evidence has since accumulated that COX-2 selective inhibitors (coxibs) not only cause gastrointestinal and renal toxicity but can also contribute to an increased risk of adverse cardiovascular events. The possible explanation of this phenomenon came from pharmacological studies that identified COX-2 inhibition as a plausible, albeit not the only, pharmacological mechanism for induction of thrombotic events. In fact, suppression of COX-2 dependent formation of PGI ₂ by coxibs, without significant concomitant inhibition of thromboxane A2 biosynthesis, can predispose patients to acute thromboembolic events. The pharmacological evidence that supports the mechanism for prothrombotic effects of coxibs has been reviewed [ ]. Figure 1 shows the physiological effects of COX-2 inhibition and the potential clinical implications [ ].

Figure 1, The physiological effects of COX-2 inhibition and the potential clinical implications [from 39 with permission].

There is also evidence from experimental and clinical studies that COX-2 may be atherogenic and thrombogenic and that selective COX-2 inhibitors may actually reduce, rather than cause, atherothrombotic vascular events [ ]. Some evidence supporting this hypothesis comes from laboratory and clinical studies. A recent randomized trial suggested that simvastatin reduces inflammation and suppresses the expression of cycloxigenase-2 and prostaglandin E synthase in plaque macrophages, and this effect may in turn contribute to plaque stabilization by inhibition of metalloproteinase-induced plaque rupture [ ]. Inhibition of COX-2 by celecoxib compared with placebo improved endothelium-dependent vasodilatation and reduced oxidative stress in men with severe coronary artery disease [ ].

Rofecoxib

Information about comes from the two large studies, CLASS and VIGOR, from a pooled cardiovascular safety analysis using individual patient data derived from all rofecoxib phase IIb to V trials conducted by the manufacturer that lasted at least 2 weeks, and from the adverse events reporting system in the USA [ ]. As a result of these and other studies the COX-2 inhibitors, rofecoxib and valdecoxib, have been voluntarily withdrawn from the market by their manufacturers, and other COX-2 inhibitors are under scrutiny.

In VIGOR, rofecoxib 50 mg/day was associated with a higher rate of non-fatal myocardial infarction (0.4%) than the non-selective COX-2 inhibitor naproxen 500 mg bd (0.1%) (RR = 0.2; CI = 0.1, 0.7) (33). In CLASS there was no difference in the rates of myocardial infarction in patients taking celecoxib (0.5%) and those taking ibuprofen or diclofenac (0.4%). However, the protocols of the two studies differed substantially with respect to the use of aspirin. In VIGOR, the patients were not allowed to take aspirin or any other antiplatelet drug, while in CLASS one-fifth of the patients took aspirin. A re-analysis of CLASS for cardiovascular thromboembolic events, including myocardial infarction, stroke, cardiovascular deaths, and peripheral events, showed no significant increase with celecoxib versus NSAIDs [ ].

The results of VIGOR might have been explained by a significant prothrombotic effect of rofecoxib and/or an antithrombotic effect of naproxen, which may have a significant antiplatelet effect. To clarify this, the annualized myocardial infarction rates in both VIGOR and CLASS were compared with those found in placebo-treated patients with similar cardiac risk factors enrolled in three meta-analyses of four aspirin primary prevention trials [ , ]. The analysis showed 0.24 and 0.30% increases over placebo in cardiovascular events for rofecoxib and celecoxib respectively, suggesting a prothrombotic potential of both COX-2 inhibitors. However, these results have been heavily criticized [ ], because of many potential pitfalls in the comparisons of patient populations in different trials. The results from a pooled analysis of individual patient data combined across rofecoxib phase IIb to phase V trials seem to be more reliable [ ]. Cardiovascular events were assessed across 23 studies. Comparisons were made between patients taking rofecoxib and those taking placebo, naproxen, or other non-selective NSAIDs (diclofenac, ibuprofen, and nabumetone). The major outcome measure was the combined end-point used by the Anti-platelet Trialist Collaboration, which includes cardiovascular death, hemorrhagic death, death of unknown cause, non-fatal myocardial infarction, and non-fatal stroke. More than 28 000 patients, representing more than 14 000 patient-years at risk, were analysed. The relative risk for an end-point was 0.84 (95% CI = 0.51, 1.38) when comparing rofecoxib with placebo; 0.79 (95% CI = 0.40, 1.55) when comparing rofecoxib with non-selective NSAIDs; and 1.69 (95% CI = 1.07, 2.69) when comparing rofecoxib with naproxen. These data provide no evidence for an excess of cardiovascular events for rofecoxib relative to either placebo or the non-selective NSAIDs that were studied. Instead, the difference between rofecoxib and naproxen showed that naproxen was associated with a reduced risk of cardiovascular events.

The cardiovascular results in VIGOR may have occurred simply by chance, given the low number of events, or because naproxen may have a cardioprotective effect similar to that of aspirin, or because rofecoxib 50 mg/day could have prothrombotic effects, especially in the absence of concomitant COX-1 inhibition in patients at increased risk of cardiovascular thromboembolic events. Because there was no untreated group in VIGOR, we do not know whether this finding suggests a protective effect of naproxen or a harmful effect of rofecoxib. All three explanations are plausible, and they are not mutually exclusive.

It has been suggested that the increase in thrombotic cardiovascular events in rofecoxib-treated patients probably represents the antiplatelet effect of naproxen [ , , ]. Naproxen has a long pharmacodynamic half-life and inhibits platelet aggregation by 88% for up to 8 hours [ ].

In VIGOR [ , ], the incidence of confirmed thrombotic cardiovascular events was 0.6% higher with rofecoxib than with naproxen (RR = 2.4; 95% CI = 3.9, 4.0).

Some data may help to answer the important question of whether the relative difference in the incidence of myocardial infarction in VIGOR was due to a harmful effect of rofecoxib or a beneficial effect of naproxen. Although contrasting data have been published [ ], the hypothesis that naproxen has a cardioprotective effect has gained wider support [ ]. Only one of the four most recently published studies negated a potential cardioprotective effect of naproxen.

However, an analysis of FDA data suggested an increase in serious cardiac events with celecoxib [ ]. The incidence of serious cardiac adverse events (myocardial infarction, combined anginal events, and atrial dysrhythmias) was 0.6% higher with celecoxib than with other NSAIDs (RR = 1.55; CI = 1.04, 2.30). The reasons for these inconsistencies are not clear.

In an 11-year observational study in new users of non-selective, non-aspirin NSAIDs (n = 181 441) and an equal number of non-users there was no evidence of a protective effect of naproxen [ ]. During 532 634 person-years of follow-up there were 6382 cases of serious coronary heart disease (11.9 per 1000 person-years). Multivariate-adjusted rate ratios for current and former use of non-aspirin NSAIDs were 1.05 (95% CI = 0.97, 1.14) and 1.02 (0.97, 1.08) respectively. Rate ratios for ibuprofen, naproxen, and other NSAIDs were 1.15 (1.02, 1.28), 0.95 (0.82, 1.09), and 1.03 (0.92, 1.16) respectively. There was no protection in long-term users with uninterrupted use; the rate ratio among current users with more than 60 days of continuous use was 1.05 (0.91, 1.21). When naproxen was directly compared with ibuprofen, the rate ratio in current users was 0.83 (0.69, 0.98). This study therefore seems to have shown no cardioprotective effect of naproxen. However, the study had a number of important limitations, including: lack of information about some important confounders (smoking, obesity), possible exposure misclassification, and lack of information about over-the-counter use of aspirin.

Opposite evidence has emerged from three case-control studies from the USA, Canada, and the UK, which showed that the rates of myocardial infarction in patients taking naproxen were lower than in patients not taking any NSAIDs [ , ] and those taking other NSAIDs [ ].

In the first study, 4425 patients hospitalized with acute myocardial infarction who used NSAIDs were compared with 17 700 controls in a large health-care database in the USA [ ]. Multivariate models were constructed to control for potential confounders. A quarter of the cases and controls had also filled a prescription for an NSAID in the 6 months before the study. Overall, the NSAID users had the same risk of acute myocardial infarction as non-users, but naproxen users had a significantly lower risk of acute myocardial infarction compared with those who were not taking NSAIDs (adjusted OR = 0.84; 95% CI = 0.72, 0.98). The cardioprotective effect of naproxen was very modest compared with aspirin (a 44% reduction in the risk of acute myocardial infarction in the Physician Health Study) [ ].

The second study was a case-control study sponsored by Merck & Co (the manufacturers of celecoxib), in which the risk of acute thromboembolic cardiovascular events among 16 937 patients aged 40–75 years with rheumatoid arthritis using naproxen was examined using the British General Practice Research Database [ ]. Each patient with a first thromboembolic event (n = 809: 435 myocardial infarctions, 347 strokes, 27 sudden deaths) was matched with four controls. The results suggested that patients with rheumatoid arthritis who currently use naproxen have a significantly lower risk of thromboembolic events relative to those who have not used naproxen in the past year (RR = 0.61; 0.39, 0.94). However, the risk was not lower with previous use of naproxen, suggesting that any effect of naproxen is likely to be short-lived. Moreover, the significantly lower risk of myocardial infarction with current naproxen was not found when myocardial infractions were analysed separately (RR = 0.57; 0.3, 1.06). There was no protective effect for thromboembolic events with current use of non-naproxen NSAIDs.

The third study was also sponsored by Merck & Co and was designed to examine the association between the use of naproxen and other non-selective NSAIDs and hospitalization for acute myocardial infarction [ ]. In a database of 14 163 patients aged 65 years or older who were hospitalized for acute myocardial infarction and an equal number of age-matched controls, concurrent exposure to naproxen had a protective effect against myocardial infarction compared with the other non-selective non-aspirin NSAIDs (RR = 0.79; CI = 0.63, 0.99). This effect was present only with concurrent naproxen exposure and was stronger in long-term users. However, this study also had several limitations: some important risk factors, such as smoking and obesity, could not be assessed, patients who died of myocardial infarction before reaching the hospital were not included, and there was uncertainty about concurrent use of over-the-counter drugs, especially aspirin.

Unexpectedly, in September 2004, Merck Sharp & Dohme announced the voluntary worldwide withdrawal of rofecoxib. This followed the release of data from a long-term, randomized, placebo-controlled trial, the Adenomatous Polyp Prevention on Vioxx the APPROVe trial, which showed that the use of rofecoxib was associated with an increased risk of thrombotic events [ , ]. The study was designed to evaluate the hypothesis that treatment with rofecoxib for 3 years would reduce the risk of recurrent adenomatous polyps among patients with a history of colorectal adenomas. A total of 2586 patients underwent randomization (1287 were assigned to rofecoxib 25 mg/day and 1299 to placebo). All investigator-reported serious adverse events that represented potential thrombotic cardiovascular events were adjudicated in a blinded fashion by an independent committee, and all the safety data were monitored by an external safety monitoring committee. The mean duration of treatment was 2.4 years with rofecoxib and 2.6 years with placebo. In the rofecoxib group 46 patients had confirmed thrombotic events during 3059 patient-years of follow-up (1.50 events per 100 patient years) compared with 26 patients who took placebo during 3327 patient-years of follow up (0.78 events per 100 patient years). The relative risk was 1.92 (95% CI = 1.19, 3.11). However, the increased relative risk became apparent only after 18 months of treatment; during the first 18 months the events rates were similar in the two groups. The difference between the two groups was mainly due to an increased number of myocardial infarctions and strokes in those taking rofecoxib.

Those who took rofecoxib also had higher risks of non-adjudicated cardiovascular events (hypertension and edema-related events) compared with those who took placebo. The Kaplan–Meier curves for the cumulative incidence of congestive heart failure, pulmonary edema, and cardiac failure showed an early separation of the two groups, at about 5 months.

Further data consistent with the results of APPROVe were found in a standard and cumulative random effects meta-analysis [ ]. This study aimed at establishing whether robust evidence on the adverse cardiovascular effects of rofecoxib had been available before September 2004, the date on which rofecoxib was withdrawn. The meta-analysis identified 18 randomized controlled trials, including 25 273 patients with chronic musculoskeletal disorders, in which rofecoxib was compared with other NSAIDs or placebo, and 11 observational studies on naproxen and cardiovascular risk. Myocardial infarction was the primary end-point. By the end of 2000 (52 myocardial infarctions, 20 742 patients) the relative risk in randomized controlled trials was 2.30 (95% CI = 1.22, 4.33), and 1 year later (64 events, 21 432 patients) it was 2.24 (1.24, 4.02). There was no evidence that the relative risk differed depending on the control group (placebo, a non-naproxen NSAID, or naproxen) or the duration of the trial. In observational studies, the cardioprotective effect of naproxen was small (combined estimate 0.86; CI = 0.75, 0.99) and could not have explained the findings of VIGOR. These findings suggest that evidence of the adverse cardiovascular effects was available before September 2004 and that rofecoxib should have been withdrawn several years earlier. The reasons why the manufacturers and regulatory authorities did not continuously monitor and summarize the accumulating evidence need to be clarified [ ].

Besides the VIGOR and APPROVe studies, a number of other clinical and epidemiological studies have signalled an increased risk of harmful cardiovascular effects with rofecoxib. Among them three epidemiological studies merit attention.

The first study [ ] was a retrospective cohort study of individuals on the expanded Tennessee Medicaid programme, in which the occurrence of serious coronary heart disease was assessed both in non-users (n = 202 916) and users of rofecoxib (n = 24 132) and other NSAIDs (n = 15 728). The patients were aged 50–84 years, lived in the community, and had no life-threatening non-cardiovascular illnesses. The incidence of acute myocardial infarction or cardiac death in new users of rofecoxib at doses above 25 mg/day was almost twice that in non-users (24 versus 13 events per 1000 patient years; RR = 1.93; 95% CI = 1.00, 3.43). In contrast, there was no evidence of an increased risk of coronary heart disease among users of rofecoxib at doses of 25 mg/day or less or among users of other NSAIDs.

The second study was a matched case-control study of the relation between coxibs, non-selective NSAIDs, and hospitalization for acute myocardial infarction in a large population of patients aged 65 years or over [ ]. The study database contained information on more than 50 000 patients and included 10 895 cases of acute myocardial infarction. Current use of rofecoxib was associated with an increased relative risk of acute myocardial infarction compared with celecoxib (OR = 1.24; CI = 1.05, 1.46) and with no NSAIDs (OR = 1.14; CI = 1.00, 1.31). A dosage of rofecoxib of over 25 mg/day was associated with a higher risk than 25 mg/day or less. The risk was increased in the first 90 days but not thereafter.

The timing of cardiovascular risks associated with the use of COX-2 inhibitors is unclear. The APPROVe trial reported a two-fold increase in cardiovascular toxicity after 18 months of use, while in the VIGOR study a similar increase in risk was reported after only 9 months of treatment. However, the risk curves in the VIGOR study started to diverge after the first month of therapy [ ]. In the adenoma prevention with celecoxib (APC) trial celecoxib was associated with a dose-dependent increase in risk after 3 years of use, but the combination of intravenous parecoxib plus oral valdecoxib resulted in an increased risk after only 10 days of exposure. A more recent study has provided new insights into the timing of the cardiovascular risks associated with the use of rofecoxib and celecoxib [ ]. The study, which was a time-matched, nested, case-control study, showed that among elderly users of rofecoxib and celecoxib the cardiovascular risks associated with the use of rofecoxib were more acute than has previously been recognized. The risk was highest in the first 6–13 days (median 9 days) after starting treatment and did not increase with further treatment. Indeed, the risk of myocardial infarction appeared to reduce over time despite continued exposure, presumably owing to attrition of susceptible individuals, i.e. the healthy survivor effect that is seen in an adverse effect of intermediate time course [ ].

Further evidence that rofecoxib increases the risk of serious coronary heart disease comes from a third case-control study commissioned by the US Food and Drug Administration [ ]. Cases of serious coronary heart disease (acute myocardial infarction and sudden cardiac death) in a cohort of NSAID-treated patients were risk-matched with four controls. Current exposure to coxibs (rofecoxib and celecoxib) and standard NSAIDs was compared with remote exposure to any NSAID, and rofecoxib was compared with celecoxib. During 2 302 029 person-years of follow up, there were 8143 cases of serious coronary heart disease, of which 2210 were fatal. Multivariate adjusted odds ratios for rofecoxib versus celecoxib were: 1.59 (all doses; CI = 1.10, 2.32); 1.47 (rofecoxib 25 mg/day or less; CI = 0.99, 2.17); and 3.58 (rofecoxib over 25 mg/day; CI = 1.00, 4.30). For naproxen versus remote NSAIDs the adjusted OR was 1.14 (1.00, 1.30). The interpretation of these results was that rofecoxib increases the risk of acute myocardial infarction and sudden death compared with celecoxib and that naproxen does not protect against serious coronary heart disease. This study also provided data relevant to some other controversies about the cardiovascular safety of coxibs. In particular there was a substantially higher risk with high-dose rofecoxib, in accordance with the results of other studies [ , ]. Moreover the mean duration of use before the occurrence of an event was identical with high doses and standard doses of rofecoxib (about 110 days), consistent with the idea that the risk of cardiovascular toxicity begins early in treatment. That is in accord with the analysis by the FDA of data from VIGOR, which showed that the survival curve for acute myocardial infarction with high-dose rofecoxib began to diverge from the naproxen curve after 1 month. There was no evidence that the relative risk differed depending on the control group (placebo, a non-naproxen NSAID, or naproxen) or the duration of treatment. This suggests that patients are at risk of myocardial infarction even if rofecoxib is taken for a few months only. Therefore, the reassuring statement by Merck that there is no excess risk in the first 18 months is not supported by this meta-analysis.

In contrast to these findings, two earlier meta-analyses from Merck laboratories showed no evidence for an excess of cardiovascular thrombotic events for rofecoxib relative to either placebo or non-naproxen NSAIDs (ibuprofen, nabumetone, diclofenac) or an increased risk in trials in which rofecoxib and naproxen were compared [ , ].

In conclusion these data provide further convincing evidence that naproxen does not have a cardioprotective effect, in accordance with the results of some studies [ , ] although not others [ ]. Indeed, the current data show even a possibility of a small increased risk of coronary heart disease.

The demonstration of a lack of protective effect of naproxen is important, because it is often used as a comparator in clinical trials of new coxibs. Thus, a result that shows that a new drug has an increased risk of cardiovascular disease relative to naproxen should alert prescribers to potential cardiotoxic effects.

There are two further crucial questions:

Is the apparent treatment-associated increase in severe cardiovascular events unique to rofecoxib or does it apply to other COX-2 selective drugs, suggesting a COX-2 class effect, and even to NSAIDs in general [ ]?
Does it apply to all patients or only to a subset, such as those with known susceptibility factors for vascular disease?

Two recently published reports help in answering these questions. The first is a meta-analysis of randomized controlled trials of selective COX-2 inhibitors versus placebo or traditional NSAIDs or both [ ]. Eligible studies were randomized trials that lasted at least 4 weeks, with information on serious vascular events (defined as myocardial infarction, stroke, and vascular death). Data were available from 138 randomized trials involving a total of 143 373 participants. NSAIDs were subdivided into naproxen and others. In placebo comparisons, allocation to a COX-2 selective inhibitor was associated with a 42% relative increase in the incidence of serious vascular events: 1.2% per year versus 0.9% per year (RR = 1.42; 95% CI = 1.13, 1.78) with no significant heterogeneity among the different coxibs. This increase was chiefly attributable to an increased risk of myocardial infarction (0.6% per year versus 0.3% per year; RR = 1.86; 95% CI = 1.93, 2.59), with little apparent difference in other vascular outcomes. Among trials of at least 1 year duration (mean 2.7 years) the rate ratio for vascular events was 1.45 (CI = 1.12, 1.89). There were too few vascular events to allow an assessment of the dose-response relation in placebo-controlled trials of all coxibs, with the exception of celecoxib, for which there was a significant trend towards an increased incidence of serious vascular events with higher daily doses. The size of the risk of vascular events in placebo-controlled trials that allowed concomitant use of aspirin was similar among aspirin users and non-users.

Comparisons of selective COX-2 inhibitors versus traditional NSAIDs showed a similar incidence of serious vascular events (1.0% per year versus 0.9% per year; RR = 1.16; CI = 0.97, 138). However, there was marked heterogeneity across the rate ratios for vascular events in trials that compared a coxib with naproxen and trials that compared a selective COX-2 inhibitor with a non-naproxen NSAID. Overall, compared with naproxen, allocation to a selective COX-2 inhibitor was associated with a highly significant increase in the incidence of a serious vascular event (RR = 1.57; CI = 1.21, 2.03) and a twofold increased risk of a myocardial infarction (RR = 2.04; CI = 1.41, 2.96), but no difference in the incidences of stroke or vascular death. In contrast the comparison of COX-2 selective inhibitors with non-naproxen NSAIDs showed non-significant differences in the incidences of vascular events, myocardial infarctions, and vascular deaths, but selective COX-2 inhibitors were associated with a significantly lower incidence of stroke than any non-naproxen NSAIDs (RR = 0.62; CI = 0.41, 0.95).

The time-course of the effects of NSAIDs in increasing the risk of cardiovascular events has been assessed in a cohort study in 74 838 users of non-selective NSAIDs or coxibs and 23 532 comparable users of other drugs [ ]. There was a significant increase in the event rate for rofecoxib (RR = 1.15; 95% CI = 1.06, 1.25) and a significant reduction in the rate for naproxen (RR = 0.75; 95% CI = 0.62, 0.92). No other coxib or NSAID had an effect. The increased rate associated with rofecoxib was seen in the first 60 days of use and persisted thereafter.

When the rate ratio of serious vascular events with traditional NSAIDs was compared with that of placebo, high-dose ibuprofen and high-dose diclofenac were both associated with an increased risk of cardiovascular events (ibuprofen RR = 1.51; CI = 0.96, 2.37; diclofenac RR = 1.63; CI = 1.12, 237), while naproxen was not.

Thus, it appears that allocation to a selective COX-2 inhibitor is associated with about three extra vascular events per 1000 patients per year. Most of this excess is attributable to myocardial infarction.

The other study that has contributed to better evaluation of the data on cardiovascular risk due to inhibition of COX is a systematic review of the available controlled pharmacoepidemiological studies [ ]. There were 23 eligible studies that reported on cardiovascular events (predominantly myocardial infarction) with COX-2 inhibitors, NSAIDs, or both, with non-use/remote use of the drug as the reference exposure for calculation of the relative risk. The studies were 17 case-control studies involving 86 193 cases with cardiovascular events and about 52 800 controls and six cohort studies involving 75 520 users of selective COX-2 inhibitors, 375 619 users of non-selective NSAIDs, and 594 720 unexposed participants. The results confirmed that there is an increased risk of cardiovascular events with rofecoxib. The summary relative risk with doses in excess of 25 mg/day was 2.19 (CI = 1.64, 2.91) compared with 1.33 (CI = 1.00, 1.79) with 25 mg/day or less. The risk was increased during the first month of treatment. In contrast to the evidence from the meta-analysis of randomized trials, this study showed no increase in risk with celecoxib 200 mg/day. However, one must remember that the randomized clinical trials only showed an increased risk with doses of celecoxib of 400 mg/day and above. The study also provided more data on the cardiovascular toxicity of traditional NSAIDs. Naproxen was not associated with a reduction in risk, as was suggested in previous comparisons with rofecoxib.

Of more concern is evidence, from both randomized and non-randomized trials, that diclofenac increases the risk of cardiovascular events (RR = 1.63; CI = 1.12, 2.37). Less convincing is the evidence of potential cardiovascular risk with other NSAIDs, such as ibuprofen, meloxicam, and indometacin.

Celecoxib

In December 2004 the results of the APC trial were presented. This was a randomized, double-blind, multicenter study of the effects of two doses of celecoxib (200 mg or 400 mg bd) or placebo in the prevention of colorectal adenomas [ ]. All potentially serious cardiovascular events among 2035 patients with a history of colorectal neoplasia were identified, with a follow up of 2.8–3.1 years. Celecoxib was associated with a dose-related increase in the composite end-point of death from cardiovascular causes, myocardial infarction, stroke, or heart failure. The composite end-point was reached in seven of 679 patients in the placebo group (1%) compared with 16 of 685 patients who took celecoxib 200 mg bd (2.3%; hazard ratio = 2.3; CI = 0.9, 5.5) and 23 of 671 patients who took celecoxib 400 mg bd (3.4%; hazard ratio = 3.4; CI = 1.4, 7.8). The annualized incidence of death due to cardiovascular causes per 1000 patient years was 3.4 events in the placebo group; there were 7.5 events in patients who took celecoxib 200 mg bd and 11.4 in those who took 400 mg bd. The hazard ratio associated with celecoxib was not significantly affected by any baseline characteristics, including use of aspirin. Moreover, the results were also consistent among the individual components of the composite end-point. Based on these statistically significant findings, the sponsor of the study, the USA National Cancer Institute suspended the study.

However, to underscore the uncertainty of the available data on cardiovascular risk, we should consider the results of another large long-term trial with celecoxib 400 mg/day, designed to investigate the role of NSAIDs in preventing Alzheimer’s disease, the Alzheimer Disease Anti-inflammatory Prevention Trial (ADAPT), in which celecoxib was compared with naproxen (440 mg/day). The unpublished results of the study showed that celecoxib did not increase cardiovascular risk, while naproxen was associated with an increased risk [ ]. The results from the APC and ADAPT studies are disturbing, because they are inconsistent with the results of most previous studies. The following data summarize the currently available information on the cardiovascular safety of celecoxib.

In two retrospective re-analyses of randomized controlled trial data for celecoxib [ , ] the cardiovascular events in patients enrolled in the CLASS study [ ] were examined, as were those reported across the entire controlled arthritis clinical trial database for celecoxib. The results of the reanalysis of CLASS showed no evidence of an increase in investigator-reported serious cardiovascular adverse events in patients taking celecoxib. In the analysis of the celecoxib comparative trials database the incidence of cardiovascular events was not significantly different between celecoxib and placebo or between celecoxib and naproxen, regardless of aspirin use. Thus, these comparative analyses failed to show an increased risk of thrombotic events associated with celecoxib compared with conventional NSAIDs, naproxen specifically, or placebo.

In a meta-analysis of studies in 7462 patients exposed to (a) celecoxib 200–800 mg/day for 1268 patient-years compared with 4057 patients treated with placebo for 585 patient-years and (b) 19 773 patients treated with celecoxib 200–800 mg/day for 5651 patient-years compared with 13 990 patients treated with non-selective NSAIDs (diclofenac, ibuprofen, naproxen, ketoprofen, and loxoprofen) for 4386 patient-years, the incidence rates of the combined cardiovascular events were not significantly different between patients treated with celecoxib and placebo or between those treated with celecoxib and non-selective NSAIDs [ ]. The dose of celecoxib, the use of aspirin, or the presence of cardiovascular risk factors did not alter these results. The authors concluded that there was no evidence of an increased cardiovascular risk with celecoxib relative to placebo and a comparable rate of cardiovascular events with celecoxib compared with non-selective NSAIDs.

Most information on the potential cardiovascular toxicity of celecoxib has come from a number of observational studies [ , , , ]. Most of them were population-based case-control studies in which the relative risk of acute myocardial infarction was assessed in patients who took celecoxib, rofecoxib, conventional NSAIDs, or no NSAIDs at all. None of these studies identified a significantly increased risk of severe cardiovascular events with celecoxib. However, the results of these observational studies must be interpreted with caution. In fact, although findings from observational studies are more generalizable, as they are larger and include less carefully selected patients than randomized controlled trials, they suffer from potential bias and confounding, which could have contributed to the failure to confirm the hypothesis of a consistent difference in severe cardiovascular risk between all the coxibs and other NSAIDs. Whether this safety concern represents a class effect requires additional information from other randomized controlled trials, not yet available [ ].

The risk of cardiovascular events with celecoxib has been studied in a systematic review and meta-analysis of randomized double-blind trials of at least 6 weeks′ duration with data on serious thromboembolic events [ ]. Four placebo-controlled trials in 4422 patients were analysed. The odds ratio for myocardial infarction with celecoxib was 2.26 (95% CI = 1.0, 5.1). There was no significant increase in risk with celecoxib for composite cardiovascular events, cardiovascular deaths, or stroke. A secondary meta-analysis of six studies with placebo, diclofenac, ibuprofen, and paracetamol as comparators in 12 780 patients showed similar findings: there was a significantly increased risk of myocardial infarction with celecoxib (OR = 1.88; 95% CI = 1.15, 3.08) but not other outcome measures.

In a randomized comparison of celecoxib and celecoxib + diflouromethylornithine in 112 subjects in the prevention of colorectal adenomas there were no adverse cardiovascular outcomes in either arm, too small a study to be helpful [ ].

In a study of the use of crowdsourcing to determine relative intensity of adverse drug reactions, celecoxib emerged as an important cause of cardiovascular morbidity [ ].

Etoricoxib

Data on the efficacy and safety of this highly selective COX-2 inhibitor are still limited. A combined analysis of all randomized, double-blind trials of long-term treatment showed a significantly lower incidence of peptic ulcer bleeding with etoricoxib (n = 9226) than with conventional NSAIDs (n = 2215), but there was no information about potential cardiovascular toxicity [ ].

There has been a systematic review and meta-analysis of double-blind, randomized, placebo-controlled trials of etoricoxib of at least 6 weeks duration (five studies with a total of 2919 patients); the main outcome measure was cardiovascular thromboembolic events [ ]. There were seven cardiovascular thromboembolic events in 1441 patients (0.5%) taking etoricoxib, and one event in 906 patients (0.1%) taking placebo. The pooled fixed-effects estimate of the absolute risk difference was 0.5% (CI = 0.1, 1.0). The odds ratio for the risk of cardiovascular events was 1.49 (0.42, 5.31). These limited data provide weak evidence of an increased cardiovascular risk with etoricoxib.

In a prespecified pooled analysis of data from three trials in which patients with osteoarthritis or rheumatoid arthritis were randomly assigned to etoricoxib 60 or 90 mg/day or diclofenac 150 mg/day for an average of 18 months (SD 11.8), there were equal numbers of thrombotic cardiovascular events with etoricoxib (n = 320) and diclofenac(n = 323), giving event rates of 1.24 and 1.30 per 100 patient-years and a hazard ratio of 0.95 (95% CI = 0.81, 1.11) for etoricoxib compared with diclofenac [ ]. Rates of upper gastrointestinal events (perforation, bleeding, obstruction, ulcer) were lower with etoricoxib (0.67 versus 0.97 per 100 patient-years; HR = 0.69; 95% CI = 0.57, 0.83), but the rates of complicated upper gastrointestinal events were similar (0.30 versus 0.32).

Lumiracoxib

Lumiracoxib is a novel COX-2 selective inhibitor, a phenylacetic acid derivative with a short half-life and a higher selectivity than any other coxib. Clinical data on its efficacy and safety are still limited.

The cardiovascular safety of lumiracoxib has been assessed in the Therapeutic Arthritis and Gastrointestinal Event Trial (TARGET) in 18 325 patients aged 50 years or older with osteoarthritis, who were randomized to lumiracoxib (400 mg/day; n = 9156), ibuprofen (800 mg tds; n = 4415), or naproxen (500 mg bd; n = 4754) for 52 weeks in two substudies of identical design (lumiracoxib versus ibuprofen and lumiracoxib versus naproxen) [ ]. Randomization was stratified for low-dose aspirin and age. The primary cardiovascular end points were the ATC end-points of non-fatal and silent myocardial infarction, stroke, or cardiovascular death. At 1 year of follow up, the incidence of the primary end-point was low and did not differ between treatment groups: lumiracoxib 59 events (0.65%), NSAIDs 50 events (0.55%; hazard ratio 1.14; CI = 0.28, 1.66). The incidence of myocardial infarction in the overall population in the individual substudies was 0.38% with lumiracoxib (18 events) versus 0.21% with naproxen (10 events), and 0.11% with lumiracoxib (5 events) versus 0.16% (7 events) with ibuprofen. In both substudies the rates of myocardial infarction did not differ significantly between lumiracoxib and NSAIDs, irrespective of aspirin use.

These TARGET trial data suggest that lumiracoxib does not have the potential to precipitate adverse cardiovascular events more often than NSAIDs. However, these results must be interpreted with caution [ ] for several reasons:

patients with known and significant pre-existing coronary artery disease were excluded, which explains the overall low frequency of cardiovascular events and introduced bias into the generalizability of the TARGET results;
the statistical power was inadequate to detect significant differences in the rates of myocardial infarction, again raising concern about an excess, albeit not a statistically significant one, of myocardial infarctions with lumiracoxib compared with naproxen: 18 events (0.38%) versus 10 events (0.21%); hazard ratio 1.77 (0.82, 3.84).
lumiracoxib significantly reduced the frequency of upper gastrointestinal ulcer complications (NNT to prevent one ulcer complication 139) but only in patients not taking low-dose aspirin [ ], confirming the results of other coxib trials (CLASS).
the data raised a concern about possible hepatotoxicity of lumiracoxib; the proportion of patients with aminotransferase activities more than three times the upper limit of the reference range differed significantly between lumiracoxib (2.57%, n = 230) and NSAIDs (0.63%, n = 56; hazard ratio 3.97; CI = 2.96, 5.32).

Valdecoxib and parecoxib

Further support of the hypothesis that the increase in cardiovascular risk is a COX-2 class effect has come from an analysis of clinical trials of valdecoxib and its prodrug parecoxib.

The cardiovascular safety of valdecoxib was initially assessed in a study that pooled results from 10 clinical trials that included nearly 8000 patients with osteoarthritis and rheumatoid arthritis and compared the incidence of cardiovascular events in patients taking valdecoxib (10–80 mg/day) with those of controls taking diclofenac, ibuprofen, naproxen, or placebo [ ]. The incidences of cardiovascular thrombotic events (cardiac, cerebrovascular, and peripheral vascular or arterial thrombotic) were similar with valdecoxib, the conventional NSAIDs, and placebo. Short-term and intermediate-term treatment with valdecoxib in therapeutic doses (10 or 20 mg/day) and supratherapeutic doses (40–80 mg/day) was not associated with an increased incidence of thrombotic events. In contrast, recent studies have raise serious doubts about the safety of valdecoxib in patients who are at high risk of thrombotic complications, such those undergoing coronary artery bypass surgery.

The cardiovascular toxicity of valdecoxib and its pro-drug parecoxib has been studied in two randomized placebo-controlled trial and one meta-analysis. The first study involved 462 patients and evaluated the safety and efficacy of intravenous parecoxib (parecoxib 40 mg intravenously every 12 hours for 3 days postoperatively), followed by oral valdecoxib (40 mg every 12 hours for a total of 14 days) [ ]. In the second study, which involved more than 1600 patients, a similar schedule of administration was used, but the dosage and duration of treatment were reduced (20 mg bd for 10 days) [ ]. In both studies there were clusters of cardiovascular adverse events, including myocardial infarction, stroke, deep vein thrombosis, and pulmonary embolism, which were more frequent in those who were given parecoxib than in the controls, although the difference was not statistically significant. However, when the coronary and cerebrovascular events were combined in a meta-analysis parecoxib/valdecoxib was associated with a three-fold risk of cardiovascular events compared with placebo: 31/1399 events versus 5/699 (RR = 3.08; CI = 1.20, 7.87) [ ].

Following this important information a new warning contraindicating the use of parecoxib/valdecoxib in patients undergoing coronary artery bypass grafting was added to the label by many drug control agencies. Meanwhile the FDA received reports of 87 cases of severe skin reactions associated with valdecoxib, including Stevens–Johnson syndrome and toxic epidermal necrolysis, with four deaths; 20 of the 87 cases involved patients with a known allergy to sulfa-containing drugs, of which valdecoxib is one; thus, patients allergic to sulfa drugs may be at greater risk of severe skin reactions with valdecoxib. These data have raised concerns not only about the cardiovascular safety of valdecoxib in the general population but also about its overall benefit to harm profile. The drug manufacturers and regulatory agencies in the USA and Europe agreed to suspend valdecoxib in April 2005 [ ].

Conclusions

At present at least three important questions about the adverse cardiovascular effects of COX-2 selective inhibitors remain unanswered.

First, the mechanism whereby COX-2 inhibitors facilitate ischemic cardiovascular events is unclear. The major unanswered question is whether unopposed COX-2 inhibition or other drug-specific mechanisms cause the increased cardiovascular risk. It is worth noting that in addition to ischemic cardiovascular disease, COX-2 inhibitors are associated with other adverse cardiovascular effects, such as heart failure [ , ], hypertension [ , ], and edema [ ].

Secondly, it is still unknown whether the cardiovascular risk is or is not a class effect. Since at least three separate drugs in the class (rofecoxib, valdecoxib, and celecoxib) have now been associated with increased cardiovascular morbidity, the burden of proof has been shifted to those who deny a class effect. We must remember that absence of evidence is not evidence of absence [ ].

Thirdly, there is uncertainty about what physicians should do if they decide to prescribe an NSAID. In light of the current uncertainty about whether cardiotoxicity is a class effect, coxibs, in particular at high dosages and for long-term use, should not be prescribed, particularly in populations at high risk, such as elderly patients and those with established cardiovascular disease. This recommendation is fully justified in light of the possible consequences of the current heavy unjustified promotion of these compounds to both patients and prescribers. A recent epidemiological study in Ontario, Canada showed that the use of NSAIDs in patients aged 66 years or older increased by 41% after the introduction of coxibs [ ]. This rise was entirely attributable to the use of coxibs and was accompanied by a 10% increase in hospitalization for upper gastrointestinal bleeding.

The continued commercial availability of coxibs is therefore troubling, and probably unjustified, in the light of their marginal efficacy, heightened risk, and higher costs compared with traditional NSAIDs.

Hypertension

The shift in hemostatic balance toward a prothrombotic state might not be the only mechanism by which COX-2 inhibitors could increase the risk of cardiovascular adverse effects. In fact, non-selective NSAIDs can raise blood pressure and antagonize the hypotensive effect of antihypertensive medications to an extent that may increase hypertension-related morbidity [ , ]. The problem is clinically relevant, as arthritis and hypertension are common co-morbid conditions in elderly people, requiring concurrent therapy.

Information on the effect of COX-2-selective inhibitors on arterial blood pressure is scanty. In VIGOR, more patients developed hypertension with rofecoxib than naproxen. For rofecoxib, the mean increase in blood pressure was 4.6/1.7 mmHg compared with a 1.0/0.1 mmHg increase with naproxen [ ]. Previous work has shown that a 2 mmHg reduction in diastolic blood pressure can result in about a 40% reduction in the rate of stroke and a 25% reduction in the rate of myocardial infarction [ ]. The effect of celecoxib on blood pressure was evaluated in a post hoc analysis using the safety database generated during the celecoxib clinical development program in more than 13 000 subjects [ ]. The incidence of hypertension after celecoxib was greater than that after placebo but similar to that after non-selective NSAIDs. Hypertension and exacerbation of pre-existing hypertension occurred, respectively, in 0.8 and 0.6% of patients. Furthermore, there was no evidence of interactions between celecoxib and other antihypertensive drugs.

The safety profiles of celecoxib (200 mg/day) and rofecoxib (25 mg/day) have recently been compared in a 6-week, randomized, parallel-group, double-blind trial in 810 patients with osteoarthritis aged 65 years or over taking antihypertensive drugs [ ]. The primary end-points were edema and changes in systolic and diastolic blood pressures, measured at baseline and after 1, 2, and 6 weeks. Systolic blood pressure rose significantly in 17% of rofecoxib-treated patients (n = 399) compared with 11% of celecoxib-treated patients (n = 411), a statistically significant difference, at any time in the study. Diastolic blood pressure rose in 2.3% of rofecoxib-treated patients compared with 1.5% of celecoxib-treated patients, a non-significant difference. At week 6, the change from baseline in mean systolic blood pressure was + 2.6 mmHg for rofecoxib compared with − 0.5 mmHg for celecoxib, a highly significant difference. Nearly twice as many rofecoxib-treated than celecoxib-treated patients had edema. Despite some limitations, this study provides some evidence that COX-2-selective inhibitors may differ in their ability to alter arterial blood pressure.

Cardiac dysrhythmias

There have been reports of celecoxib-associated torsade de pointes in three patients who had never had any complaints before celecoxib administration; the dysrhythmia did not recur after the drug was withdrawn [ ]. However, all three patients had cardiac abnormalities that might have predisposed them to the development of torsade de pointes, and the follow-up period was too short to evaluate possible spontaneous recurrence. The hypothesis that celecoxib is dysrhythmogenic requires confirmation.

Respiratory

Most people tolerate aspirin well, but not patients with asthma, of whom there is a subgroup in whom aspirin precipitates asthmatic attacks [ , ]. This is a distinct clinical syndrome, called aspirin-induced asthma, which affects about 10% of adults with asthma [ ]. Aspirin-induced asthma is usually accompanied by naso-ocular symptoms and can be triggered not only by aspirin, but by several NSAIDs, a fact that makes immunological cross-reactivity most unlikely. The propensity of an NSAID to precipitate an attack of asthma is probably related to inhibition of COX. There is evidence that potent inhibitors of COX-1 (such as ibuprofen, indometacin, and naproxen) are more likely to precipitate bronchoconstriction than NSAIDs that inhibit COX-2 preferentially (such as meloxicam and nimesulide) [ , ]. A widely accepted hypothesis is that in patients with asthma and aspirin intolerance, NSAID-induced COX inhibition results in increased products from the 5-lipoxygenase pathway, the leukotrienes, which are both potent bronchoconstrictors and also inducers of mucous hypersecretion and airway edema. The leukotrienes implicated in aspirin intolerance are cysteinyl leukotrienes [ , ], but leukotriene release is probably not the only pathogenic mechanism.

The hypothesis that in aspirin-induced asthma the attacks are triggered by inhibition of COX-1 and not COX-2 has been tested in three small studies, two of which were double-blind and placebo-controlled. In the first study [ ] 12 patients with aspirin-induced asthma were challenged with increasing doses of rofecoxib (1.25–25 mg/day for 5 days); no patients had any adverse symptoms, and biochemical markers that reflect intolerance to aspirin in asthma (urinary leukotriene E ₄ and 9α-11β-PGF ₂ ) were unchanged.

In the second study, 60 aspirin-sensitive asthmatics were challenged with oral rofecoxib (12.5 and 25 mg/day for 2 days) [ ]. There were no signs or symptoms of asthma in any patient and no reductions in FEV ₁ .

In the third study, 27 patients with stable chronic asthma in whom inhalation of lysine aspirin caused a 20% fall in FEV ₁ were challenged with increasing doses of oral celecoxib (from 10 to 200 mg/day); they do not develop bronchoconstriction or other extrapulmonary reactions [ ].

So, is the safety of COX-2 inhibitors in patients with aspirin-induced asthma sufficiently well documented? Avoidance of NSAIDs is crucial in patients with asthma who have aspirin sensitivity, and hitherto two alternative analgesic options have been available: paracetamol and non-acetylated salicylates. Paracetamol is a generally safe substitute; however, it is a weak COX inhibitor and, albeit very rarely, patients who are sensitive to aspirin have an adverse reaction to high-dose paracetamol [ ]. Choline magnesium salicylate and salicylic acid are also weak COX inhibitors and should therefore be used with care in these patients. The results of the above-mentioned studies suggest that rofecoxib and celecoxib can be taken safely by patients with aspirin-induced asthma. However, this conclusion must be treated with caution, as only a few patients were studied and this does not exclude COX-2 inhibitors from participating in other types of reactions, including immune recognition after prior treatment.

You're Reading a Preview

Become a Clinical Tree membership for Full access and enjoy Unlimited articles

Become membership

If you are a member. Log in here