Elsevier

The Lancet

Volume 364, Issue 9435, 21–27 August 2004, Pages 675-684
The Lancet

Articles
Comparison of lumiracoxib with naproxen and ibuprofen in the Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET), cardiovascular outcomes: randomised controlled trial

https://doi.org/10.1016/S0140-6736(04)16894-3Get rights and content

Summary

Background

The potential for cyclo-oxygenase 2 (COX2)-selective inhibitors to increase the risk for myocardial infarction is controversial. The Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET) aimed to assess gastrointestinal and cardiovascular safety of the COX2 inhibitor lumiracoxib compared with two non-steroidal anti-inflammatory drugs, naproxen and ibuprofen.

Methods

18 325 patients age 50 years or older with osteoarthritis were randomised to lumiracoxib 400 mg once daily (n=9156), naproxen 500 mg twice daily (4754), or ibuprofen 800 mg three times daily (4415) in two substudies of identical design. Randomisation was stratified for low-dose aspirin use and age. The primary cardiovascular endpoint was the Antiplatelet Trialists' Collaboration endpoint of non-fatal and silent myocardial infarction, stroke, or cardiovascular death. Analysis was by intention to treat.

Findings

81 (0·44%) patients did not start treatment and 7120 (39%) did not complete the study. At 1-year follow-up, incidence of the primary endpoint was low, both with lumiracoxib (59 events [0·65%]) and the non-steroidal anti-inflammatory drugs (50 events [0·55%]; hazard ratio 1·14 [95% CI 0·78–1·66], p=0·5074). Incidence of myocardial infarction (clinical and silent) in the overall population in the individual substudies was 0·38% with lumiracoxib (18 events) versus 0·21% with naproxen (ten) and 0·11% with lumiracoxib (five) versus 0·16% with ibuprofen (seven). In the naproxen substudy, rates of myocardial infarction (clinical and silent) did not differ significantly compared with lumiracoxib in the population not taking low-dose aspirin (hazard ratio 2·37 [95% CI 0·74–7·55], p=0·1454), overall (1·77 [0·82–3·84], p=0·1471), and in patients taking aspirin (1·36 [0·47–3·93], p=0·5658). In the ibuprofen substudy, these rates did not differ between lumiracoxib and ibuprofen in the population not taking low-dose aspirin (0·75 [0·20–2·79], p=0·6669), overall (0·66 [0·21–2·09], p=0·4833), and in patients taking aspirin (0·47 [0·04–5·14], p=0·5328).

Interpretation

The primary endpoint, including incidence of myocardial infarction, did not differ between lumiracoxib and either ibuprofen or naproxen, irrespective of aspirin use. This finding suggests that lumiracoxib is an appropriate treatment for patients with osteoarthritis, who are often at high cardiovascular risk and taking low-dose aspirin.

Introduction

The potential increased risk of cardiovascular adverse events with cyclo-oxygenase 2 (COX2)-selective inhibitors is controversial.1, 2, 3, 4, 5, 6 The Vioxx GI Outcomes Research (VIGOR) study7 compared the COX2-selective inhibitor rofecoxib with the non-selective, non-steroidal anti-inflammatory drug naproxen in 8000 patients with rheumatoid arthritis. These investigators recorded a difference in the rate of a composite endpoint of non-fatal myocardial infarction, non-fatal stroke, and sudden death between the treatment groups favouring naproxen (0·8% for rofecoxib vs 0·4% for naproxen, p<0·05), which was largely attributable to the difference in the incidence of myocardial infarction (0·4% vs 0·1%, p<0·01).

Data from three, large case-control studies suggest a possible antithrombotic effect of naproxen.8, 9, 10 One hypothesis is that this benefit arises mainly from the ability of naproxen to inhibit platelet aggregation11 to a level comparable with that recorded with aspirin. Since, to our knowledge, no placebo or aspirin-controlled trials of naproxen have been done in patients with known cardiovascular disease, definitive data are scarce about naproxen's possible clinical antithrombotic effects. Conversely, two large observational studies did not show any clinically meaningful reduction in myocardial infarction rates in users of naproxen compared with those taking non-steroidal anti-inflammatory drugs.4, 12

Additionally, researchers are unsure whether or not cardioprotection can be provided while maintaining adequate gastric protection in patients taking COX2-selective inhibitors with concomitant low-dose aspirin.13

Since COX2-selective inhibition has no effect on platelet thromboxane production, it might reduce vascular prostacyclin synthesis, thereby possibly altering the balance between thromboxane and prostacyclin, promoting a prothrombotic state.14, 15 However, clinical data to support this hypothesis are absent.

Besides the concern about an increase in myocardial infarction and thrombotic events, some evidence from an observational study6 has questioned whether the different pharmacokinetic and pharmacodynamic profiles of COX2-selective inhibitors confer in them differing cardiovascular adverse effects. Patients taking rofecoxib had a higher incidence of congestive heart failure and use of antihypertensive drugs than did those taking celecoxib or non-steroidal anti-inflammatory drugs versus controls.6

Osteoarthritis affects a large proportion of the population (18% women and 10% men worldwide),16 and coronary heart disease and cardiovascular risk factors are frequent in these people. Clinicians need evidence balancing any possible increased risk of myocardial infarction with potentially favourable gastrointestinal outcomes to make rational decisions about whom to treat with a COX2-selective inhibitor. A survey by the US National Center for Health Statistics reported that 40% of patients with osteoarthritis have hypertension compared with 25% in the general population without arthritis.17 Additionally, 20% of patients with osteoarthritis smoked, 11% had diabetes, and 32% had high total cholesterol (≥6·2 mmol/L). These statistics highlight the need to determine whether COX2-selective inhibitors differ in their adverse cardiovascular effects and whether this difference can alter their overall individual risk-benefit profiles.

Lumiracoxib is a novel COX2-selective inhibitor that has a different structure from others in the class, which are typically sulfonamides (celecoxib and valdecoxib) or sulfones (rofecoxib and etoricoxib).18, 19 It has a low lipophilicity, high selectivity, and a fairly short plasma half-life (3–6 h)20 compared with other COX2-selective inhibitors.18 These properties lead to less systemic drug exposure.

The Therapeutic Arthritis Research and Gastrointestinal Event Trial (TARGET) was designed to assess the gastrointestinal and cardiovascular safety of lumiracoxib compared with naproxen and ibuprofen.

Details of the methodology and implementation of the TARGET study have been published elsewhere21 and are reported here briefly and elsewhere in this issue.22

TARGET was an international double-blind study of more than 18 000 patients with osteoarthritis who received treatment with lumiracoxib 400 mg once daily (two or four times the recommended chronic dose for osteoarthritis), naproxen 500 mg twice daily (maximum therapeutic dose), or ibuprofen 800 mg three times daily (maximum therapeutic dose) for 52 weeks. Patients age 50 years or older were divided into one of six strata before randomisation by low-dose aspirin use (75–100 mg daily) and by age (50–64, 65–74, ≥75 years) and were followed up for 52 weeks, with clinic visits at 4, 13, 20, 26, 39, and 52 weeks. The study was undertaken in accordance with the principles of Good Clinical Practice and the Declaration of Helsinki. The TARGET study was approved by the ethics committees of all participating institutions and all patients gave their informed consent at the time of enrolment.

For logistical and masking reasons, TARGET was divided into two substudies, one with naproxen as the comparator and the other with ibuprofen. Within each substudy randomisation was stratified by age and low-dose aspirin use. The sponsor prepared a computer-generated randomisation list with appropriate blocks. The study was centrally randomised according to strata with an interactive voice response system in all countries to ensure age and low-dose aspirin stratification. Allocation of treatment was done via the interactive system and all information was verified by this system before allocation of the patient to a treatment and assignment of the drug packs. To ensure allocation concealment all treatment packs were identically designed and all study drugs were supplied as tablets with matching placebo. We prespecified that data from the two substudies would be pooled for analysis.

The primary objective was to test the hypothesis that lumiracoxib reduces the risk of developing upper gastrointestinal ulcer complications compared with the non-steroidal anti-inflammatory drugs naproxen and ibuprofen. A key secondary objective was to measure and compare a composite endpoint of cardiovascular morbidity and mortality across the randomised treatment groups.

By design, TARGET included patients at high cardiovascular risk who were taking low-dose aspirin (75–100 mg daily) for primary or secondary prevention of coronary heart disease. We intended to include about 24% of patients in the overall study population who were taking low-dose aspirin, based on projections from other trials. Low-dose aspirin use was assessed at baseline with the interactive voice response system and stratification was done accordingly. The low-dose aspirin strata information in this system was cross-checked for a matching entry on the concomitant medication case report form; any discrepancies were queried and the information corrected in the appropriate database. The decision for administration of aspirin was ultimately left to the discretion of the individual investigator.

We enrolled patients age 50 years or older with primary osteoarthritis. We excluded those with a history of myocardial infarction, stroke, coronary-artery bypass graft surgery, percutaneous coronary intervention, or new-onset angina within the 6 months before screening; patients with electrocardiogram evidence of silent myocardial ischaemia; those with congestive heart failure with symptoms at rest or minimal activity (New York Heart Association class III–IV); or patients who were receiving anticoagulation therapy (apart from low-dose aspirin).

Patients at increased risk for coronary heart disease were eligible for study entry provided that they had been receiving low-dose aspirin (75–100 mg daily) for primary or secondary cardiovascular prevention for a minimum of 3 months before randomisation. The definition of high cardiovascular risk was based on major independent cardiovascular risk factors assessed with Framingham risk assessment equations (for primary prevention)23 or a previous history of cardiovascular or cerebrovascular events (for secondary prevention). Patients not at high cardiovascular risk by these criteria but who were taking low-dose aspirin were also enrolled.

We obtained electrocardiograms at the screening visit and at the end of study or early termination visit and analysed them centrally. To assess any cardiac ischaemic events arising during the study, serial electrocardiograms were also obtained and analysed locally.

Predefined clinical events of myocardial infarction, unstable angina, cardiovascular death, cardiac arrest, stroke (ischaemic and haemorrhagic), transient ischaemic attack, deep vein thrombosis, and pulmonary embolism were independently adjudicated by a cardiovascular and cerebrovascular safety committee who were unaware of treatment allocation. We defined myocardial infarction as either established or acute. Established myocardial infarction was defined as development of new pathological Q waves on serial electrocardiograms. Acute or recent myocardial infarction was defined as a rise and gradual fall (troponin) or more rapid increase and decline (muscle-brain fraction of creatine kinase) of biochemical markers of myocardial necrosis with at least one of the following: ischaemic symptoms; development of pathological Q waves on electrocardiogram; changes on electrocardiogram indicative of ischaemia (ST segment elevation or depression); coronary artery intervention (eg, coronary angioplasty); or pathological findings of acute myocardial infarction. Events were assigned to one of the following categories: confirmed, probable, possible, not enough information for adjudication, or no event.

In addition to clinical events, the safety committee reviewed cases of electrocardiogram-detected myocardial infarctions—reported as a new finding on the end-of-study or post-baseline electrocardiogram by the central electrocardiogram reading laboratory. These events were categorised as either confirmed silent (electrocardiogram-detected) myocardial infarction or no silent myocardial infarction.

Other adverse events of interest included blood pressure measurements and congestive heart failure. Blood pressure was measured at every study visit after study drug was taken. The protocol stated that measurement should be taken after 5 min of rest using the same arm, the same device, and whenever possible at the same time of day (preferably between 0800 h and 1100 h). Blood-pressure changes were calculated for every patient as the average of their post-baseline values at every study visit and compared between treatment groups. Congestive heart failure data were obtained with case report forms as part of the monitoring and recording of all adverse events and serious adverse events. Congestive heart failure data were not adjudicated and analyses were done post hoc.

All cardiovascular and cerebrovascular events, as adjudicated by the safety committee, were summarised in frequency tables by type and adjudication category. The primary endpoint for analysis of cardiovascular adverse events was a composite cardiovascular endpoint, as defined by the Antiplatelet Trialists' Collaboration.24 This endpoint included confirmed silent (electrocardiogram-detected) myocardial infarctions, confirmed or probable clinical myocardial infarction, stroke (ischaemic and haemorrhagic), and cardiovascular death. Other cardiovascular events reported and adjudicated included cardiac arrest, transient ischaemic attack, unstable angina, deep vein thrombosis, and pulmonary embolism. Differences between treatment groups were measured with the Cox proportional-hazards regression model (with substudy, age, and low-dose aspirin as covariates), including estimation of hazard ratios and their associated 95% CIs. The Kaplan-Meier approach was used to generate estimates of the cumulative probability of events arising up to a particular timepoint. Analysis was by intention to treat. Analysis of blood pressure data used ANCOVA on average blood pressure changes across all post-baseline assessments, with baseline values and substudy as covariates.

The study was powered for the primary endpoint of confirmed or probable upper gastrointestinal ulcer complications, and these results are presented elsewhere in this issue.22

The study was designed interactively between an advisory board and the sponsor. The sponsor managed the data and did all final analyses. Authors had full access to all data and were involved in data interpretation and wrote the first draft of the report, which was further developed in collaboration with the sponsor.

Section snippets

Results

Figure 1 shows the disposition of patients included in the study, and Table 1 shows the baseline characteristics of randomised patients who received at least one dose of study medication (safety population). Compliance (defined as patients consuming full daily dose of study drug on at least 75% of days) was similar between lumiracoxib (6965, 76%) and non-steroidal anti-inflammatory drugs (6915, 76%) and within substudies (3189 [73%] ibuprofen vs 3213 [73%] lumiracoxib and 3726 [79%] naproxen vs

Discussion

We have shown that the incidence of non-fatal and silent myocardial infarction, stroke, or cardiovascular death, which was low in our population, did not differ between treatment groups (lumiracoxib 0·65% vs non-steroidal anti-inflammatory drugs 0·55%) or when analysed by aspirin use, age, sex, high cardiovascular risk, or cerebrovascular history. More events arose in the lumiracoxib versus naproxen substudy than in the lumiracoxib versus ibuprofen substudy, which could represent the difference

References (33)

  • D Mukherjee et al.

    Risk of cardiovascular events associated with selective COX-2 inhibitors

    JAMA

    (2001)
  • MA Konstam et al.

    Cardiovascular thrombotic events in controlled, clinical trials of rofecoxib

    Circulation

    (2001)
  • V Strand et al.

    The risk of cardiovascular thrombotic events with selective cyclooxygenase-2 inhibitors

    Arthritis Rheum

    (2002)
  • M Mamdani et al.

    Effect of selective cyclooxygenase 2 inhibitors and naproxen on short-term risk of acute myocardial infarction in the elderly

    Arch Intern Med

    (2003)
  • DH Solomon et al.

    Relationship between selective cyclooxygenase-2 inhibitors and acute myocardial infarction in older adults

    Circulation

    (2004)
  • C Bombardier et al.

    Comparison of upper gastrointestinal toxicity of rofecoxib and naproxen in patients with rheumatoid arthritis

    N Engl J Med

    (2000)
  • Cited by (0)

    View full text