The pain and the gain of treating patients with acute coronary syndromes—can the two be separated?
Article Outline
More than 1.5 million patients are admitted every year to acute care hospitals in the United States with a diagnosis of acute coronary syndrome (ACS) without ST-segment elevation.1 They consume a large amount of resources, and 1 in 6 patients develops cardiac ischemic events within 30 days. Furthermore, 5% to 6% of patients develop major bleeding during the hospitalization, which is associated with a substantially higher (2- to 4-fold) risk of immediate and more long-term death. The definition of major bleeding varies among clinical trials and registries. The more inclusive the definition of bleeding is, the more elaborate the models predicting bleeding are and the more difficult it is to identify causality versus confounding.2 Thus, for the physician caring for patients with ACS, 3 important questions need a prompt and discriminating answer: (1) Are there baseline demographic and clinical characteristics that predict the risk for ischemic events? (2) Are there baseline demographic and clinical characteristics that predict the risk of major bleeding in response to therapies directed against ischemic events? (3) If the answer is affirmative to the previous 2 questions, are the 2 sets of predictors sufficiently different so that we can reduce the risk of the latter without affecting the effectiveness of the treatments designed to prevent the former? Moreover, are different strategies preferable in certain subsets of patients? Conceptually, one can envision that these interrelated questions translate into a few distinct scenarios: A treatment strategy prevents ischemic events and causes more bleeding without net clinical benefit, or a treatment reduces bleeding without affecting ischemic events with little or no clinical benefit. The ideal strategy would, obviously, minimize ischemic and hemorrhagic complications independently of each other.
In this issue of the Journal (p. 502-508), Dr. Joyner and coinvestigators from the Organization to Assess Strategies in Ischemic Syndromes-5 (OASIS-5) study make important strides in clarifying these issues.3 Among more than 20,000 patients randomized to fondaparinux or enoxaparin as the antithrombin regimen administered for ACS, the primary end point of independently adjudicated death, myocardial infarction, or refractory ischemia was very similar among the 2 groups by 9 days (5.9% vs 5.8%, respectively, P = .85), whereas major bleeding was reduced by nearly 50% (2.1% vs 4.0%, P < .001). Using the well-validated (C-statistic 0.83) global registry of acute coronary events (GRACE) score,4 the investigators divided the study cohort into tertiles of risk (<118, 118-142, and >143) and compared the efficacy and safety of the allocated treatment in each risk category. Notably, because information regarding cardiac arrest on admission was not available, a value of 0 (vs 39 if documented) was imputed for all patients. As expected, the observed 30-day mortality was increased in the higher-risk tertiles and correlated well with the predicted values. More surprising to the authors was the increasing risk of major bleeding in the higher risk strata. With moderate accuracy (C-statistic 0.6), the GRACE risk score correlated with major bleeding rates: 2.2%, 3.2%, and 4.1%, respectively, for increasing tertiles of risk. The composite end point of death, infarction, or stroke at 180 days was also associated with baseline risk in a graded fashion. As compared with enoxaparin, fondaparinux had a similar effect on ischemic events (P for interaction, 0.43) and on hemorrhagic events (P for interaction, 0.77) in the different risk groups. Driven predominantly by a reduction in major bleeding, the net adverse clinical event rates at 6 months (integration of bleeding and ischemic events) were reduced by approximately 20% in each tertile of risk, without evidence of statistical interaction between risk and treatment allocation.
In summary, the authors emphasize that the balance of risk and benefit from antithrombin therapy favors fondaparinux at any given level of risk, predominantly because of the reduction in major bleeding, which is greater, in absolute terms, when the baseline risk is elevated (indeed, “no pain, no gain,” or “more pain, more gain”). Should we be surprised by these findings? Have the authors provided the elusive link between early bleeding and late(r) mortality, elegantly distilled by recent analyses of major randomized clinical trials using different antithrombotic regimens?5, 6, 7
A closer look at the GRACE score and at the OASIS-5 protocol might clarify the findings. The highest possible score of 333 (absent the 39 points contributed for cardiac arrest) is disproportionately affected by higher age (up to 100 points) and heart failure (maximum 59 points). More than three quarters of all patients were older than 60 years in this cohort. Renal dysfunction, assessed by serum creatinine rather than the preferred creatinine clearance, contributed equally to the score compared with ST-segment changes and twice more than elevated cardiac markers. Because increasing age and decreasing renal function are tightly linked to risk of bleeding, it is not surprising that the GRACE score predicts bleeding with at least modest accuracy. Furthermore, even though patients with creatinine levels >3.0 mg/dL were excluded from enrollment, because the dosing of enoxaparin was not adjusted by calculated creatinine clearance, it is likely that many elderly patients received higher than optimal doses of the drug for longer than desirable to prevent accumulation of the drug (5 days on average). Indeed, at 30 days, 57 more patients treated with enoxaparin died (of a total of 647 deaths), and nearly 70% of this excess occurred in those with bleeding events. The excess in mortality at 180 days was 69 (of a total of 1,202 deaths), and 85% of them occurred in bleeding patients.8
The GRACE registry also provided data for risk of bleeding in patients with ACS. Independent predictors of bleeding complications were increasing age, female gender, history of bleeding, and renal insufficiency. Three of these 4 factors are shared by the GRACE risk score for prediction of ischemic events. In a different dataset from the Acute Catheterization and Urgent Intervention Triage Strategy (ACUITY) trial, major bleeding predictors were, not surprisingly, advanced age, female gender, diabetes, hypertension, renal insufficiency, no prior revascularization, elevated cardiac markers, ST-segment changes, and treatment with heparin/glycoprotein inhibitor combination versus bivalirudin. Many of these variables feature prominently in the GRACE risk score.
Thus, we should expect that patients at higher risk for ischemic events will be also at risk for major bleeding because the predictors for both are not too dissimilar. It is the last of the questions posed at the beginning of this commentary that is the most critical: Can we uncouple this risk by applying a certain antithrombotic strategy rather than another? In other words, can we eliminate or attenuate the increasing risk of bleeding conferred by less favorable baseline demographic and clinical characteristics while attempting to reduce ischemic events? Obviously, this dual approach is predicated on the empiric observation from some,6, 7 but not all,9, 10 studies linking major bleeding with subsequent death (Figure 1). We learn from the OASIS-5 investigators in this issue that fondaparinux reduces bleeding without relation to the baseline hazard, measured as tertiles of GRACE risk score, by nearly half. As a result, the absolute improvement in net clinical outcome is stable at 2 to 3 events (death, infarction, stroke, or major bleeding) prevented by fondaparinux for every 100 patients treated. The sparing of bleeding is not accompanied by a significant reduction in overall ischemic events (death, infarction, or refractory ischemia), without heterogeneity among the tertiles. The most prominent effect of bleeding reduction is on in-hospital and early (30-day) death.
Figure 1.
Adjusted hazard ratios for death at 6 to 12 months from randomization according to presence of major bleeding before hospital discharge. Data from Thrombolysis in Myocardial Infarction (TIMI) 25 pertain to death between 30 and 365 days.
We conclude that, as of now, the gain and the pain of ACS management cannot be entirely uncoupled, because both ischemic and hemorrhagic events are predicted by shared and frequently nonmodifiable characteristics of the patients experiencing it.11 Most deaths after hospital discharge still occur in the approximately 90% to 95% of patients who did not experience major bleeding, although the latter group experiences a higher mortality rate. Unless we are able to minimize primarily recurrent ischemic events without increasing bleeding, we will not be able to profoundly affect outcome. Just reducing bleeding (and death indirectly), although highly desirable and an important step forward, is not fully sufficient to reduce the morbidity and mortality of ACS.
Disclosures
Dr Brener discloses the following relationships: Consultant/Advisory Board (moderate)—Sanofi Aventis (Paris, France), Eli Lilly (Indianapolis, IN). Dr Bhatt discloses the following relationships: Research Grants (significant, directly to the institution)—Bristol Myers Squibb (Princeton, NJ), Eisai (Tokyo, Japan), Ethicon (New Brunswick, NJ), Heartscape (Columbia, MD), Sanofi Aventis, The Medicines Company (Parsippany, NJ); Consultant/Advisory Board (modest, waived or donated to nonprofits for >2 years)—Arena (San Diego, CA), Astellas (Tokyo, Japan), Astra Zeneca (London, UK), Bayer (Leverkusen, Germany), Bristol Myers Squibb, Cardax (Honolulu, HI), Centocor (Horsham, PA), Cogentus (Palo Alto, CA), Daiichi-Sankyo (Tokyo, Japan), Eisai, Eli Lilly, GlaxoSmithKline (Philadelphia, PA), Johnson & Johnson (New Brunswick, NJ), McNeil (New Brunswick, NJ), Medtronic (Minneapolis, MN), Millennium (Cambridge, MA), Molecular Insights (Cambridge, MA), Otsuka (Rockville, MD), Paringenix (Tucson, AZ), PDL (Redwood City, CA), Philips (Amsterdam, The Netherlands), Portola (San Francisco, CA), Sanofi Aventis, Schering Plough (Kenilworth, NJ), Scios (New Brunswick, NJ), Takeda (Osaka, Japan), The Medicines Company, Vertex (Cambridge, MA).
References
- Heart disease and stroke statistics—2008 update: a report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation. 2008;117:e25–e146
- Impact of major bleeding on 30-day mortality and clinical outcomes in patients with acute coronary syndromes: an analysis from the ACUITY Trial. J Am Coll Cardiol. 2007;49:1362–1368
- Fondaparinux compared to enoxaparin in patients with acute coronary syndromes without ST-segment elevation: outcomes and treatment effect across different levels of risk. Am Heart J. 2009;157:502–508
- Predictors of hospital mortality in the global registry of acute coronary events. Arch Intern Med. 2003;163:2345–2353
- Periprocedural bleeding and 1-year outcome after percutaneous coronary interventions appropriateness of including bleeding as a component of a quadruple end point. J Am Coll Cardiol. 2008;51:690–697
- Budaj A, Eikelboom JW, Mehta SR, et al. Improving clinical outcomes by reducing bleeding in patients with non–ST-elevation acute coronary syndromes. Eur Heart J [Electronic Publication 2008 Aug 18].
- Antithrombotic strategies in patients with acute coronary syndromes undergoing early invasive management: one-year results from the ACUITY Trial. JAMA. 2007;298:2497–2506
- Comparison of fondaparinux and enoxaparin in acute coronary syndromes. N Engl J Med. 2006;354:1464–1476
- The relative contribution of clinical factors, treatments bleeding and CV complication to mortality in 20,323 patients receiving fibrinolysis for STEMI. J Am Coll Cardiol. 2008;51:A209
- Does comorbidity account for the excess mortality in patients with major bleeding in acute myocardial infarction?. Circulation. 2007;116:2793–2801
- . Intensifying platelet inhibition—navigating between Scylla and Charybdis. N Engl J Med. 2007;357:2078–2081
PII: S0002-8703(08)01030-2
doi:10.1016/j.ahj.2008.11.012
© 2009 Mosby, Inc. All rights reserved.
