Peter A. Kavsak, 1* Andrew R. MacRae, 2 Glenn E. Palomaki, 3 Alice M. Newman, 4 Dennis T. Ko, 4 Viliam Lustig, 2 Jack V. Tu, 4 and Allan S.

Transcription

1 Clinical Chemistry 52: (2006) Evidence-Based Laboratory Medicine and Test Utilization Health Outcomes Categorized by Current and Previous Definitions of Acute Myocardial Infarction in an Unselected Cohort of Troponin-Naïve Emergency Department Patients Peter A. Kavsak, 1* Andrew R. MacRae, 2 Glenn E. Palomaki, 3 Alice M. Newman, 4 Dennis T. Ko, 4 Viliam Lustig, 2 Jack V. Tu, 4 and Allan S. Jaffe 5 Background: In a population originally classified for acute myocardial infarction (AMI) by the World Health Organization (WHO) definition, we compared the health outcomes after retrospectively reclassifying with the European Society of Cardiology and the American College of Cardiology (ESC/ACC) AMI definition, using the peak cardiac troponin I (ctni) concentrations. The health outcomes were based on the WHO definition and occurred in an era that preceded the use of cardiac troponin biomarkers. Methods: For 448 patients who presented to the emergency department with symptoms suggestive of cardiac ischemia in 1996, we obtained data for all-cause mortality and recurrent AMI for up to 1 year after the initial presentation. We performed retrospective analysis of the patients frozen plasma samples to measure ctni (AccuTnI, Beckman Coulter). Results: At 30, 120, and 360 days, the risk for AMI/death in patients positive for AMI by only the ESC/ACC criteria was significantly lower than the risk in patients positive by both ESC/ACC and WHO criteria, and significantly higher than in patients negative according to both criteria. In a separate analysis, patients with a 1 Department of Pathology and Molecular Medicine, McMaster University, Hamilton, ON, Canada. 2 Department of Laboratory Medicine and Pathobiology and 4 Institute for Clinical Evaluative Sciences, University of Toronto, Toronto, ON, Canada. 3 Department of Pathology, Women and Infants Hospital of Rhode Island, Providence, RI. 5 Cardiovascular Division and Division of Laboratory Medicine, Mayo Clinic, Rochester, MN. * Address correspondence to this author at: McMaster University Medical Centre, 1200 Main St. W., HSC 2N52, Hamilton, ON L8N 3Z5, Canada. Fax ; kavsakp@mcmaster.ca. Received May 11, 2006; accepted August 17, Previously published online at DOI: /clinchem peak ctni >0.10 g/l were at greater risk for AMI/death than patients with ctni concentrations of g/l. Patients negative by both definitions or with peak ctni concentrations <0.04 g/l had the highest eventfree survival rates (92% and 94%, respectively) at 1 year. Conclusion: In a troponin-naïve population, patients classified as positive for AMI by only the ESC/ACC criteria have a prognosis that appears to be intermediate between those classified positive by both the WHO and ESC/ACC definitions and those who meet neither criteria American Association for Clinical Chemistry In 2000, the European Society of Cardiology/American College of Cardiology (ESC/ACC) 6 modified the criteria defining acute myocardial infarction (AMI) to rely more substantially on measurements of a highly sensitive biomarker, cardiac troponin (ctn) (1). As expected, this newer definition resulted in an increased prevalence of AMI compared with the previous WHO criteria (2, 3); however, whether the risk of death or recurrent AMI in this newly defined AMI group was as high as for those diagnosed by the previous WHO criteria remained unclear (4, 5). Two reports provided early risk estimates. One study suggested an increased risk of all-cause mortality and the combined endpoint of cardiovascular death or nonfatal AMI for those with AMI defined by the ESC/ACC criteria (5). The other found no differences in 6 Nonstandard abbreviations: AMI, acute myocardial infarction; ESC, European Society of Cardiology; ACC, American College of Cardiology; ctn, cardiac troponin; ctni, cardiac troponin I; NN, negative according to both WHO and ESC/ACC definitions; NP, negative according to WHO definition and positive according to ESC/ACC definition; PP, positive according to both WHO and ESC/ACC definitions; STEMI, ST-segment elevation myocardial infarction; MI, myocardial infarction; AHA, American Heart Association. 2028

2 Clinical Chemistry 52, No. 11, major adverse cardiac events between the newly defined group and those meeting the criteria of both definitions (4). Neither study measured the relative risk of recurrent AMI alone for either group. The purpose of our study was to make an unbiased assessment of the risk of recurrent AMI and/or all-cause mortality in patients negative by both definitions, positive by both, or positive only by the ESC/ACC criteria. Because our study consisted of patients originally classified and treated in 1996 on the basis of the WHO criteria, without knowledge of the ctn concentrations, the results were not affected by currently available interventions such as percutaneous coronary intervention and antiplatelet therapies that are now commonly used in response to ctn increases (6). This type of study is important because it is not known whether the new ESC/ACC criteria led to identification of more low-risk AMI patients, thus improving the overall prognosis of AMI, or more high-risk patients, in which case the overall prognosis may deteriorate (4, 5). Evidence from an unbiased population is essential to allow for continued epidemiologic surveillance of populations. Furthermore, our retrospective measurement of ctni also allows a comparison of different ctni concentrations grouped according to published cutoff values (7). Materials and Methods patient selection The study population and its characteristics have been reported previously (8, 9). Briefly, the study cohort consisted of 448 unique patients who were enrolled in 1996 as part of an emergency department cardiac marker study at Oshawa General Hospital (now Lakeridge Health Oshawa). The study included all patients presenting with symptoms suggestive of cardiac ischemia as assessed by triage staff (8, 9). The leftover plasma specimens from 1996 were frozen at 70 C, with the exception of brief periods when they were at 20 C. In 2003, the specimens were thawed and analyzed for cardiac troponin I (ctni) by the AccuTnI assay (Beckman Coulter Inc.) without knowledge of the previous clinical diagnoses (8, 9). Specimen stability with the AccuTnI troponin assay after multiple freeze-thaw cycles has been well documented (10, 11). This study received research ethics approval. ami classification The WHO-based AMI group in our study population was obtained in 1996 by independent review of charts and creatine kinase isoenzyme MB (CKMB) mass biomarker concentrations by a cardiologist and a specialist in emergency medicine on the basis of the WHO MONICA (Multinational MONItoring of trends and determinants in CArdiovascular disease) project criteria (8). For the retrospective ESC/ACC diagnosis, we evaluated the peak ctni concentration in each patient against 2 cutoffs: the 99th percentile (0.04 g/l) and 10% CV (0.06 g/l). Concentrations exceeding the cutoff were classified as AMI positive (8, 9, 12). For 24% of our population (n 108) only 1 specimen, collected at presentation, was available, and the measured ctni of this specimen was considered the peak concentration. On the basis of the 99th percentile cutoff and the ESC/ACC definition for AMI, AMI was excluded in 307 of the 448 patients by both WHO and ESC/ACC criteria [negative according to both WHO and ESC/ACC definitions (NN); Fig. 1A]; 80 patients, all of whom had non ST-segment elevation myocardial infarction (STEMI), met only the ESC/ACC definition [negative according to WHO and positive according to ESC/ACC (NP)], and 60 patients met both definitions for AMI [positive according to both definitions (PP); this group included all 31 patients with STEMI]. Only 1 patient was classified as positive by the WHO definition and negative by the ESC/ACC definition. This patient had only 1 specimen measured, with CKMB 15 g/l and ctni 0.02 g/l; this result could reflect either a noncardiac source for CKMB or a sample obtained very soon after onset (13). The median ctni concentrations in the NN, NP, and PP groups were 0.01, 0.11, and 5.38 g/l, respectively. The 99th percentile is the desired cutoff for AMI classification (1, 14). The 99th percentile is derived from a reference control group that did not take sex into account. One study, however, has reported a sex difference in the 99th percentile with the AccuTnI assay (e.g., male 99th 0.10 g/l vs female 99th 0.04 g/l; P 0.034) (7). We therefore used the peak ctni concentrations to assemble our study cohort into 3 groups based on the lowest and highest reported 99th percentiles: a low-ctni group (i.e., ctni 0.04 g/l; also the manufacturer s indicated 99th percentile), an intermediate-ctni group (ctni g/l), and a high-ctni group (ctni 0.10 g/l). The median peak ctni concentrations and numbers of patients in the 3 groups were 0.01, 0.05, and 1.12 g/l in the low (n 284), intermediate (n 65), and high (n 99) groups, respectively. health outcomes Research ethics approval was granted to obtain all-cause mortality and recurrent AMI for up to 1 year after the initial presentation via linkage with the Canadian Institute for Health Information database (containing information on all hospital discharges in Ontario) and the registered persons database for mortality outcomes. The Canadian Institute for Health Information database has been validated as a source for obtaining myocardial infarction (MI) outcomes in Ontario hospitals (15). The database captures hospital discharge abstracts and includes information on sex, date of birth, date of hospital admission, admitting institution, and most responsible diagnosis, with the diagnoses coded via the International Classification of Diseases (9th revision) coding scheme (e.g., code for AMI is 410) (15).

3 2030 Kavsak et al.: Health Outcomes by AMI Definitions Fig. 1. Kaplan Meier curves for mortality from all causes and/or subsequent AMI. Among 448 patients stratified by 2 definitions of AMI (A), patients were stratified into 3 groups (NN, NP, and PP). Solid line indicates use of the 99th percentile cutoff level for ctni in the ESC/ACC criteria; dotted line indicates use of the 10% CV cutoff level for ctni in the ESC/ACC criteria. (B), ctni levels. Patients were stratified into 3 groups (low: peak ctni 0.04 g/l; intermediate: g/l; high: 0.10 g/l). statistical analysis All statistical analyses were performed with SAS, version Privacy constraints prohibit the display of cells from groups of 6 individual patients. A P value of 0.05 was considered statistically significant. Between-group comparisons of central tendency (means and medians) were based on 1-way ANOVA and the Kruskal Wallis test. The Pearson 2 test statistic was used to compare proportions unless the expected value was 5, in which case the Fisher exact test was substituted. Percentages were based on group totals, with the exception of readmission rates for AMI, for which the denominator was the number of patients who had not died up to the time point of interest. Time to an adverse event was assessed by Kaplan Meier survival curves with differences between groups determined by the log rank test. The Cox proportional hazard model was used for comparing adjusted survival curves (adjusted for age and sex). Logistic regression analysis (adjusted for age 65, time from onset of symptoms to hospital presentation of 6 h, sex, and previous MI) was performed to assess the role of ctni as a predictor of adverse outcome, with the odds ratio reflecting the strength of association and 95% confidence intervals indicating statistical significance if they did not include The C statistic, which is equivalent to the area under the ROC curve, was used to compare the predictive accuracy. The Kaplan Meier and Cox proportional hazard analyses were performed on the basis of the AMI groups (NN, NP, and PP) and the ctni groups (low, intermediate, and high). The Cox proportional hazard model was used for comparing adjusted survival curves of the PP, NP, and NN groups. The 99th percentile (ctni 0.04 g/l) was used to classify patients positive for AMI with the ESC/ ACC definition. In addition, hazard ratios were provided for the intermediate and high-ctni groups vs the low-ctni group. In a separate analysis, the Cox proportional hazard

4 Clinical Chemistry 52, No. 11, model was used for comparing adjusted survival curves of either the NP group or intermediate ctni group. The logistic regression analyses were performed with either the presentation or peak ctni concentration. Results The classification of our study population on the basis of the American Heart Association (AHA) definition, which is intended to be used for clinical and epidemiologic Variable Value Table 1. Study population. AHA case definition a n 228 At least 2 biomarker measurements b n Biomarker measurement n 98 P value Total population c n 448 Demographics Age at presentation Mean (SD) (12.97) (15.19) (15.29) (14.36) Median (IQR) 67 (56 76) 54 (46 72) 60 (49 73) (51 74) Sex, n (%) F 96 (42.1) 42 (37.8) 40 (40.8) (40.8) M 132 (57.9) 69 (62.2) 58 (59.2) 265 (59.2) Previous MI, n (%) No 162 (71.1) 90 (81.1) 77 (78.6) (75.2) Yes 66 (28.9) 21 (18.9) 21 (21.4) 111 (24.8) In-hospital course, n (%) Thrombolytic therapy No 209 (91.7) 107 (96.4) 97 (99.0) Yes 19 (8.3) 5 (3.6) 5 (1.0) ASA No 170 (74.6) 97 (87.4) 81 (82.7) (79.7) Yes 58 (25.4) 14 (12.6) 17 (17.3) 91 (20.3) Nitro No 96 (42.1) 74 (66.7) 71 (72.4) (55.1) Yes 132 (57.9) 37 (33.3) 27 (27.6) 201 (44.9) Length of hospital stay, days Mean (SD) 6.04 (7.36) 1.85 (4.67) 1.65 (3.53) (6.36) Median (IQR) 4 (2 7) 0 (0 0) 0 (0 2) (0 6) AMI Diagnosis, n (%) 1996 Expert Clinical Diagnosis No 182 (79.8) 105 (94.6) 91 (92.9) (86.4) Yes 46 (20.2) 6 (5.4) 7 (7.1) 61 (13.6) 1994 WHO MONICA (1 CKMB No 179 (78.5) 105 (94.6) 92 (93.9) (85.7) measurement 2XULN) Yes 49 (21.5) 6 (5.4) 6 (6.1) 64 (14.3) ESC/ACC 2000 (1 CKMB No 171 (75.0) 104 (93.7) 92 (93.9) (83.7) measurement 99th) Yes 57 (25.0) 7 (6.3) 6 (6.1) 73 (16.3) ESC/ACC 2000 (1 ctni No 126 (55.3) 92 (82.9) 83 (84.7) (68.8) measurement 99th) Yes 102 (44.7) 19 (17.1) 15 (15.3) 140 (31.3) ctni peak terciles ctni 0.04 g/l 110 (48.2) 88 (79.3) 79 (80.6) (63.4) ctni g/l 43 (18.9) 11 (9.9) 11 (11.2) 65 (14.5) ctni 0.10 g/l 75 (32.9) 12 (10.8) 8 (8.2) 99 (22.1) Group based on WHO and ESC/ACC definition NN negative for both NN 126 (55.3) 92 (82.9) 82 (84.5) (68.7) NP negative for WHO but NP 56 (24.6) 13 (11.7) 9 (9.3) 80 (17.9) positive for ESC/ACC PP positive for both PP 46 (20.2) 6 (5.4) 6 (6.2) 60 (13.4) Outcomes, n (%) Death within 30 d of presentation 6 (2.6) Death within 1 yr of presentation 29 (12.7) 7 (6.3) 10 (10.2) (10.3) Admitted for AMI within 30 d of 15 (6.8) presentation Admitted for AMI within 1 year of 29 (14.1) 6 6 (6.7) presentation Either death or AMI within 30 d 21 (9.2) 7 (6.3) 7 (7.1) (8.0) Either death or AMI within 1 year 52 (22.8) 10 (9.0) 14 (14.3) (17.4) a 2 specimens at least 6 h apart. Time interval between specimens 6 h. c Includes 11 patients who presented 24 h after onset of pain.

5 2032 Kavsak et al.: Health Outcomes by AMI Definitions research, is shown in Table 1 (2). The study population is presented as those patients who fulfilled the AHA definition of an adequate set of biomarkers, as well as those who did not have the minimum time length between specimens and those with only 1 specimen. For privacy reasons, values for the total columns are suppressed if any of the 3 groups consisted of 6 patients. Accordingly, the 11 patients with a self-reported onset of pain 24 h before presentation are not presented as a group. Only 51% of the patients (228 of 448; Table 1) conform to the AHA requirement of 2 specimens at least 6 h apart; the other half of our population had specimens collected with an interval 6 h, had only 1 specimen collected, or presented to the emergency department after 24 h from their selfreported onset of symptoms. For this reason, we opted to classify our population according to the maximum concentration of ctn, consistent with the ESC/ACC definition (1, 7). The Kaplan Meier curves constructed for the NN, NP, and PP groups (Fig. 1A) with the 99th percentile cutoff show clear differences in event-free (i.e., absence of AMI/ death) survival after 1 year between the 3 groups (92%, 70%, and 50%, respectively; P 0.001). The ctni cutoff concentrations of 0.04 g/l (99th percentile) and 0.06 g/l (10% CV cutoff) led to nearly identical outcomes, as expected from the proximity of these 2 cutoffs with this ctni assay. The groups classified according to peak ctni concentrations (low, intermediate, and high; Fig. 1B) also displayed differences in event-free survival after 1 year (low cutoff, 94%; intermediate cutoff, 75%; high cutoff, 55%; P 0.001). Cox proportional hazard modeling was used to compare adjusted survival curves of the AMI defined groups and the peak ctni groups (Tables 2 and 3). Setting either the NN group or the low-ctni group as the reference, both the NP and PP groups, or the intermediate and high-ctni groups, had significant hazard ratios for the combined endpoint of death/ami recurrence at 30, 120, and 360 days (Table 2). When the NP group or the intermediate ctni group was set as the reference, significant proportional hazards ratios associated with AMI recurrence and the combined endpoint of death or AMI occurred at 30, 120, and 360 days for the PP and high-ctni groups. In contrast, the NN and low-ctni groups were at less risk for the combined endpoint of death or AMI at 30, 120, and 360 days (Table 3). In our population, after we adjusted for age, sex, time from symptom onset to presentation, and history of previous MI, logistic regression analysis demonstrated that a peak ctni concentration 0.10 g/l was a strong predictor for both AMI recurrence and death at 1 year. Use of the presentation ctni values from our population in the analysis indicated that presentation ctni concentrations 0.04 g/l were also predictive of AMI recurrence at 1 year (Table 4). Discussion Our findings indicate that the newly defined AMI group under the ESC/ACC definition (i.e., the NP group: negative according to WHO criteria but positive according ESC/ACC criteria) have risks for adverse events that appear, for the most part, to be intermediate between the risks observed in the group negative for both ESC/ACC and WHO criteria (i.e., the NN group) and the group positive for both ESC/ACC and WHO criteria (i.e., the PP group) (Tables 2 and 3). Each category exhibited least a trend toward a significant difference, and for most outcomes, statistical significance was achieved. These data Days since presentation Table 2. Hazard ratios for the AMI positive groups (i.e., NP, PP groups and intermediate, high ctni groups). Outcome AMI classification Hazard ratio relative to NN (95% CI) Marker concentration ( g/l) Hazard ratio relative to ctni <0.04 g/l (95% CI) 30 Death NP 1.64 (0.45, 5.97) (1.02, 19.22) PP 2.12 (0.53, 8.49) (0.88, 15.74) AMI recurrence NP (2.22, 167) No events in ref. category PP (12.33, 703) 0.10 Death/AMI NP 4.21 (1.54, 11.52) (2.16, 34.21) PP (6.32, 35.73) (7.28, 80.39) 120 Death NP 4.86 (2.02, 11.69) (1.97, 16.99) PP 4.15 (1.50, 11.46) (2.42, 18.80) AMI recurrence NP 7.64 (1.88, 31.06) (1.31, 126) PP (13.64, 152) (11.98, 652) Death/AMI NP 6.55 (3.00, 14.33) (3.42, 28.18) PP (7.90, 34.00) (9.42, 61.42) 360 Death NP 2.89 (1.49, 5.60) (1.37, 6.43) PP 2.51 (1.13, 5.60) (1.84, 7.60) AMI recurrence NP 4.71 (1.72, 12.89) (0.55, 10.06) PP (9.30, 50.65) (8.13, 54.30) Death/AMI NP 3.33 (1.87, 5.96) (1.65, 6.70) PP 8.19 (4.77, 14.05) (4.97, 15.35)

6 Clinical Chemistry 52, No. 11, Days since presentation Table 3. Hazard ratios compared with either the NP group or the intermediate ctni group ( g/l). Outcome AMI classification Hazard ratio relative to NP (95% CI) Marker g/l Hazard ratio relative to ctni g/l (95% CI) 30 Death NN 0.61 (0.17, 2.22) (0.05, 0.98) PP 1.29 (0.28, 5.89) (0.24, 2.96) AMI recurrence NN 0.05 (0.01, 0.45) 0.04 No events PP 4.82 (1.76, 13.21) (1.59, 30.16) Death/AMI NN 0.24 (0.09, 0.65) (0.03, 0.46) PP 3.57 (1.61, 7.94) (1.20, 6.59) 120 Death NN 0.21 (0.09, 0.50) (0.06, 0.51) PP 0.85 (0.34, 2.14) (0.52, 2.60) AMI recurrence NN 0.13 (0.03, 0.53) (0.01, 0.76) PP 5.98 (2.41, 14.86) (2.05, 23.00) Death/AMI NN 0.15 (0.07, 0.33) (0.04, 0.29) PP 2.50 (1.38, 4.55) (1.29, 4.67) 360 Death NN 0.35 (0.18, 0.67) (0.16, 0.73) PP 0.87 (0.39, 1.95) (0.63, 2.54) AMI recurrence NN 0.21 (0.08, 0.58) (0.10, 1.82) PP 4.61 (2.12, 10.03) (2.71, 29.45) Death/AMI NN 0.30 (0.17, 0.54) (0.15, 0.61) PP 2.46 (1.42, 4.25) (1.48, 4.69) are in contrast to previous findings that suggested that the risks of the group positive by ESC/ACC criteria only are greater than or equivalent to the risks of the group positive by both WHO and ESC/ACC criteria (4, 5). Specifically, these data vary from the report of Salomaa et al. (5), who reported that the all-cause mortality rate in patients meeting only the new AMI definition was worse than in patients meeting both criteria. Their population consisted of suspected MI events included in the Finnish myocardial infarction register study, FINAMI, during Troponin concentrations were used for diagnosis, but troponin did not uniformly determine management. Their patients who met only the ESC/ACC criteria were older and had diabetes more often, but most did not receive urgent revascularization. Troponin values were not available in our study population, and aggressive revascularization was not used widely at that time except for patients with STEMIs. Thus, the outcomes in our study are less biased by the use of modern therapies. Our NP group represents a previously undiagnosed subpopulation at equal risk for all-cause mortality but with a slightly lower AMI recurrence rate than the WHO-defined AMI population (Table 3), despite the fact our AMI recurrence rate in 1997 was based on the less sensitive WHO AMI criteria. We expect that this effect would have been greater had the ESC/ACC definition been used in 1997 for Table 4. Logistic regression model for prediction of AMI recurrence or death within 1 year of presentation based on either presentation or peak ctni concentration. A. AMI recurrence within 1 year Reference g/l Effect g/l OR, 95% CI C-Statistic Presentation value (3.17, 30.31) (1.91, 25.81) (6.71, 38.47) Peak value (0.64, 12.79) (11.00, 81.73) B. Death within 1 year Reference g/l Effect g/l OR, 95% CI C-Statistic Presentation value (0.89, 6.62) (1.61, 12.85) (0.44, 3.55) Peak value (1.61, 8.75) (2.00, 9.35)

7 2034 Kavsak et al.: Health Outcomes by AMI Definitions AMI classification (8). These data are consistent with the increased frequency of recurrent infarction and mortality over time in patients with non-stemi, which are enriched in our population (16). Patients with STEMI have a greater initial hazard but thereafter tend to have a more stable course (16, 17). In most series, by 6 months after presentation mortality rates for non-stemi and STEMI patients are similar (17). In addition, in keeping with the literature on acute coronary syndromes, our data confirm that a higher peak ctni is associated with a greater risk for death/ami. Our observation of a concentration-dependent effect was in an unselected population of all presenting patients and did not exclude STEMI patients. Studies by Kontos et al. (18) and Sabatine et al. (19) had similar findings, although their studies excluded STEMI patients who met criteria for fibrinolytic treatment, thereby lessening the overall severity of outcomes in their high-ctni group, which was analogous to our PP group or our high-peak ctni group. We did not exclude these patients, and therefore our PP and high-peak ctni groups might have had more severe outcomes. Our data are different from those of some acute coronary syndrome therapy trials, which have tended to select higher-risk patients and cannot be extrapolated easily to societal screening. For example, Morrow et al. (20) used the same AccuTnI assay in a population in which 64% of the patients had ctni values at or above the 99th percentile. Our data set comes from a nonselected cohort of emergency department patients in whom the frequency of ctni elevations at or above the 99th percentile was 37%, which is closer to the AMI prevalence in this population (8, 21). Thus, the present study provides new information from a population more representative of patients who present to emergency departments. In a troponin-naïve population, diagnoses made according to only the more sensitive ESC/ACC criteria identify patients with intermediate risks (lower than the risks in patients positive by WHO criteria), a finding that restores the expected relationship of biomarker concentrations, severity of disease, and prognosis (22, 23). Our study has several limitations. We are unable to discern the cause of death in our study population; therefore, we cannot compare the rates for cardiovascular deaths in our derived cohorts. Also, our small study population and consequent scarcity of events (i.e., AMI and/or death) not only led to wide confidence intervals around the hazard ratio estimates but also prevented secondary analyses, such as those based on sex. Our study demonstrates a difference in the frequency of adverse events during follow-up, related to the magnitude of the peak concentration of ctni and AMI definition used. However, the absolute adverse event frequencies in our study should not be extrapolated to the present. Currently, on the basis of the ESC/ACC definition, most of the patients with increased troponin values, including those in the group with only troponin increases, would be identified as high-risk patients (8) and would receive early invasive interventions and more potent antiplatelet/ antithrombin therapy, which in the vast majority of trials has been found to lower rates of adverse events (6). In conclusion, these data, from an unselected troponinnaïve population, before the identification and treatment of patients with non-stemi for aggressive intervention, validates the use of the ESC/ACC criteria for evaluating AMI in the population at large. This work was supported by the Research Trust Small Grants program of the Ontario Association of Medical Laboratories. The AccuTnI reagent was contributed for the study by an unrestricted grant from Beckman-Coulter Inc. References 1. Alpert JS, Thygesen K, Antman E, Bassand JP. Myocardial infarction redefined a consensus document of the Joint European Society of Cardiology/American College of Cardiology Committee for the Redefinition of Myocardial Infarction. J Am Coll Cardiol 2000;36: Luepker RV, Apple FS, Christenson RH, Crow RS, Fortmann SP, Goff D, et al. AHA Scientific Statement. Case definitions for acute coronary heart disease in epidemiology and clinical research studies. Circulation 2003;108: Lin JC, Apple FS, Murakami MM, Luepker RV. Rates of positive cardiac troponin I and creatine kinase MB mass among patients hospitalized for suspected acute coronary syndromes. Clin Chem 2004;50: Trevelyan J, Needham EW, Smith SC, Mattu RK. Impact of the recommendations for the redefinition of myocardial infarction on diagnosis and prognosis in an unselected United Kingdom cohort with suspected cardiac chest pain. Am J Cardiol 2004;93: Salomaa V, Koukkunen H, Ketonen M, Immonen-Raiha P, Karja- Koskenkari P, Mustonen J, et al. A new definition for myocardial infarction: what difference does it make? Eur Heart J 2005;26: Babuin L, Jaffe AS. Troponin: the biomarker of choice for the detection of cardiac injury. Can Med Assoc J 2005;173: Apple FS, Quist HE, Doyle PJ, Otto AP, Murakami MM. Plasma 99th percentile reference limits for cardiac troponin and creatine kinase MB mass for use with European Society of Cardiology/ American College of Cardiology Consensus Recommendations. Clin Chem 2003;49: Kavsak PA, MacRae AR, Lustig V, Bhargava R, Vandersluis R, Palomaki GE, et al. The impact of the ESC/ACC redefinition of myocardial infarction and new sensitive troponin assays on the frequency of acute myocardial infarction. Am Heart J 2006;152: MacRae AR, Kavsak PA, Lustig V, Bhargava R, Vandersluis R, Palomaki GE, et al. Assessing the Requirement for the 6-hour interval between specimens in the American Heart Association classification of Myocardial Infarction in Epidemiology and Clinical Research Studies. Clin Chem 2006;52: Venge P, Lindahl B, Wallentin L. New generation cardiac troponin I assay for the access immunoassay system. Clin Chem 2001; 47:

8 Clinical Chemistry 52, No. 11, Apple FS, Murakami MM, Pearce LA, Herzog CA. Multi-biomarker risk stratification of N-terminal pro-b-type natriuretic peptide, high-sensitivity C-reactive protein, and cardiac troponin T and I in end-stage renal disease for all-cause death. Clin Chem 2004;50: Panteghini M, Pagani F, Yeo KT, Apple FS, Christenson RH, Dati F, et al. Evaluation of imprecision for cardiac troponin assays at low-range concentrations. Clin Chem 2004;50: Wu AH, Apple FS, Gibler WB, Jesse RL, Warshaw MM, Valdes R. National Academy of Clinical Biochemistry Standards of Laboratory Practice: recommendations for the use of cardiac markers in coronary artery diseases. Clin Chem 1999;45: Apple FS, Parvin CA, Buechler KF, Christenson RH, Wu AH, Jaffe AS. Validation of the 99th percentile cutoff independent of assay imprecision (CV) for cardiac troponin monitoring for ruling out myocardial infarction. Clin Chem 2005;51: Austin PC, Daly PA, Tu JV. A multicenter study of the coding accuracy of hospital discharge administrative data for patients admitted to cardiac care units in Ontario. Am Heart J 2002;144: Furman MI, Dauerman HL, Goldberg RJ, Yarzebski J, Lessard D, Gore JM. Twenty-two year (1975 to 1997) Trends in the incidence, in-hospital and long-term case fatality rates from initial Q-wave and non-q-wave myocardial infarction: a multi-hospital, communitywide perspective. J Am Coll Cardiol 2001;37: Armstrong PW, Fu Y, Chang WC, Topol EJ, Granger CB, Betriu A, et al. Acute coronary syndromes in the GUSTO-IIb trial: prognostic insights and impact of recurrent ischemia. Circulation 1998;98: Kontos MS, Shah R, Fritz LM, Anderson FP, Tatum JL, Ornato JP, et al. Implication of different cardiac troponin I levels for clinical outcomes and prognosis of acute chest pain patients. J Am Coll Cardiol 2004;43: Sabatine MS, Morrow DA, McCabe CH, Antman EM, Gibson CM, Cannon CP. Combination of quantitative ST deviation and troponin elevation provides independent prognostic and therapeutic information in unstable angina and non-st-elevation myocardial infarction. Am Heart J 2006;151: Morrow DA, Rifai N, Sabatine MS, Ayanian S, Murphy SA, de Lemos JA, et al. Evaluation of the AccuTnI cardiac troponin I assay for risk assessment in acute coronary syndromes. Clin Chem 2003;49: Kavsak P, Lustig V, MacRae AR. Clinical evaluation of the Beckman Coulter AccuTnI assay for the early detection of acute myocardial infarction [Abstract]. Clin Chem 2004;50:A Geltman EM, Ehsani AA, Campbell MK, Schechtman K, Roberts R, Sobel BE. The influence of location and extent of myocardial infarction on long-term ventricular dysrhythmia and mortality. Circulation 1979;60: Alexander JH, Sparapani RA, Mahaffey KW, Deckers JW, Newby LK, Ohman EM, et al. Association between minor elevations of creatine kinase-mb level and mortality in patients with acute coronary syndromes without ST-segment elevation. JAMA 2000; 283: