The Charisma of Subgroups and the Subgroups of CHARISMA
http://www.100md.com
《新英格兰医药杂志》
Antiplatelet therapy with aspirin, an irreversible inhibitor of platelet cyclooxygenase, has earned its rightful place as a cornerstone of treatment for reducing cardiovascular events in patients with established vascular disease.1,2 This statement is based on consistent results favoring aspirin in a large number of randomized, controlled trials in populations with either acute or long-term disease. Clopidogrel, by inhibiting the adenosine diphosphate P2Y12 receptor, offers a distinctly different mechanism to reduce platelet activation and aggregation. As monotherapy, clopidogrel has been shown to be effective in reducing the risk of vascular events in patients with established vascular disease.3
Even more important, when clopidogrel is given in addition to aspirin, major incremental clinical benefits have been achieved with relatively short-term use in the setting of acute vascular injury. During coronary stenting, dual antiplatelet therapy with ticlopidine (generally supplanted by clopidogrel) results in a significant reduction in clinical events.4 Similarly, across the spectrum of acute coronary syndromes, the addition of clopidogrel to aspirin results in measurable clinical benefits.5,6,7,8 Most convincingly, COMMIT (the Clopidogrel and Metoprolol in Myocardial Infarction Trial, also known as the Second Chinese Cardiac Study ), which included more than 45,000 patients, demonstrated that therapy with clopidogrel plus aspirin resulted in a clear reduction in mortality during acute myocardial infarction.8 In these relatively short-term studies, the cardiovascular benefits were achieved without a significant excess of bleeding. Thus, the data support a recommendation for this dual antiplatelet strategy in the setting of acute vascular injury, whether the injury is catheter-induced (during percutaneous coronary intervention) or associated with a spontaneous acute coronary syndrome.
Investigators continue to conduct randomized trials of proven therapies in order to refine risk–benefit ratios precisely. The Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial, reported in this issue of the Journal,9 was designed to test the hypothesis that "long-term treatment with a combination of clopidogrel plus aspirin may provide greater protection against cardiovascular events than aspirin alone in a broad population of patients at high risk." Therefore, the trial was probing uncharted territory in determining whether long-term dual antiplatelet therapy would be of incremental clinical value in patients with vascular disease without an acute manifestation of injury. If the addition of clopidogrel to low-dose aspirin resulted in a reduction in the risk of the primary efficacy end point (death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke), the finding would certainly represent a major clinical advance, especially if the study population was receiving the benefits of other proven therapies, such as antihypertensive agents and lipid-lowering agents (particularly statins at evidence-based doses). Any potential efficacy would have to be weighed against the increased risk of bleeding, which is an inherent consequence of antiplatelet therapy. The event-driven design of the trial, with continuation until at least 1040 primary efficacy events had occurred, ensured that the primary hypothesis would be adequately tested.
CHARISMA went to completion; 15,603 randomly assigned patients were followed for a median of 28 months. During that time, 1107 patients had at least one of the major cardiovascular events constituting the composite primary outcome. The distribution of these events did not differ significantly between the study groups (534 in the clopidogrel group and 573 in the placebo group, P=0.22). As such, the trial did not reject the null hypothesis. Thus, in this population of patients with established coronary, cerebral, or peripheral arterial disease or with multiple risk factors for atherothrombosis, the addition of clopidogrel to aspirin did not result in a lowering of risk. The investigators also found an increased risk of at least moderate bleeding (defined as bleeding requiring transfusion) — a finding consistent with the results of other long-term trials of dual antiplatelet therapy.5 Indeed, in the recently presented ACTIVE (Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events), which involved patients with atrial fibrillation, the risk of bleeding complications with dual antiplatelet treatment was as high as that associated with oral anticoagulation.10 In CHARISMA, the absence of a clear benefit in terms of clinical outcome, coupled with the increased rate of bleeding (as well as the economic considerations related to the long-term use of clopidogrel), provides a robust answer to the central question of the trial and argues against the use of dual antiplatelet therapy in this patient population.
There is understandable desire on the part of both practitioners and investigators for refinement of the overall findings of any randomized trial. For practitioners, there is the desire to know more about risks and benefits for their specific patients. For investigators, the massive effort of a major international clinical trial seems to warrant more than one hypothesis-testing P value. Unfortunately, this exploration of the data, generally termed subgroup analysis, is notoriously fraught with multiple hazards, especially the play of chance and uncontrollable confounders.11,12,13 Indeed, there are numerous examples from randomized trials in which an apparently important differential response to therapy suggested by a subgroup analysis generated a hypothesis that was subsequently refuted in a trial designed to test that hypothesis.13
Of the multiple subgroups defined in CHARISMA, the investigators emphasize the group designated as having either multiple risk factors or clinically manifest atherothrombosis (and termed "asymptomatic" or "symptomatic," respectively). Among the patients classified as symptomatic, there was a marginal reduction in the primary end point, whereas among those classified as asymptomatic, the trend favored the placebo group; an unadjusted test for interaction indicated borderline significance (P=0.045). As the authors point out, these observations should be interpreted with caution, but not just because the interaction term was only marginally significant and not adjusted for multiple analyses. We are also concerned that the characteristics that differentiate patients in this subgroup are not sufficiently distinct. As reported, some of the patients termed "asymptomatic" had major cardiovascular events, and one would assume that a large proportion of the patients termed "symptomatic" also had multiple risk factors. If anything, this subgroup analysis could be used to generate a hypothesis for a future study to help define the currently blurry distinction between the benefits and risks of dual antiplatelet therapy in patients with acute injury and those in patients with more stable vascular disease.
Of the data presented, we find the overall point estimate and 95 percent confidence limits for the entire population convincing. These data show no significant benefit associated with long-term clopidogrel therapy in addition to aspirin. Admittedly, this "one size fits all" approach leaves much to be desired. However, more homogeneous genotypic and phenotypic (pathophysiological) characterization of patients will be required before clinical trials can be "personalized."14 Until then, the charisma of extracting favorable P values from subgroups should be resisted and dual antiplatelet therapy avoided in these patients with stable disease.
Dr. Pfeffer reports having received research grants to Brigham and Women's Hospital from Hamilton University for a project funded by Sanofi-Aventis and Bristol-Myers Squibb. He also reports having received consulting fees from Bristol-Myers Squibb. No other potential conflict of interest relevant to this article was reported.
Source Information
From Brigham and Women's Hospital and Harvard Medical School (M.A.P.) and the Journal (J.A.J.) — all in Boston.
This article was published at www.nejm.org on March 12, 2006.
References
Patrono C, Rodríguez LAG, Landolfi R, Baigent C. Low-dose aspirin for the prevention of atherothrombosis. N Engl J Med 2005;353:2373-2383.
Antithrombotic Trialists' Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ 2002;324:71-86.
CAPRIE Steering Committee. A randomised, blinded trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1996;348:1329-1339.
Leon MB, Baim DS, Popma JJ, et al. A clinical trial comparing three antithrombotic-drug regimens after coronary-artery stenting. N Engl J Med 1998;339:1665-1671.
The Clopidogrel in Unstable Angina to Prevent Recurrent Events Trial Investigators. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494-502.
Mehta SR, Yusuf S, Peters RJ, et al. Effects of pretreatment with clopidogrel and aspirin followed by long-term therapy in patients undergoing percutaneous coronary intervention: the PCI-CURE study. Lancet 2001;358:527-533.
Sabatine MS, Cannon CP, Gibson M, et al. Addition of clopidogrel to aspirin and fibrinolytic therapy for myocardial infarction with ST-segment elevation. N Engl J Med 2005;352:1179-1189.
COMMIT Collaborative Group. Addition of clopidogrel to aspirin in 45,852 patients with acute myocardial infarction: randomised placebo-controlled trial. Lancet 2005;366:1607-1621.
Bhatt DL, Fox KAA, Hacke W, et al. Clopidogrel and aspirin versus aspirin alone for the prevention of atherothrombotic events. N Engl J Med 2006;354:1706-1717.
Connolly SJ. The Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events (ACTIVE-W)–Non-inferiority trial versus oral anticoagulation. Presented at the Late-Breaking Clinical Trials III, American Heart Association Scientific Sessions 2005, Dallas, November 13–16, 2005.
Yusuf S, Wittes J, Probstfield J, Tyroler HA. Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials. JAMA 1991;266:93-98.
Moyé LA. Statistical reasoning in medicine: the intuitive P-value primer. New York: Springer-Verlag, 2000.
Rothwell PM. Treating individuals. 2. Subgroup analysis in randomised controlled trials: importance, indications, and interpretation. Lancet 2005;365:176-186.
Maseri A. Atillio Maseri, MD, FACC, discusses the role of biological clinical research in cardiovascular disease and the move from megatrials to personalized patient management. Circulation 2006;113:29-30.(Marc A. Pfeffer, M.D., Ph)
Even more important, when clopidogrel is given in addition to aspirin, major incremental clinical benefits have been achieved with relatively short-term use in the setting of acute vascular injury. During coronary stenting, dual antiplatelet therapy with ticlopidine (generally supplanted by clopidogrel) results in a significant reduction in clinical events.4 Similarly, across the spectrum of acute coronary syndromes, the addition of clopidogrel to aspirin results in measurable clinical benefits.5,6,7,8 Most convincingly, COMMIT (the Clopidogrel and Metoprolol in Myocardial Infarction Trial, also known as the Second Chinese Cardiac Study ), which included more than 45,000 patients, demonstrated that therapy with clopidogrel plus aspirin resulted in a clear reduction in mortality during acute myocardial infarction.8 In these relatively short-term studies, the cardiovascular benefits were achieved without a significant excess of bleeding. Thus, the data support a recommendation for this dual antiplatelet strategy in the setting of acute vascular injury, whether the injury is catheter-induced (during percutaneous coronary intervention) or associated with a spontaneous acute coronary syndrome.
Investigators continue to conduct randomized trials of proven therapies in order to refine risk–benefit ratios precisely. The Clopidogrel for High Atherothrombotic Risk and Ischemic Stabilization, Management, and Avoidance (CHARISMA) trial, reported in this issue of the Journal,9 was designed to test the hypothesis that "long-term treatment with a combination of clopidogrel plus aspirin may provide greater protection against cardiovascular events than aspirin alone in a broad population of patients at high risk." Therefore, the trial was probing uncharted territory in determining whether long-term dual antiplatelet therapy would be of incremental clinical value in patients with vascular disease without an acute manifestation of injury. If the addition of clopidogrel to low-dose aspirin resulted in a reduction in the risk of the primary efficacy end point (death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke), the finding would certainly represent a major clinical advance, especially if the study population was receiving the benefits of other proven therapies, such as antihypertensive agents and lipid-lowering agents (particularly statins at evidence-based doses). Any potential efficacy would have to be weighed against the increased risk of bleeding, which is an inherent consequence of antiplatelet therapy. The event-driven design of the trial, with continuation until at least 1040 primary efficacy events had occurred, ensured that the primary hypothesis would be adequately tested.
CHARISMA went to completion; 15,603 randomly assigned patients were followed for a median of 28 months. During that time, 1107 patients had at least one of the major cardiovascular events constituting the composite primary outcome. The distribution of these events did not differ significantly between the study groups (534 in the clopidogrel group and 573 in the placebo group, P=0.22). As such, the trial did not reject the null hypothesis. Thus, in this population of patients with established coronary, cerebral, or peripheral arterial disease or with multiple risk factors for atherothrombosis, the addition of clopidogrel to aspirin did not result in a lowering of risk. The investigators also found an increased risk of at least moderate bleeding (defined as bleeding requiring transfusion) — a finding consistent with the results of other long-term trials of dual antiplatelet therapy.5 Indeed, in the recently presented ACTIVE (Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events), which involved patients with atrial fibrillation, the risk of bleeding complications with dual antiplatelet treatment was as high as that associated with oral anticoagulation.10 In CHARISMA, the absence of a clear benefit in terms of clinical outcome, coupled with the increased rate of bleeding (as well as the economic considerations related to the long-term use of clopidogrel), provides a robust answer to the central question of the trial and argues against the use of dual antiplatelet therapy in this patient population.
There is understandable desire on the part of both practitioners and investigators for refinement of the overall findings of any randomized trial. For practitioners, there is the desire to know more about risks and benefits for their specific patients. For investigators, the massive effort of a major international clinical trial seems to warrant more than one hypothesis-testing P value. Unfortunately, this exploration of the data, generally termed subgroup analysis, is notoriously fraught with multiple hazards, especially the play of chance and uncontrollable confounders.11,12,13 Indeed, there are numerous examples from randomized trials in which an apparently important differential response to therapy suggested by a subgroup analysis generated a hypothesis that was subsequently refuted in a trial designed to test that hypothesis.13
Of the multiple subgroups defined in CHARISMA, the investigators emphasize the group designated as having either multiple risk factors or clinically manifest atherothrombosis (and termed "asymptomatic" or "symptomatic," respectively). Among the patients classified as symptomatic, there was a marginal reduction in the primary end point, whereas among those classified as asymptomatic, the trend favored the placebo group; an unadjusted test for interaction indicated borderline significance (P=0.045). As the authors point out, these observations should be interpreted with caution, but not just because the interaction term was only marginally significant and not adjusted for multiple analyses. We are also concerned that the characteristics that differentiate patients in this subgroup are not sufficiently distinct. As reported, some of the patients termed "asymptomatic" had major cardiovascular events, and one would assume that a large proportion of the patients termed "symptomatic" also had multiple risk factors. If anything, this subgroup analysis could be used to generate a hypothesis for a future study to help define the currently blurry distinction between the benefits and risks of dual antiplatelet therapy in patients with acute injury and those in patients with more stable vascular disease.
Of the data presented, we find the overall point estimate and 95 percent confidence limits for the entire population convincing. These data show no significant benefit associated with long-term clopidogrel therapy in addition to aspirin. Admittedly, this "one size fits all" approach leaves much to be desired. However, more homogeneous genotypic and phenotypic (pathophysiological) characterization of patients will be required before clinical trials can be "personalized."14 Until then, the charisma of extracting favorable P values from subgroups should be resisted and dual antiplatelet therapy avoided in these patients with stable disease.
Dr. Pfeffer reports having received research grants to Brigham and Women's Hospital from Hamilton University for a project funded by Sanofi-Aventis and Bristol-Myers Squibb. He also reports having received consulting fees from Bristol-Myers Squibb. No other potential conflict of interest relevant to this article was reported.
Source Information
From Brigham and Women's Hospital and Harvard Medical School (M.A.P.) and the Journal (J.A.J.) — all in Boston.
This article was published at www.nejm.org on March 12, 2006.
References
Patrono C, Rodríguez LAG, Landolfi R, Baigent C. Low-dose aspirin for the prevention of atherothrombosis. N Engl J Med 2005;353:2373-2383.
Antithrombotic Trialists' Collaboration. Collaborative meta-analysis of randomised trials of antiplatelet therapy for prevention of death, myocardial infarction, and stroke in high risk patients. BMJ 2002;324:71-86.
CAPRIE Steering Committee. A randomised, blinded trial of clopidogrel versus aspirin in patients at risk of ischaemic events (CAPRIE). Lancet 1996;348:1329-1339.
Leon MB, Baim DS, Popma JJ, et al. A clinical trial comparing three antithrombotic-drug regimens after coronary-artery stenting. N Engl J Med 1998;339:1665-1671.
The Clopidogrel in Unstable Angina to Prevent Recurrent Events Trial Investigators. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494-502.
Mehta SR, Yusuf S, Peters RJ, et al. Effects of pretreatment with clopidogrel and aspirin followed by long-term therapy in patients undergoing percutaneous coronary intervention: the PCI-CURE study. Lancet 2001;358:527-533.
Sabatine MS, Cannon CP, Gibson M, et al. Addition of clopidogrel to aspirin and fibrinolytic therapy for myocardial infarction with ST-segment elevation. N Engl J Med 2005;352:1179-1189.
COMMIT Collaborative Group. Addition of clopidogrel to aspirin in 45,852 patients with acute myocardial infarction: randomised placebo-controlled trial. Lancet 2005;366:1607-1621.
Bhatt DL, Fox KAA, Hacke W, et al. Clopidogrel and aspirin versus aspirin alone for the prevention of atherothrombotic events. N Engl J Med 2006;354:1706-1717.
Connolly SJ. The Atrial Fibrillation Clopidogrel Trial with Irbesartan for Prevention of Vascular Events (ACTIVE-W)–Non-inferiority trial versus oral anticoagulation. Presented at the Late-Breaking Clinical Trials III, American Heart Association Scientific Sessions 2005, Dallas, November 13–16, 2005.
Yusuf S, Wittes J, Probstfield J, Tyroler HA. Analysis and interpretation of treatment effects in subgroups of patients in randomized clinical trials. JAMA 1991;266:93-98.
Moyé LA. Statistical reasoning in medicine: the intuitive P-value primer. New York: Springer-Verlag, 2000.
Rothwell PM. Treating individuals. 2. Subgroup analysis in randomised controlled trials: importance, indications, and interpretation. Lancet 2005;365:176-186.
Maseri A. Atillio Maseri, MD, FACC, discusses the role of biological clinical research in cardiovascular disease and the move from megatrials to personalized patient management. Circulation 2006;113:29-30.(Marc A. Pfeffer, M.D., Ph)