Journal of the American College of Cardiology Vol. 60, No. 8, 2012 © 2012 by the American College of Cardiology Foundation ISSN 0735-1097/$36.00 Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jacc.2012.05.012

make indirect comparisons between the new agents. Be- EDITORIAL COMMENT cause each trial had a common comparator arm, one could assume similarity in outcomes of that arm and then look at the relative benefit (or harm) that each drug displays versus Danger Ahead: Watch Out the standard arm. They have used the Bucher method to compare hazard ratios of the new agent to the comparator to for Indirect Comparisons!* try to determine the relative efficacy between the 2 newer agents. Christopher P. Cannon, MD,† Payal Kohli, MD‡ For such comparisons, however, we need to proceed with extreme caution: attempts at indirect comparisons were Boston, Massachusetts made among fibrinolytic agents—tissue plasminogen acti- vator, streptokinase, and anistreplase—where relative reduc- tions in mortality in placebo-controlled trials were 25%, We have entered an exciting new era of anticoagulation— 25%, and 50%, respectively (6–8). On the basis of these with multiple new oral anticoagulants becoming available results, one might be tempted to conclude that anistreplase for use as potential replacements for warfarin, the vitamin K is superior, but in a direct comparison of these drugs in antagonist that has been the only oral agent available for 50 ISIS-3 (Third International Study of Infarct Survival), no years! The most prominent of the indications for anticoag- difference in endpoints was seen (9). As such, indirect ulation is for stroke prevention in patients with atrial comparisons can be misleading, and extreme caution should fibrillation, where long-term, often life-long therapy is be exercised when using such methods to draw definitive needed in millions of patients. Two agents have already conclusions. been approved by the Food and Drug Administration After evaluating the efficacy of each new drug individu- (FDA), and a third is anticipated later this year. There are ally, Lip et al. (5) then went on to calculate a weighted 2 main classes of drugs that are undergoing testing or have average of all the new agents combined to estimate the been tested in large randomized trials: direct thrombin relative benefit of “any new oral anticoagulant” versus inhibitors (dabigatran [1]) and factor Xa inhibitors (rivar- warfarin. This latter approach is akin to pooling or meta- oxaban [2], apixaban [3], edoxaban [4]); in these trials, these analysis of all the studies, and provides a reasonable over- drugs are being compared with warfarin. view of the “class effect.” For all the new oral anticoagulants, they found that this class as a whole significantly reduced See page 738 the risk of: 1) stroke or systemic embolism by 21% (p Ͻ 0.001); 2) stroke by 23% (p Ͻ 0.001); 3) hemorrhagic stroke Three large randomized trials (totaling over 50,000 pa- by 53% (p Ͻ 0.001); and 4) all-cause mortality by 12% (p Ͻ tients with atrial fibrillation) comparing one of the newer 0.001). In addition, major bleeding was lower for any new oral anticoagulants to warfarin have already been published. oral anticoagulant by 13% (p Ͻ 0.001) compared with Each new drug has shown equal or superior efficacy to warfarin. This straightforward and statistically sound anal- warfarin, suggesting that these agents offer excellent alter- ysis of the data strongly supports the use of the newer agents natives to warfarin for stroke prevention. But faced with over warfarin, as a class. these growing new options, how do physicians choose For the indirect comparisons between individual agents, between them? Are there differences between the drugs in however, the picture remains a bit confusing. Lip et al. their efficacy and/or safety that would help inform our report a “significantly lower risk of stroke or systemic clinical decision to prescribe one over the other? Unfortu- embolism (by 26%) for dabigatran 150 mg twice daily nately, no trials have been conducted (yet) that directly compared with rivaroxaban.” On the other hand, they compare one new agent versus another. conclude “we found no profound significant differences in In this issue of the Journal, Lip et al. (5) have tried to efficacy between apixaban, and dabigatran (both doses) or answer these questions by applying statistical methods to rivaroxaban.” They also report no significant difference in myocardial infarction event rates among the 3 agents. Yet, within the individual trials, the rate of myocardial infarction *Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the was definitely higher with dabigatran (1), but not with American College of Cardiology. either of the factor Xa inhibitors (2,3). These conflicting From the †Cardiovascular Division, Brigham and Women’s Hospital, Boston, results lend support to the conclusion that such methods for Massachusetts; and the ‡Division of Cardiology, University of California San Francisco, San Francisco, California. Dr. Cannon is a member of the advisory boards indirect comparisons may not be the most accurate due to for Alnylam, Bristol-Myers Squibb, and Pfizer; has received financial support from several sources of confounding. them, which has been donated to charity; and he has also received grant support from The authors acknowledge that there are multiple limita- Accumetrics, AstraZeneca, Essentialis, GlaxoSmithKline, Merck, Regeneron, Sanofi, Takeda; and has equity in Automedics Medical systems. Dr. Kohli is a member of the tions in comparing different trials, despite the common Advisory Board to Daiichi Sankyo. comparator arm—notably the patient population differs. 748 Cannon and Kohli JACC Vol. 60, No. 8, 2012 Danger of Indirect Comparisons August 21, 2012:747–8

Variations in the baseline risk of patients in 1 study could appear to be saying that there are more similarities between influence the absolute rates of the endpoints in that study, these agents than differences, as has also been previously confounding the drug–drug comparison being studied. For noted by Mega (10). However, because of the statistical example, the ROCKET trial (Rivaroxaban-once daily, oral, limitations of such comparisons, although of some interest, direct factor Xa inhibition compared with vitamin K antag- we feel the differences they report on some endpoints are onism for prevention of stroke and Embolism Trial in Atrial not robust enough to be relied upon for the clinical care of Fibrillation) trial enrolled older patients with higher patients. Instead, we would turn to direct evidence from

CHADS2 (Congestive heart failure, Hypertension, Age, trials and the indications put forth by the FDA to select the Diabetes, Prior Stroke) risk scores, and as such, would be appropriate agent, at the dose tested, for use in the patient expected to have higher rates of bleeding and stroke—which population studied within the trial. might influence the relative drug effect seen in this trial. Therefore, 2 variables would differ when comparing a Reprint requests and correspondence: Dr. Christopher P. Can- similar endpoint between these 2 trials—the drug being non, TIMI Study Group, Brigham and Women’s Hospital, 350 tested and the patient population—and it would be unclear Longwood Avenue, 1st Floor, Boston, Massachusetts 02115. whether any apparent difference observed in the relative E-mail: [email protected]. efficacy could be attributed to the drug alone. Similarly, the optimal use of warfarin (as measured by the “time in REFERENCES therapeutic range”) also differed between these trials, result- 1. Connolly SJ, Ezekowitz MD, Yusuf S, et al. Dabigatran versus ing in another possible source of confounding that could warfarin in patients with atrial fibrillation. N Engl J Med 2009;361: influence the relative efficacy of the drugs when comparing 1139–51. 2. Patel MR, Mahaffey KW, Garg J, et al. Rivaroxaban versus warfarin in the trials. nonvalvular atrial fibrillation. N Engl J Med 2011;365:883–91. Traditional metrics for determining efficacy derive di- 3. Granger CB, Alexander JH, McMurray JJ, et al. Apixaban versus rectly from the results of the randomized trial. Indeed, this warfarin in patients with atrial fibrillation. N Engl J Med 2011;365: 981–92. is exactly what the FDA does in the prescribing insert: the 4. Ruff CT, Giugliano RP, Antman EM, et al. Evaluation of the novel FDA approves and lists only those findings and indications factor Xa inhibitor edoxaban compared with warfarin in patients with that were directly reported in the randomized trials. As atrial fibrillation: design and rationale for the Effective aNticoaGula- tion with factor xA next GEneration in Atrial Fibrillation- such, we would find that dabigatran 150 mg twice daily Thrombolysis In Myocardial Infarction study 48 (ENGAGE AF- significantly reduced stroke or systemic embolism (and TIMI 48). Am Heart J 2010;160:635–41. within that endpoint, ischemic stroke alone), but with a 5. Lip GYH, Larsen TB, Skjøth F, Rasmussen KH. Indirect compari- sons of new oral anticoagulant drugs for efficacy and safety when used similar rate of major bleeding to warfarin. The 110-mg dose for stroke prevention in atrial fibrillation. J Am Coll Cardiol 2012;60: (unfortunately not approved by the FDA) had similar 738–46. efficacy for stroke or systemic embolism but a lower risk of 6. Wilcox RG, von der Lippe G, Olsson CG, Jensen G, Skene AM, Hampton JR. Trial of tissue plasminogen activator for mortality major bleeding compared with warfarin (1). For apixaban, reduction in acute myocardial infarction. Anglo-Scandinavian Study of significant reductions in stroke or systemic embolism, major Early Thrombolysis (ASSET). Lancet 1988;2:525–30. bleeding, and mortality were seen (3). Rivaroxaban showed 7. ISIS-2 (Second International Study of Infarct Survival) Collaborative Group. Randomised trial of intravenous streptokinase, oral aspirin, similar rates of stroke or systemic embolism and major both, or neither among 17,187 cases of suspected acute myocardial bleeding versus warfarin (2). But whether apixaban and infarction: ISIS-2. Lancet 1988;2:349–60. dabigatran 150 mg are superior to rivaroxaban cannot 8. AIMS Trial Study Group. Long-term effects of intravenous anis- treplase in acute myocardial infarction: final report of the AIMS study. readily be concluded by comparing the relative efficacy of Lancet 1990;335:427–31. each agent versus comparator within its own respective trial. 9. ISIS-3 (Third International Study of Infarct Survival) Collaborative We can say though that these 2 agents have been shown to Group. ISIS-3: a randomised comparison of streptokinase vs tissue plasminogen activator vs anistreplase and of aspirin plus heparin vs be superior for stroke prevention, and apixaban for mortality aspirin alone among 41,299 cases of suspected acute myocardial reduction, whereas neither can be said for rivaroxaban. infarction. Lancet 1992;339:753–70. So what are we to do? Should we use the indirect 10. Mega JL. A new era for anticoagulation in atrial fibrillation. N Engl J Med 2011;365:1052–4. comparisons put forth by Lip et al. (5) since that provides the only comparative data we have? In general, the authors Key Words: anticoagulation y atrial fibrillation y bleeding y stroke.