Use of Anticoagulant Agents and Risk of Bleeding Among Patients Admitted With Myocardial Infarction

Transcription

1 JACC: CARDIOVASCULAR INTERVENTIONS VOL. 3, NO. 11, BY THE AMERICAN COLLEGE OF CARDIOLOGY FOUNDATION ISSN /$36.00 PUBLISHED BY ELSEVIER INC. DOI: /j.jcin CLINICAL RESEARCH Use of Anticoagulant Agents and Risk of Bleeding Among Patients Admitted With Myocardial Infarction A Report From the NCDR ACTION Registry GWTG (National Cardiovascular Data Registry Acute Coronary Treatment and Intervention Outcomes Network Registry Get With the Guidelines) Mitul B. Kadakia, MD,* Nihar R. Desai, MD, MPH,* Karen P. Alexander, MD, Anita Y. Chen, MS, JoAnne M. Foody, MD, Christopher P. Cannon, MD,* Stephen D. Wiviott, MD,* Benjamin M. Scirica, MD, MPH,* on behalf of the National Cardiovascular Data Registry Boston, Massachusetts; and Durham, North Carolina Objectives The aim of this study was to evaluate anticoagulant use patterns and bleeding risk in a contemporary population of patients with acute coronary syndrome. Background Current practice guidelines support the use of unfractionated heparin, low molecular weight heparin, bivalirudin, or fondaparinux in non ST-segment elevation myocardial infarction (NSTEMI) and ST-segment elevation myocardial infarction (STEMI). Little is known about how these agents are selected in clinical practice. Methods Between January 2007 and June 2009, data were captured for 72,699 patients with NSTEMI and 48,943 patients with STEMI at 360 U.S. hospitals for the NCDR ACTION Registry GWTG (National Cardiovascular Data Registry Acute Coronary Treatment and Intervention Outcomes Network Registry Get With the Guidelines). Patients were categorized based on anticoagulant strategy selected during hospitalization and their CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of ACC/AHA [American College of Cardiology/American Heart Association] Guidelines) bleeding risk category. Results At least 1 anticoagulant was administered to 66,279 patients (91.2%) with NSTEMI and 46,149 patients (94.3%) with STEMI. Among STEMI patients, unfractionated heparin was most commonly used (66%), followed by bivalirudin (14%) and low molecular weight heparin (8%). In NSTEMI patients, unfractionated heparin was also the most commonly used anticoagulant (42%), followed by low molecular weight heparin (27%) and then bivalirudin (13%). There were significant differences in anticoagulant use by age, risk factors, concomitant medications, and invasive care. There was a 5-fold difference in the rate of bleeding between patients in the lowest and highest CRUSADE bleeding risk groups, which was consistently observed in most anticoagulant groups. Conclusions There is a wide variability in the use of anticoagulant regimens with significant differences according to baseline characteristics and concomitant therapies. Major bleeding is common, though a great degree of the variability in the rate of bleeding is largely based on differences in baseline characteristics, comorbidities, and invasive treatment strategies, rather than specific anticoagulant regimens. (J Am Coll Cardiol Intv 2010;3: ) 2010 by the American College of Cardiology Foundation

2 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, Substantial clinical trial experience has compared different anticoagulant regimens in patients with ST-segment elevation myocardial infarction (STEMI) (1 3) and non STsegment elevation acute coronary syndrome (ACS) (4 8), yet relatively little is known about the current use of available agents in clinical practice. Anticoagulant therapy remains a cornerstone of management of ACS and is endorsed by practice guidelines for STEMI and unstable angina/non-stemi (NSTEMI) (9 12). Based primarily on clinical trial results, unfractionated heparin (UFH), low molecular weight heparin (LMWH), fondaparinux, and bivalirudin have all received the highest guideline recommendations in various clinical presentations of ACS (9 12). Compared with the general population, subjects in clinical trials are younger and by design have fewer of the comorbidities that increase their risk of complications of anticoagulant therapy, including bleeding. Thus, greater appreciation of real-world practice patterns of anticoagulant See page 1178 therapy and the associated risk of bleeding is important in understanding the circumstances and consequences of their use. We sought to examine the use of available anticoagulant agents across the spectrum of patients admitted with myocardial infarction in a contemporary cohort of over 121,000 patients in the NCDR ACTION Registry GWTG (National Cardiovascular Data Registry Acute Coronary Treatment and Intervention Outcomes Network Registry Get With the Guidelines). Methods ACTION Registry GWTG. The ACTION registry is a quality improvement registry of patients with myocardial infarction in the U.S. Participating sites enroll consecutive patients; therefore, it is an unselected population. Eligibility included patients with STEMI and NSTEMI. STEMI and NSTEMI are defined for registry inclusion by: 1) ischemic symptoms at rest, lasting 10 min, occurring within 24 h before admission or up to 72 h for STEMI; 2) electrocardiogram changes associated with STEMI (new left bundlebranch block or persistent ST-segment elevation 1 mm in 2 or more contiguous electrocardiographic leads); or 3) positive cardiac markers associated with NSTEMI (creatine kinase-mb isoenzyme or troponin I/T local laboratory upper limit of normal values) within 24 h after initial presentation. Patients are ineligible for the ACTION registry if they develop ischemic symptoms that meet the diagnostic criteria for STEMI and NSTEMI during hospitalization but were originally admitted for clinical conditions unrelated to STEMI and NSTEMI diagnosis (13). Hospitals in this registry are diverse in size, teaching status, capacity, and region. Participating hospitals collect data through retrospective chart review using standardized data collection tools that do not require direct contact with individual patients. Data collected include patient demographics, presenting features, pre-hospital therapy, inhospital therapy, hospital discharge therapy, timing of care delivery, laboratory tests, procedure use, and in-hospital patient outcomes. The Institutional Review Board of each hospital approved its organization s participation in the registry (13). The population included 51,980 patients with STEMI and 80,000 patients with NSTEMI enrolled at 360 U.S. hospitals from January 2007 to June Patients were excluded sequentially if they had a contraindication to receiving an anticoagulant agent or if the agent administered was blinded (n 762 in STEMI, n 3,390 in NSTEMI) and if anticoagulant agent used was missing or was not Abbreviations and Acronyms ACS acute coronary syndrome(s) GP glycoprotein LMWH low molecular weight heparin NSTEMI non ST-segment elevation myocardial infarction PCI percutaneous coronary intervention STEMI ST-segment elevation myocardial infarction UFH unfractionated heparin captured on the data collection instrument (n 2,112 in STEMI, n 3,624 in NSTEMI). Furthermore, due to the very infrequent use of fondaparinux, these patients were also excluded (n 163 [0.3%] for STEMI, n 287 [0.4%] for NSTEMI). The remaining study population includes 48,943 patients with STEMI and 72,699 patients with NSTEMI. From the *TIMI Study Group, Brigham and Women s Hospital and Harvard Medical School, Boston, Massachusetts; Cardiovascular Division, Brigham and Women s Hospital and Harvard Medical School, Boston, Massachusetts; and Duke Clinical Research, Institute and Division of Cardiology, Duke University Medical Center, Durham, North Carolina. The NCDR ACTION Registry GWTG is administered by the American College of Cardiology Foundation (ACCF) and sponsored by Bristol-Myers Squibb/Sanofi Pharmaceuticals Partnership, Genentech, and Schering-Plough Corporation, who provide material support for the operation of the data collection and infrastructure. The sponsors had no additional role in this project including the selection of topic, analysis of data, decision to publish, or approval of the manuscript before publication. Dr. Foody reports receiving consulting fees from Bristol-Myers Squibb, Merck, Pfizer, and Sanofi-Aventis. Dr. Cannon reports receiving research grants from Accumetrics, AstraZeneca, GlaxoSmithKline, Intekrin Therapeutics, Merck, and Takeda; he is a clinical adviser for and holds equity in Automedics Medical Systems; he is on the Advisory Board (but funds donated to charity) for Bristol-Myers Squibb/Sanofi, Novartis, and Alnylam; and he received honorarium for development of independent educational symposia: Pfizer, AstraZeneca. Dr. Wiviott reports receiving honoraria for educational presentations from Bristol-Myers Squibb, Daiichi Sankyo, Eli Lilly, Schering-Plough, Merck, and The Medicines Company; consulting fees from AstraZeneca, Bristol-Myers Squibb, and Sanofi-Aventis; and research grant support from Daiichi Sankyo, Eli Lilly, and Schering-Plough. Dr. Scirica has received honoraria for educational presentations from CV Therapeutics, Merck, Novartis, and Schering-Plough; consulting fees from AstraZeneca, Cogentus, and Novartis; and received research grant support from AstraZeneca, Bristol-Myers Squibb, CV Therapeutics, Merck, Daiichi Sankyo, Novartis, Johnson & Johnson, and Bayer Healthcare, as well as an unrestricted research grant from the Michael Lerner Foundation. All other authors have reported that they have no relationships to disclose. Manuscript received July 28, 2010, accepted August 6, 2010.

3 1168 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, 2010 For the major bleeding analysis, patients were also excluded sequentially if they were transferred to another facility as data were unavailable for non-action hospitals (n 8,172); were on home warfarin (n 5,504); had a baseline hematocrit recorded after a lowest hematocrit (n 474); had a baseline hematocrit recorded after a packed red blood cell transfusion (n 18); had missing major bleeding data (n 47); had missing age, sex, and/or race (n 1,174); or if they died within 48 h of presentation (n 1,795). There were 42,918 STEMI and 61,540 NSTEMI patients in the final analysis set. Definitions. Creatinine clearance was estimated with the Cockroft-Gault formula and was derived from data on admission, before any procedures (14). Major bleeding was determined if an event met 1 of the following criteria: 1) absolute hematocrit drop of 12%; 2) intracranial hemorrhage; 3) retroperitoneal bleeding; 4) red blood cell transfusion if baseline hematocrit 28%; or 5) red blood cell transfusion if baseline hematocrit 28% with report of a witnessed bleeding event. For patients undergoing surgical revascularization, criteria for bleeding had to occur before coronary artery bypass graft surgery. A listing of specific data fields and additional definitions are available on NCDR s website (15). Statistical analysis. STEMI and NSTEMI were analyzed separately. Patients were then divided into 5 groups based on anticoagulant strategy used during hospitalization: 1) no anticoagulant therapy; 2) UFH alone; 3) LMWH alone; 4) both UFH and LMWH; or 5) any bivalirudin (alone or in combination with other anticoagulant therapies) (Fig. 1). Bivalirudin patients were further divided based on whether they received bivalirudin alone or in combination with other anticoagulant therapies. Baseline patient demographics, medical history, presentation features, and concomitant therapies were examined across the treatment groups. Median values with interquartile ranges (IQR: 25th, 75th percentiles) were used to describe continuous variables, and numbers (percentages) were reported for categorical variables. Continuous and ordinal categorical variables were compared using the stratum adjusted Wilcoxon rank-sum test (2 samples) and Kruskal-Wallis test (more than 2 samples), whereas nominal categorical variables were compared using the stratum adjusted chi-square test where stratification is by hospital. Furthermore, selection of patients on each anticoagulant therapy was explored according to therapeutic strategy (e.g., primary percutaneous coronary intervention [PCI] and fibrinolytics for STEMI, and invasive and conservative strategy for NSTEMI). Finally, use of anticoagulants was also divided by quarter to examine time trends in usage. To evaluate the risk of bleeding observed in the different anticoagulant strategies, we used the CRUSADE (Can Rapid Risk Stratification of Unstable Angina Patients Figure 1. Flow of Subjects Included in Analysis From the ACTION Registry GWTG and Definitions of CRUSADE Bleeding Risk Score Anticoagulant strategy and CRUSADE bleeding risk score were used to categorize 121,642 patients with ST-segment elevation myocardial infarction (STEMI) or non ST-segment elevation myocardial infarction (NSTEMI). ACTION Acute Coronary Treatment and Intervention Outcomes Network; CHF congestive heart failure; CrCl creatinine clearance (determined with Cockroft-Gault formula); CRUSADE Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of ACC/AHA [American College of Cardiology/American Heart Association] Guidelines; DM2 diabetes mellitus type 2; GWTG Get With the Guidelines; Hct hematocrit; HR heart rate; LMWH low molecular weight heparin; SBP systolic blood pressure; UFH unfractionated heparin. Suppress Adverse Outcomes With Early Implementation of ACC/AHA [American College of Cardiology/American Heart Association] Guidelines) bleeding risk score to account for patient-specific variables known at baseline to affect bleeding rates. The construction of this score has previously been described and validated (16). Briefly, the CRUSADE bleeding risk score assigns points based on 8 criteria: 1) baseline hematocrit; 2) creatinine clearance; 3) heart rate; 4) sex; 5) signs of congestive heart failure; 6) systolic blood pressure; 7) prior vascular disease (prior stroke or peripheral artery disease); and 8) diabetes mellitus. Age is not included separately in this model as it is a component of the Cockroft-Gault formula for estimating creatinine clearance. In the original derivation and validation of the risk score, it did not remain a significant covariate in the model in the presence of other covariates and, therefore, was not included in the final model (16). Patients were divided into 5 groups based on the CRUSADE bleeding risk score ( 20, 21 to 30, 31 to 40, 41 to 50, and 50) (Fig. 1). Overall rates of major bleeding and were presented across the CRUSADE bleeding risk score categories for STEMI and NSTEMI. In addition, the distribution of the CRUSADE bleeding risk score categories by anticoagulant strategies was displayed in

4 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, patients with STEMI and NSTEMI. A graphical presentation of unadjusted rates of major bleeding according to anticoagulant and the CRUSADE bleeding risk score category was shown in patients with STEMI and NSTEMI. Rates of bleeding in STEMI by the CRUSADE bleeding risk score category were further stratified based on reperfusion strategy and use of anticoagulant strategies. Lastly, usage trends during this period (January 2007 to June 2009) were explored. A p value of 0.05 was considered statistically significant for all tests. All analyses were performed using SAS software (version 9.2, SAS Institute, Cary, North Carolina). Results Overall anticoagulant use. Among STEMI patients, UFH was the most commonly used anticoagulant agent (66%) followed by bivalirudin (14%) and LMWH (8%). In NSTEMI patients, UFH was also the most commonly used anticoagulant (42%) though there was a relative greater use of LMWH (27%) and bivalirudin (13%) (Figs. 2A and 2B). STEMI. Among the STEMI cohort, there were significant differences in baseline demographics between patients in each of the different anticoagulant groups (Table 1). In the STEMI cohort, patients who received LMWH were more likely to be 75 years of age, women, and have reduced renal function (estimated glomerular filtration rate 60 ml/min) than patients receiving either UFH or bivalirudin. Concomitant use of a glycoprotein (GP) IIb/IIIa inhibitor was more frequent in patients receiving UFH (79.1%) and LMWH (61.8%), though almost one-half (44.5%) of patients treated with bivalirudin also received a GP IIb/IIIa inhibitor (p ). Concomitant use of clopidogrel was more frequent in patients receiving bivalirudin (94.5%) than patients receiving UFH (88.0%) or LMWH (73.7%) (p ). Fibrinolytics were most commonly used alongside LMWH (27.9%) as opposed to UFH (11.6%) or bivalirudin (9.7%) (p ). However, fewer patients receiving LMWH had primary PCI (58.0%) than patients receiving UFH (81.8%) and bivalirudin (83.9%) (p ) (Table 1). In patients who received bivalirudin, there were significant differences in reperfusion strategy based on whether the patient received bivalirudin alone or in combination with other agents. Of patients who received bivalirudin alone, 92.7% underwent primary PCI versus 79.3% of patients who received bivalirudin in combination with other agents. In contrast, only 2.5% of patients who received bivalirudin alone underwent fibrinolytic therapy versus 13.5% of patients who received bivalirudin in combination with other agents (Online Appendix, Online Table 1A). Figure 2. Usage of Anticoagulant Regimens in STEMI and NSTEMI Patients The usage of each of the 5 anticoagulant strategies is presented in STEMI (A) and NSTEMI (B) patients. Abbreviations as in Figure 1. Anticoagulant in STEMI according to reperfusion strategy. Among patients undergoing primary PCI only (n 32,544, 79.9% of eligible STEMI patients without contraindications to reperfusion), UFH was the most commonly used anticoagulant (68.3%), followed by bivalirudin (16.8%), no anticoagulant agent (5.1%), both UFH and LMWH (5.0%), and LMWH (4.8%). In patients receiving a fibrinolytic (n 5,123, 12.6% of eligible STEMI patients without contraindications to reperfusion), UFH was also the most commonly used agent (61.6%), followed by LMWH (14.5%), bivalirudin (12.3%), and both UFH and LMWH (9.2%) (Fig. 3). NSTEMI. In the NSTEMI cohort, patients receiving LMWH were more likely to be elderly, women, or have

5 1170 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, 2010 Table 1. Baseline Characteristics and Concomitant Therapies by Anticoagulant Agent: STEMI Patients Overall None UFH LMWH UFH LMWH Bivalirudin Patients, n 48,943 2,794 32,214 4,044 2,946 6,945 Baseline characteristics Age 75 yrs, % Men, % Diabetes mellitus, % Prior vascular disease, % Signs of CHF, % Heart rate, beats/min* 78 (65 92) 79 (66 94) 78 (65 92) 79 ( ) 80 (66 96) 78 (65 92) SBP, mm Hg* 139 ( ) 137 ( ) 139 ( ) 138 ( ) 137 ( ) 140 ( ) BMI, kg/m 2 * 28.1 ( ) 27.8 ( ) 28.1 ( ) 28.0 ( ) 28.0 ( ) 28.3 ( ) Baseline hematocrit, % 42.9 ( ) 42.0 ( ) 43.0 ( ) 42.6 ( ) 42.5 ( ) 42.9 ( ) egfr, ml/min/1.73 m 2 * 85.5 ( ) 81.8 ( ) 86.3 ( ) 81.2 ( ) 79.8 ( ) 87.3 ( ) egfr 60 ml/min/1.73 m 2, % Concomitant therapies, % Clopidogrel mg mg mg mg Other GP IIb/IIIa inhibitor No reperfusion Fibrinolytic only Primary PCI only *Continuous variables displayed as median values with 25th, 75th percentiles. egfr determined with Cockroft-Gault formula; dialysis patients excluded from calculations. Clopidogrel dose information was missing for a small percentage of patients, thus the percentage of distribution by dose does not equal 100%. BMI body mass index; CHF congestive heart failure; egfr estimated glomerular filtration rate; GP glycoprotein; LMWH low molecular weight heparin; PCI percutaneous coronary intervention; SBP systolic blood pressure; STEMI ST-segment elevation myocardial infarction; UFH unfractionated heparin. impaired renal function than those patients receiving UFH or bivalirudin. Concomitant pharmacologic and invasive therapies also varied between groups. Those receiving bivalirudin were less likely to be simultaneously treated with a GP IIb/IIIa inhibitor (23.8%) than patients receiving UFH (51.4%) or LMWH (37.3%) (p ). Clopidogrel was used more frequently in the bivalirudin group (76.7%) than the UFH (62.0%) and LMWH (51.2%) groups (p ) (Table 2). In patients who received bivalirudin, there were no clinically significant differences between those who received bivalirudin alone or in combination with other agents (Online Appendix, Online Table 1B). Anticoagulant use according to invasive versus conservative strategy. Among patients undergoing an invasive strategy with cardiac catheterization with or without PCI in the first 48 h after admission (n 46,012, 74.2% of eligible NSTEMI patients without contraindications to cardiac catheterization), UFH was the most commonly used anticoagulant agent (46.0%), followed by LMWH (20.8%), bivalirudin (16.8%), and both UFH and LMWH (10.5%). In patients treated with a conservative strategy without cardiac catheterization or cardiac catheterization after the first 48 h of admission (n 15,631, 25.2% of eligible NSTEMI patients without contraindications to cardiac catheterization), LMWH was the most commonly used anticoagulant (35.0%) followed by UFH (34.6%), bivalirudin (9.4%), and both LMWH and UFH (9.2%) (Fig. 3). Time trends in anticoagulant use. Anticoagulant use was also examined over time, by quarter, from January 2007 to June In STEMI, bivalirudin use has risen from 10.2% to 20.9%, whereas UFH alone and LMWH alone have dropped (66.9% to 62.3% and 11.1% to 5.9%, respectively). In NSTEMI, bivalirudin use has increased from 10.2% to 16.5%, whereas LMWH alone has dropped from 30.8% to 23.4%. Use of UFH alone has remained stable at approximately 40% (Figs. 4A and 4B). Major bleeding. More than 50% of patients with either STEMI or NSTEMI were in the 2 lowest ( 20 or 21 to 30) of the 5 CRUSADE bleeding risk groups. More NSTEMI patients (33.7%), compared with only 17.3% of STEMI patients, were in the highest 2 risk groups (41 to 50 or 50) (Table 2). This was due to the NSTEMI population having a greater proportion of subjects who

6 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, Figure 3. Anticoagulant Use According to Therapeutic Strategy Use of anticoagulant by therapeutic strategy in both STEMI and NSTEMI were examined. PCI percutaneous coronary intervention; other abbreviations as in Figures 1 and 2. were women (38.5% in NSTEMI vs. 30% in STEMI) and with more comorbidities including diabetes mellitus (34.1% in NSTEMI vs. 22.4% in STEMI), prior vascular disease (18.6% in NSTEMI vs. 9.5% in STEMI), estimated glomerular filtration rate 60 (36.5% in STEMI vs. 23.6% in STEMI), signs of congestive heart failure (20.4% vs. 11.6% in STEMI) than the STEMI population. Despite their overall lower risk profile, patients with STEMI had higher rates of bleeding than those with NSTEMI across all the CRUSADE bleeding risk groups (Table 3). The distribution of the CRUSADE bleeding risk groups among each anticoagulant group is presented in Figures 5A and 5B. A higher percentage of STEMI patients in the lowest CRUSADE bleeding risk group ( 20) received UFH (45.6%) and bivalirudin (45.7%) than LMWH (38.8%). Conversely, a higher percentage of patients in the highest CRUSADE risk group ( 50) received LMWH (12.5%) compared with UFH (6.6%) and bivalirudin (6.5%). Overall, the greatest variation in anticoagulant use occurred in the low and high CRUSADE bleeding risk groups. In both STEMI and NSTEMI cohorts, there was roughly a 5-fold difference in the rate of bleeding between patients in the lowest and highest CRUSADE bleeding risk groups that was consistently observed in most anticoagulant groups. A similar gradient of bleeding risk was even observed among patients who received no anticoagulant therapy during their hospitalization (Figs. 6A and 6B). We further examined the rates of bleeding in STEMI by CRUSADE bleeding risk score based on reperfusion strategy selected (no reperfusion, fibrinolytics alone, primary PCI alone, or both fibrinolytics and PCI) and use of reperfusion strategies in STEMI based on the CRUSADE bleeding risk score. Among each subgroup, there is an increased rate of major bleeding events in the higher CRUSADE bleeding risk score categories. There is perhaps a trend toward higher rates of absolute bleeding seen within the fibrinolytics alone and primary PCI alone groups. There is a very small sample size of patients who received both fibrinolytics and PCI, and it is difficult to make any conclusions about this group (Online Appendix, Online Figs. 1A to 1E). Analysis of the usage of therapeutic strategies based on the CRUSADE bleeding risk score demonstrates that those in the

7 1172 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, 2010 Table 2. Baseline Characteristics and Concomitant Therapies by Anticoagulant Agent: NSTEMI Patients Overall None UFH LMWH UFH LMWH Bivalirudin Patients, n 72,699 6,420 30,611 19,425 6,957 9,286 Baseline characteristics Age 75 yrs, % Men, % Diabetes mellitus, % Prior vascular disease, % Signs of CHF, % Heart rate, beats/min* 83 (70 98) 85 (72 102) 82 (70 97) 84 (71 101) 83 (70 98) 80 (69 94) SBP, mm Hg* 145 ( ) 141 ( ) 145 ( ) 145 ( ) 146 ( ) 148 ( ) BMI, kg/m 2 * 28.4 ( ) 27.3 ( ) 28.4 ( ) 28.1 ( ) 28.8 ( ) 28.9 ( ) Baseline hematocrit, %* 41.0 ( ) 39.0 ( ) 41.4 ( ) 40.5 ( ) 41.4 ( ) 41.9 ( ) egfr, ml/min/1.73 m 2 * 74.0 ( ) 56.9 ( ) 77.1 ( ) 68.4 ( ) 78.4 ( ) 81.6 ( ) egfr 60 ml/min/1.73 m 2, % Concomitant therapies, % Clopidogrel mg mg mg mg Other GP IIb/IIIa inhibitor Catheterized within 48 h PCI *Continuous variables displayed as median values with 25th, 75th percentiles. egfr determined with Cockroft-Gault formula; dialysis patients excluded from calculations. Clopidogrel dose information was missing for a small percentage of patients, thus the percentage of distribution by dose does not equal 100%. NSTEMI non ST-segment elevation myocardial infarction; other abbreviations as in Table 1. higher CRUSADE bleeding risk score categories were less likely to be treated with reperfusion therapy than were those in the lower CRUSADE bleeding risk score groups (primary PCI: 77.5% vs. 81.7%; fibrinolytics 8.7% vs. 13.1%; no reperfusion: 13.4% vs. 4.3% in the high CRUSADE bleeding risk score 50 group vs. the low [ 20] group) (Online Appendix, Online Table 2). Discussion In this large contemporary cohort of patients admitted to U.S. hospitals with myocardial infarction, we find that there is a wide variability in provider use of anticoagulant regimens with significant differences according to baseline characteristics and concomitant therapies. Unfractionated heparin remains the most commonly used anticoagulant in current practice and is given in over two-thirds of patients with STEMI and nearly one-half with NSTEMI. The translation of clinical trial data into clinical practice is complex, yet patterns of use of anticoagulants relatively preferred with invasive care among NSTEMI and STEMI (bivalirudin and UFH) and conservative care (LMWH) among NSTEMI are discernable. Major bleeding is common among patients with STEMI and NSTEMI, though a large degree of the variability in the rate of bleeding is likely based on differences in baseline characteristics and comorbidities as categorized using the CRUSADE bleeding risk score and invasive treatment strategies, rather than specific anticoagulant regimens. Multiple randomized studies have compared UFH to LMWH, bivalirudin, and fondaparinux, and many have shown superiority of the latter agents in both STEMI and NSTEMI. One exception is the SYNERGY (Superior Yield of the New Strategy of Enoxaparin, Revascularization, and Glycoprotein IIb/IIIa Inhibitors) trial, which showed no difference in ischemic outcomes between UFH and LMWH and increased bleeding with LMWH in a population that underwent an early invasive strategy (5). Despite most trials showing benefit for agents other than UFH, UFH remains the most commonly used agent in ACS. In our analysis, even patients who underwent conservative strategy (a population different from that of the SYNERGY trial and 1 in which UFH has not ever been shown to be superior), there was significant use of UFH. Consensus guidelines integrate the results of multiple trials but are often unable to provide definitive guidance. Thus they give high recommendations for multiple anticoagulant strategies and leave open the choice among the 5

8 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, Figure 4. Time Trends in Use of Anticoagulant Regimens in STEMI and NSTEMI Patients Usage of each of the 5 anticoagulant strategies were examined individually in each quarter from January 2007 to June 2009 for STEMI (A) and NSTEMI (B) patients. Trends in use are plotted by anticoagulant strategy. Q quarter; other abbreviations as in Figures 1, 2, and 3. available agents. Patients in the ACTION registry were older and more likely to have comorbidities including diabetes, hypertension, or a prior myocardial infarction than patients in several of the randomized clinical trials of anticoagulant strategies (1,2,4,6,8). Thus, physicians may feel the results of these trials are less generalizable to the

9 1174 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, 2010 Table 3. Distribution of CRUSADE Bleeding Risk Scores and Associated Bleeding Rates STEMI, n 42,918 NSTEMI, n 61,540 CRUSADE Bleeding Score n (%) Major Bleeding Events (%) 20 19,050 (44.4) 1,036 (5.4) ,864 (23.0) 1,002 (10.2) ,565 (15.3) 956 (14.6) ,250 (9.9) 857 (20.2) 50 3,189 (7.4) 928 (29.1) 20 19,803 (32.2) 605 (3.1) ,134 (18.1) 670 (6.0) ,831 (16.0) 822 (8.4) ,944 (14.5) 1,099 (12.3) 50 11,828 (19.2) 2,314 (19.6) CRUSADE Can Rapid Risk Stratification of Unstable Angina Patients Suppress Adverse Outcomes With Early Implementation of ACC/AHA [American College of Cardiology/American Heart Association] Guidelines; other abbreviations as in Tables 1 and 2. overall clinical population and, therefore, may choose UFH, an agent they are most comfortable using. Not surprisingly, there was an observed association between invasive strategies and type of anticoagulant used. Patients undergoing invasive strategies were relatively more likely to receive UFH or bivalirudin, due to their easy titratability and ability to control levels of anticoagulation during complicated procedures. LMWH was less frequently used in patients undergoing catheterization despite data that suggest that it is efficacious and safe in this setting (17,18). LMWH was the preferred agent among patients with NSTEMI managed with a conservative strategy. Additionally, those receiving bivalirudin or UFH were more likely to receive concomitant clopidogrel, also suggesting that these anticoagulant agents were more commonly used as a part of an overall treatment strategy that included more potent antiplatelet agents and intervention. When looking at time trends, the increase in bivalirudin usage over the last year was notable. This is likely related to the publication of the ACUITY (Acute Catheterization and Urgent Intervention Triage Strategy Trial) and HORIZONS-AMI (Harmonizing Outcomes with Revascularization and Stents in Acute Myocardial Infarction) trials in November 2006 and May 2008, respectively. There were some notable examples of real-world practice not supported by clinical evidence. For example, 28.4% of STEMI patients and 41.3% of NSTEMI patients who received LMWH had estimated glomerular filtration rate 60 ml/min/1.73 m 2, even though using LMWH in renal insufficiency is associated with an increased risk of bleeding. Another example was patients who receive both bivalirudin and GP IIb/IIIa inhibitors simultaneously. Although fewer patients on bivalirudin received GP IIb/IIIa inhibitors than those on other anticoagulants, it was still a sizeable result and not consistent with recent clinical trial data. Additionally, despite the concern regarding bleeding associated with the crossover between LMWH and UFH in the SYNERGY and ACUITY trials (5,19) and guideline recommendations to avoid crossing over, between 5% and 10% of patients received both agents during their hospitalization. The reason for these practice patterns is unclear, but the need for further education and dissemination of clinical trial results to all practicing physicians persists. The GRACE (Global Registry of Acute Coronary Events) registry reported anticoagulant use in 12,665 patients admitted with ACS between 1999 and 2001, a period almost 8 years earlier than the cohort evaluated in this analysis (20). Despite the difference in time and the introduction of bivalirudin, the use of UFH is essentially identical in patients with STEMI (66.8% in the GRACE registry receiving UFH compared with 66% in the ACTION registry). Comparing the use of LMWH between registries is more complicated because, in contrast to the ACTION registry, the GRACE registry predominately includes non-u.s. sites, in which LMWH is more commonly used. However, the use of LMWH in the ACTION registry does appear to have increased from 13% reported in U.S. sites from the GRACE registry to almost 30% in patients with NSTEMI in the ACTION registry. A concern in clinical practice is the comparative risk of major bleeding as the result of treatment strategies. We found that the risk of bleeding in patients with ACS varies significantly according to baseline characteristics and use of invasive strategies, rather than specific anticoagulant regimen used. This is an important consideration for clinicians as concern for bleeding events contributes significantly to therapeutic decision making during ACS. Thus, a careful assessment of the patient s baseline risk of bleeding relative to the potential benefit may be more important than deciding between different anticoagulants. In our analysis, we both confirm the validity of the CRUSADE bleeding risk score to predict bleeding in patients with NSTEMI and extend the application of the bleeding score to patients with STEMI in which there is a clear stepwise increase in bleeding with each higher risk group. Interestingly, the overall incidence of bleeding in each risk group is higher in STEMI than NSTEMI despite that patient population being younger and having fewer comorbidities. This may be because compared with NSTEMI patients, STEMI patients were more universally managed with cardiac catheterization within 48 h of arrival (92.0% vs. 74.2% in NSTEMI) and PCI (86.7% vs. 55.3% in NSTEMI), again emphasizing the importance of invasive strategies in increasing bleeding risk. Additionally, the CRUSADE bleeding risk score was developed in NSTEMI patients, so it may not fully translate to STEMI patients.

10 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, Figure 5. Distribution of CRUSADE Bleeding Risk Scores by Anticoagulant Strategy in Patients With STEMI and NSTEMI Patients were stratified by anticoagulant strategy and distribution of CRUSADE bleeding risk score categories was examined in each subgroup: (A) STEMI, and (B) NSTEMI. Abbreviations as in Figures 1, 2, 3, and 4. We specifically do not compare the rates of bleeding between different anticoagulant treatment regimens because the large differences in baseline characteristics and concomitant therapy between patients receiving different anticoagulants would bias and confound any true association in this observational study. Some may find the lower rates of bleeding in patients treated with bivalirudin consistent with clinical trial data (2,6). However, there is a similarly low rate of bleeding in patients treated with LMWH, compared with UFH, that is also contrary to randomized trial results (1). Additionally, there was an increased use of GP IIb/IIIa inhibitors in all groups

11 1176 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, 2010 Figure 6. Rates of Major Bleeding According to Anticoagulation Regimen and CRUSADE Bleeding Risk Score Category in Patients With STEMI and NSTEMI Patients were first divided by anticoagulant strategy. Within each anticoagulant strategy, patients were divided by CRUSADE bleeding risk score and rates of major bleeding were observed. (A) STEMI; (B) NSTEMI. Abbreviations as in Figures 1, 2, 3, 4, and 5. relative to the bivalirudin group, possibly increasing bleeding rates in these groups. This highlights the confounding that can arise when comparing nonrandomized treatment regimens. Study limitations. Limitations of this analysis are that this is not a randomized control cohort, and thus associations between baseline characteristics, anticoagulant use, and outcomes is likely to be confounded. Additionally, the CRUSADE bleeding risk score model was developed in NSTEMI patients, and, although we apply it to STEMI patients as well, it is not validated in this group. We could not determine the rate of bleeding based on GUSTO (Global Utilization of Streptokinase and Tissue Plasminogen Activator for Occluded Coronary Arteries) trial and TIMI (Thrombolysis In Myocardial Infarction) criteria because the ACTION registry does not collect all the required criteria for these definitions. However, the major bleeding definition used in the ACTION registry was applied similarly to all patients, thus there should be no significant classification bias between treatment regimens. The data come from hospitals that report to the ACTION registry, and though this does reflect a variety of hospitals by size, location, private/public, and academic/ nonacademic, it does not reflect all hospitals in the U.S. Finally, these data are only from the U.S., and global practice patterns likely vary. Conclusions This large, contemporary cohort with myocardial infarction provides insight into real-world use of anticoagulant strategies in ACS. There are significant differences in anticoagulant agent usage, which can be explained by patient characteristics and concomitant pharmacologic and invasive therapies. Use of anticoagulant therapy not consistent with clinical guidelines appears to be present in current practice and, therefore, identifies an important focus for quality improvement programs. Risk of bleeding

12 JACC: CARDIOVASCULAR INTERVENTIONS, VOL. 3, NO. 11, can be evaluated using a simple risk score in both NSTEMI and STEMI, and across anticoagulant strategies, providing important prognostic information for the clinician. Variability in rates of bleeding is likely based on differences in baseline characteristics, comorbidities, and invasive treatment strategies rather than specific anticoagulation regimens. Future clinical trials perhaps should also categorize subjects by bleeding risk (as is already done for ischemic risk, that is, TIMI and GRACE risk scores) to understand better any heterogeneity in risk of bleeding for a particular agent. A better understanding of how clinicians choose anticoagulant regimens among available recommended options would be important in improving the translation of clinical trial data into real-world practice. Reprint requests and correspondence: Dr. Benjamin M. Scirica, Cardiovascular Division, Brigham and Women s Hospital, 75 Francis Street, Boston, Massachusetts bscirica@ partners.org. REFERENCES 1. Antman EM, Morrow DA, McCabe CH, et al., on behalf of ExTRACT-TIMI 25 Investigators. Enoxaparin versus unfractionated heparin with fibrinolysis for ST-elevation myocardial infarction. N Engl J Med 2006;354: Stone GW, Witzenbichler B, Guagliumi G, et al., on behalf of HORIZONS-AMI Trial Investigators. Bivalirudin during primary PCI in acute myocardial infarction. N Engl J Med 2008;358: Yusuf S, Mehta SR, Chrolavicius S, et al., on behalf of OASIS-6 Trial Group. Effects of fondaparinux on mortality and reinfarction in patients with acute ST-segment elevation myocardial infarction: the OASIS-6 randomized trial. JAMA 2006;295: Antman EM, McCabe CH, Gurfinkel EP, et al. Enoxaparin prevents death and cardiac ischemic events in unstable angina/non Q-wave myocardial infarction. Results of the Thrombolysis In Myocardial Infarction (TIMI) 11B trial. Circulation 1999;100: Ferguson JJ, Califf RM, Antman EM, et al., on behalf of SYNERGY Trial Investigators. Enoxaparin vs unfractionated heparin in high-risk patients with non ST-segment elevation acute coronary syndromes managed with an intended early invasive strategy: primary results of the SYNERGY randomized trial. JAMA 2004;292: Stone GW, McLaurin BT, Cox DA, et al., on behalf of ACUITY Investigators. Bivalirudin for patients with acute coronary syndromes. N Engl J Med 2006;355: Yusuf S, Mehta SR, Chrolavicius S, et al., on behalf of Fifth Organization to Assess Strategies in Acute Ischemic Syndromes Investigators. Comparison of fondaparinux and enoxaparin in acute coronary syndromes. N Engl J Med 2006;354: Cohen M, Demers C, Gurfinkel EP, et al. Low-molecular-weight heparins in non ST-segment elevation ischemia: the ESSENCE trial. Efficacy and Safety of Subcutaneous Enoxaparin versus intravenous unfractionated heparin, in non Q-wave Coronary Events. Am J Cardiol 1998;82:19L 24L. 9. Anderson JL, Adams CD, Antman EM, et al. ACC/AHA 2007 guidelines for the management of patients with unstable angina/non ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 2002 Guidelines for the Management of Patients With Unstable Angina/Non ST-Elevation Myocardial Infarction). J Am Coll Cardiol 2007;50:e Antman EM, Hand M, Armstrong PW, et al focused update of the ACC/AHA 2004 guidelines for the management of patients with ST-elevation myocardial infarction: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2008;51: Bassand JP, Hamm CW, Ardissino D, et al., on behalf of Task Force for Diagnosis and Treatment of Non ST-Segment Elevation Acute Coronary Syndromes of European Society of Cardiology. Guidelines for the diagnosis and treatment of non ST-segment elevation acute coronary syndromes. Eur Heart J 2007;28: Van de Werf F, Bax J, Betriu A, et al., on behalf of Task Force on the Management of ST-Segment Elevation Acute Myocardial Infarction. ESC guidelines on management of acute myocardial infarction in patients presenting with persistent ST-segment elevation. Rev Esp Cardiol 2009;62:293e Peterson ED, Roe MT, Rumsfeld JS, et al. A call to ACTION (Acute Coronary Treatment and Intervention Outcomes Network): a national effort to promote timely clinical feedback and support continuous quality improvement for myocardial infarction. Circ Cardiovasc Qual Outcomes 2009;2: Cockroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976;16: National Cardiovascular Data Registry. What Data Is Collected in the ACTION Registry GWTG? [downloadable forms]. April December Available at: ACTION/Elements.aspx. Accessed February 1, Subherwal S, Bach RG, Chen AY, et al. Baseline risk of major bleeding in non-st-segment-elevation myocardial infarction: the CRUSADE (Can Rapid risk stratification of Unstable angina patients Suppress ADverse outcomes with Early implementation of the ACC/AHA Guidelines) bleeding score. Circulation 2009;119: Montalescot G, White HD, Gallo R, et al., on behalf of STEEPLE Investigators. Enoxaparin versus unfractionated heparin in elective percutaneous coronary intervention. N Engl J Med 2006;355: Gibson CM, Murphy SA, Montalescot G, et al., on behalf of ExTRACT-TIMI 25 Investigators. Percutaneous coronary intervention in patients receiving enoxaparin or unfractionated heparin after fibrinolytic therapy for ST-segment elevation myocardial infarction in the ExTRACT-TIMI 25 trial. J Am Coll Cardiol 2007;49: White HD, Chew DP, Hoekstra JW, et al. Safety and efficacy of switching from either unfractionated heparin or enoxaparin to bivalirudin in patients with non-st-segment elevation acute coronary syndromes managed with an invasive strategy: results from the ACUITY (Acute Catheterization and Urgent Intervention Triage strategy) trial. J Am Coll Cardiol 2008;51: Budaj A, Brieger D, Steg PG, et al., on behalf of GRACE Investigators. Global patterns of use of antithrombotic and antiplatelet therapies in patients with acute coronary syndromes: insights from the Global Registry of Acute Coronary Events (GRACE). Am Heart J 2003;146: Key Words: anticoagulants bleeding heparin myocardial infarction registries. APPENDIX For supplemental data, please see the online version of this article.