Cardiovascular Safety Signal Evaluation

CARDIOVASCULAR SAFETY SIGNAL EVALUATION

As results from VIGOR indicated that rofecoxib was associated with increased risk of myocardial infarction, a Food and Drug Administration (FDA) advisory committee meeting was convened in February 2001 to review gastrointestinal and cardiovascular safety data of rofecoxib and celecoxib. These data are avail-able from the FDA website (FDA, 2001) and some were subsequently published in peer-reviewed journals. In the US, revised package insert of rofecoxib with updated safety information was released in April 2002 (Kweder, 2004).

BIOLOGICAL MECHANISM

As the COX-2 inhibitors have minimal effects on COX-1 in gastric epithelium, they have more favor-able gastrointestinal safety profile than the non-selective NSAIDs. However, at the time of approval of celecoxib and rofecoxib, pharmacologic stud-ies indicated that selective inhibition of COX-2 may contribute to imbalance in the prostacyclin (prostaglandin I₂) to thromboxane ratio and result in thrombotic events (FitzGerald, 2004). COX-2 in endothelium is responsible for the production of prostacyclin, which inhibits platelet aggregation, causes vasodilatation, and prevents proliferation of vascular smooth-muscle cells. On the other hand, the COX-1 in platelets, which is not affected by the COX-2 inhibitors, is responsible for thromboxane A₂ synthesis, and thromboxane A₂ has the opposite effects of prostacyclin, causing platelet aggregation, vasoconstriction, and vascular proliferation. Theoret-ically, selective inhibition of COX-2 would allow the physiologic effects of thromboxane to predomi-nate and result in adverse cardiovascular outcomes (FitzGerald, 2003).

The VIGOR investigators cited a study in healthy volunteers that showed different effects on thrombox-ane A₂ production and platelet aggregation associated with different non-selective NSAIDs (Van Hecken et al., 2000). Ibuprofen 800 mg three times a day and sodium naproxen 550 mg twice a day in healthy volun-teers showed substantial inhibition effects on throm-boxane A₂ production and platelet aggregation, but how these pharmacologic findings translate to preven-tion of myocardial infarction by naproxen needed to be verified in clinical studies. The cardio-protective effects of naproxen would have to be very power-ful in order to result in the VIGOR findings and support the hypothesis that rofecoxib 50 mg per day was not associated with increased risk of myocardial infarction.

CLINICAL TRIALS OF CELECOXIB AND ROFECOXIB

After the results of clinical trials discussed during the February 2001 FDA advisory committee meet-ing were made public, a group of independent inves-tigators reviewed cardiovascular events observed in CLASS, VIGOR, and two unpublished clinical trials of rofecoxib (Mukherjee, Nissen and Topol, 2001). The review covered more cardiovascular data than that was published in the original CLASS and VIGOR reports in 2000 as patients contributed more follow-up time and cardiovascular events were independently adjudicated. Reviews of cardiovascular safety data in the clinical development programs of celecoxib and rofecixb were also published by manufacturers of the drugs from 2001 through 2003.

Celecoxib

No increased risk of myocardial infarction, stroke, or death was found in the celecoxib group in CLASS. The same finding was observed in patients who received aspirin and those who did not receive aspirin during the trial (Mukherjee, Nissen and Topol, 2001). The CLASS investigators carried out a safety analy-sis and evaluated the risk of cardiovascular outcomes, cerebrovascular outcomes, and peripheral vascular outcomes (White et al., 2002). No increased risk of thromboembolic events was found among the cele-coxib group (Table 47.1). White and colleagues combined data from 15 trials of celecoxib and used the Anti-Platelet Trialists’ Collaboration (APTC) defini-tion of adverse cardiovascular outcomes (Antiplatelet Trialists’ Collaboration, 1994), including cardiovas-cular, hemorrhagic and unknown death, myocardial infarction, and cerebrovascular accident, as the end point of interest (White et al., 2003). A relative risk of 1.06 (95% CI, 0.70–1.61) was found for the celecoxib (all doses) versus non-selective NSAID comparison.

Rofecoxib

Two unpublished studies, 085-2001 and 090-2001, of rofecoxib reported by Mukherjee and colleagues in 2001 were three-arm trials of rofecoxib 12.5 mg per day, nabumetone 1000 mg per day, and placebo for six weeks among patients with osteoarthritis. Using low dose aspirin was not an exclusion criterion in both trials. No difference in cardiovascular events was found between the rofecoxib- and nabumetone-treated groups in 085-2001 and there was a non-significant increase in the incidence of cardiovascular events in the rofecoxib versus nabumetone comparison (1.5% versus, 0.5%) in 090-2001.

For VIGOR, Mukherjee and colleagues reported incidence of serious thrombotic cardiovascular events, including myocardial infarction, unstable angina, cardiac thrombus, resuscitated cardiac arrest, sudden or unexplained death, ischemic stroke, and transient ischemic attacks in the treatment groups (Mukherjee, Nissen and Topol, 2001). In addition, more adju-dicated cases of serious cardiovascular events were reported at the February 2001 FDA advisory commit-tee, which did not appear in the original VIGOR publication (Curfman, Morrissey and Drazen, 2005). The VIGOR study protocol stated that history of cere-brovascular events in the two years before the study, history of myocardial infarction or coronary bypass in the year before the study, and requirement for or ongo-ing treatment with aspirin, ticlopidine, or anticoagu-lants were exclusion criteria. However, ‘requirement for aspirin treatment’ was not objectively defined and was subjectively assessed by the physicians who enrolled the patients. Mukherjee and colleagues reported a post-hoc stratification of study subjects according to cardiovascular indication for aspirin use, which was operationally defined as prior history of stroke, transient ischemic attack, myocardial infarc-tion, unstable angina, angina pectoris, coronary artery bypass graft surgery, or percutaneous coronary inter-ventions. Only 321 (4%) of the 8076 enrolled subjects were aspirin-indicated and no myocardial infarction occurred among aspirin indicated patients. Relative risk of serious cardiovascular events was 4.89 (95% CI, 1.41–16.88) among the rofecoxib (50 mg per day) group in the full cohorts and for the non-aspirin indi-cated subjects, the relative risk was 1.89 (95% CI, 1.03–3.45).

Scientists from the manufacturer of rofecoxib and academic investigators analyzed the safety database of clinical trials of rofecoxib and published three safety reports after VIGOR. In the first report, Konstam and colleagues reviewed combined data of more than 28 000 patients from 23 trials and used the APTC defi-nition for cardiovascular end point (Konstam et al., 2001). Rofecoxib was associated with increased risk of APTC events when compared with naproxen (rela-tive risk, 1.69, 95% CI, 1.07–2.69) but there was no increased risk when rofecoxib was compared with placebo or non-naproxen NSAIDs. Stratifying the rofecoxib group by dose and compared with all NSAIDs, relative risk was 2.08 (95% CI, 0.57, 7.51) for rofecoxib 50 mg per day and was 1.16 (95% CI, 0.25, 7.18) for rofecoxib 25 mg per day. Reicin and colleagues reported results from pooled data from eight phase IIB or III trials of rofecoxib, ibuprofen, diclofenac, nabumetone, or placebo for osteoarthri-tis, with a total of 5435 patients (Reicin et al., 2002). Using the APTC end point and any arterial or venous thrombotic cardiovascular adverse event as the outcome of interest, no significantly increased cardiovascular risk was observed in the rofecoxib– non-selective-NSAID and rofecoxib–placebo compar-isons. However, statistical power was limited to detect a modest increase in cardiovascular risk among a dataset of this size and these trials were short-term studies with no information on potential effects among long-term use. In 2003, Weir and colleagues reported safety data from the rofecoxib development program that included short-term trials, long-term studies like VIGOR, and placebo-controlled studies for the prevention of Alzheimer’s disease (Weir et al., 2003). The results confirmed the increased cardio-vascular risk for the rofecoxib–naproxen comparison, with a relative risk of 1.61 (95% CI, 1.04–2.50) in pooled analysis for trials among patients with rheuma-toid arthritis and/or osteoarthritis. No increased cardiovascular risk was found for the rofecoxib– non-naproxen-NSAID comparison or the rofecoxib– placebo comparison. These data pooling exercises were helpful in the safety assessment of drugs, but the heterogeneity of the study populations and treat-ment periods are major limitations. In the report by Weir and colleagues, there were no stratified results based on dose and duration of use for rofecoxib.

CARDIOPROTECTIVE EFFECTS OF NAPROXEN IN EPIDEMIOLOGY STUDIES

While there has been pharmacologic basis for cardio-protective effects of naproxen (Van Hecken et al., 2000), there has been no reported clinical trial that specifically evaluated the risk of adverse cardiovas-cular outcomes in a naproxen versus placebo compar-ison before VIGOR or through 2005. The VIGOR results prompted more than 10 reports of observa-tional studies that evaluated the association between naproxen use and myocardial infarction, all of them were included in a meta-analysis conducted by Jüni and colleagues (Jüni et al., 2004). Reduced risk of myocardial infarction associated with naproxen was reported in three case-control studies; reduced but non-significant risk was reported in four studies; and small increased risk associated with naproxen was reported in three studies. All these data taken together suggested a small reduced risk, on the order of 15%, of myocardial infarction associated with naproxen use. However, these findings did not support the hypothesis that rofecoxib 50 mg per day did not increase risk of myocardial infarc-tion, as the small reduction in risk associated with naproxen use could not adequately explain the more than fourfold increase in risk of myocardial infarc-tion in the rofecoxib–naproxen comparison observed in VIGOR.

OBSERVATIONAL STUDIES OF CELECOXIB AND ROFECOXIB

After VIGOR, important information about population-based cardiovascular risk associated with rofecoxib and celecoxib came from large observa-tional studies with sufficient numbers of COX-2 inhibitors users in a real life setting. Ray and colleagues reported that new users of high dose rofecoxib (more than 25 mg per day) had an almost twofold increase in the risk of hospital admission for acute myocardial infarction or death from coronary heart disease when compared with those who did not receive any non-selective NSAID or COX-2 inhibitor (Ray et al., 2002; Table 47.2). No statistically significant increased risk was observed among new users of lower dose rofecoxib (25 mg or less per day), celecoxib, naproxen, or ibuprofen.

Three other observational studies, one each from Canada, the UK, and the US, were published before the market withdrawal of rofecoxib in September 2004 and are summarized in Table 47.2. Mamdani compared new users of COX-2 inhibitors or non-selective NSAIDs with subjects who did not use any COX-2 inhibitors or non-selective NSAID and reported no increased risk in acute myocardial infarc-tion in new users of celecoxib or rofecoxib (Mamdani et al., 2003).

The UK study was based on the Prescription-Event Monitoring system and was reported as two companion articles in the same journal in 2003 – one was a comparison between celecoxib and meloxicam (Layton et al., 2003a) and the other was a compari-son between rofecoxib and meloxicam (Layton et al., 2003b). Outcomes of interest were cardiovascular, cerebrovascular, and peripheral venous thrombotic events. Comparing with meloxicam, a preferential but not selective COX-2 inhibitor, and only adjusted for age and sex, celecoxib and rofecoxib were both associated with increased risk of cerebrovascular events. Age-sex-adjusted relative risk for cardiovas-cular events suggested an increased risk for both drugs but the 95% confidence intervals for both relative risks included one.

The manufacturer of rofecoxib funded a case-control study among members of state-sponsored pharmacy benefit programs for residents aged 65 or older of New Jersey and Pennsylvania in the US (Solomon et al., 2004). Comparing rofecoxib with no use of non-selective NSAIDs or COX-2 inhibitors, naproxen, ibuprofen, or other NSAIDs, relative risk estimates for developing acute myocardial infarction were 1.14, 0.95, 1.21, and 1.17 respectively. Compar-ing celecoxib against the same drug groups, the relative risk estimates ranged between 0.93 and 0.98. 95% CIs for these eight relative risks all included one. Consis-tently higher relative risk of acute myocardial infarc-tion was associated with the use of more than 25 mg per day of rofecoxib than that for the use of 25 mg per day or less of rofecoxib when compared with the same groups. No dose-dependent finding was observed for celecoxib. Rofecoxib was associated with a 24% increased risk of hospitalized myocardial infarction when compared with celecoxib (Table 47.2).

NON-THROMBOEMBOLIC ADVERSE CARDIOVASCULAR EVENTS

Other adverse cardiovascular effects of COX-2 inhibitors have also been reported after their approval. Whelton and colleagues conducted a 6-week clin-ical trial of COX-2 inhibitors among patients 65 years or older with osteoarthritis and stable medication-controlled hypertension and showed that the incidence of increased systolic blood pressure was higher among patients randomly assigned to receive 25 mg of rofecoxib per day than among those who received celecoxib 200 mg per day (Whelton et al., 2002). Using automated administrative data from Ontario, Canada, Mamdani and colleagues reported an increased risk of congestive heart failure among rofe-coxib users but not celecoxib users (Mamdani et al., 2004).

SUMMARY OF SAFETY INFORMATION OF CELECOXIB AND ROFECOXIB THROUGH JULY 2004

Results from VIGOR clearly indicated that rofecoxib 50 mg per day was associated with higher risk of adverse cardiovascular thrombotic events. Epidemiol-ogy studies reported by Ray and colleagues in 2002 and by Solomon and colleagues in 2004 corrobo-rated this finding. Whether the use of rofecoxib 25 mg or less per day increased the risk of serious cardio-vascular thrombotic events was less certain. On the other hand, no compelling evidence suggested that celecoxib use was associated with increased risk of adverse cardiovascular outcomes and as of mid-2004 these results did not support a class effect of adverse cardiovascular effects for the COX-2 inhibitors. In the label revision for rofecoxib in 2002, increased cardiovascular risk observed from the VIGOR trial was noted and recommendation that high dose rofe-coxib not to be used chronically was added (Kweder, 2004).