Causes of Elevated Cardiac Troponins in the Emergency Department and Their Associated Mortality

Transcription

1 ORIGINAL CONTRIBUTION Causes of Elevated Cardiac Troponins in the Emergency Department and Their Associated Mortality Stephen Meigher, Henry C. Thode, PhD, W. Frank Peacock, MD, Jay L. Bock, MD, Louis Gruberg, MD, and Adam J. Singer, MD Abstract Objective: Cardiac troponins (ctn) are structural components of myocardial cells and are expressed almost exclusively in the heart. Elevated ctn levels indicate myocardial cell damage/death but not reflect the underlying etiology. The third universal definition of myocardial infarction (MI) differentiates MI into various types. Type 1 (T1MI) is due to plaque rupture with thrombus, while type 2 (T2MI) is a result of a supply:demand mismatch. Non-MI ctn elevations are also common. We determined the causes of elevated ctn in a tertiary care emergency department (ED) and the associated in-hospital mortality. Methods: We performed a structured, retrospective review of all consecutive adult ED patients with elevated troponin I (defined as > 99th percentile of the normal population, as run on the ADVIA Centaur platform; Siemens USA) during 1 year. Causes of elevated ctn were classified based on the third universal definitions. Comparisons between groups were performed using chi-square and Mann-Whitney U-tests. Results: Of 96,612 ED patients presenting from May 2012 to April 2013, a total of 13,502 (14%) had ctn measured, of which 1,310 (9.7%) were elevated. Of these, 340 (26.5%, 95% confidence interval [CI], 24.2% to 29.0%) were T1MI, 452 (35.2%, 95% CI = 32.7% to 37.9%) T2MI, 458 (35.7%, 95% CI = 33.1% to 38.4%) multifactorial, and 33 (2.5%, 95% CI = 1.8% to 3.5%) due to nonischemic injury. Non-T1MI patients were slightly older, more likely female, and had higher blood urea nitrogen and creatinine. Comorbidities were more common in non-t1mi while cardiac risk factors were more common in T1MI. Non-T1MI patients were less likely to have diagnostic ECGs and had lower initial and subsequent ctn levels. In-hospital mortality rates were similarly high for T1MI and non-t1mi (11% [95% CI = 8% to 15%] vs. 10% [95% CI = 8% to 12%], p = 0.48). Conclusions: Of all ED patients with elevated ctn, ~75% have a non-t1mi. The mortality of patients with non-t1mi is similar to the mortality in patients with T1MI. ACADEMIC EMERGENCY MEDICINE 2016;23: by the Society for Academic Emergency Medicine Of all biomarkers of myocardial injury currently used in clinical practice, cardiac troponins (ctni and ctnt) are the most sensitive and specific. As a result, they have replaced all other markers and multimarker strategies to become a critical part of the diagnosis of myocardial infarction (MI). 1 With the advent of improved assay platforms able to detect very low troponin concentrations, the sensitivity of ctn for detecting myocardial cell necrosis has greatly increased. 2 In contrast, the specificity for diagnosing classical, spontaneous, coronary artery disease (CAD)- related MI has decreased. This may result in diagnostic confusion if only the troponin concentration is considered. In recognition that many etiologies cause myocardial cell injury and elevated ctn, the definition of MI and its subtypes have been standardized. The universal definition of MI was introduced in and most recently revised in This classifies MI into five groups based on the clinical scenario and underlying pathophysiology. Type 1 MI (T1MI) represents a spontaneous MI due to CAD with plaque rupture/thrombosis, whereas Type 2 MI (T2MI) results from a mismatch in myocardial oxygen supply and demand. Type 3 MIs are diagnosed in the setting of sudden death From the Department of Emergency Medicine (SM, HCT, AJS), the Department of Clinical Pathology (JLB), and the Department of Cardiology (LG), Stony Brook University, Stony Brook, NY; and the Department of Emergency Medicine, Baylor College of Medicine (WFP), Houston, TX. Received March 2, 2016; revision received May 19, 2016; accepted June 12, The authors have no relevant financial information or potential conflicts to disclose. Supervising Editor: Brian C. Hiestand, MD, MPH. Address for correspondence and reprints: Adam J Singer, MD; adam.singer@stonybrook.edu by the Society for Academic Emergency Medicine ISSN doi: /acem PII ISSN

2 1268 Meigher et al. CAUSES OF ELEVATED TROPONIN and are generally not diagnosed without an autopsy. Types 4 and 5 represent iatrogenic MIs occurring during percutaneous coronary interventions (PCI) or coronary artery bypass graft (CABG) surgery, respectively. In addition to MI due to ischemia, other causes of elevated ctn include nonischemic myocardial cell necrosis (e.g., myocarditis, toxins) and multifactorial or indeterminate etiologies. 4,5 Early identification of a T1MI is important since an early and prompt intervention has been shown to be a major determinant of outcomes. Thus in the setting of an elevated ctn, emergency physicians must be able to distinguish between a T1M1 and other causes for its increase. We are unaware of any emergency department (ED)-based studies that have systematically studied the underlying etiologies in patients with elevated ctn, especially in the era of increasing ctn assay platform sensitivities. The purpose of this study was to determine the etiology of ctn elevations in consecutive ED patients with levels above the upper limit of normal. We also compared characteristics and outcomes in ED patients with T1MI and non-t1mi. We hypothesized that non-t1mi would be more common than T1MI in ED patients, but that mortality would not differ between these patient groups. METHODS Study Design We conducted a structured, retrospective chart review of all patients that presented to the ED and had elevated ctn levels. Standard methods for structured chart reviews were followed 6,7 except that the chart abstractors were not blinded to the study goals. All study variables were defined a priori and documented in a coding guide for the chart abstractors. A standard data collection instrument that had been tested was used. All data abstractors received specific training in the definitions and use of the data collection instrument and jointly reviewed the first 20 cases. Monitoring of data abstraction was performed periodically by the principal investigator and on a subset of 20 randomly selected charts that were independently reviewed by two abstractors; interobserver agreement on the primary outcome (type of MI) and ECG was 100% for both. The study was approved by the institutional review board with waiver of informed consent due to the retrospective nature and minimal risk of the study. Subjects Consecutive adult patients presenting to the ED between May 2012 to April 2013, having an i-stat Point-of-Care (POC) Troponin I (Abbott Point of Care) obtained, and with a result exceeding the 99th percentile upper reference limit for a normal reference population (0.04 ng/ml) were included in this study. Patients determined to be experiencing type 4 or type 5 MI were excluded from further data analysis. Setting The study was conducted at a regional Level 1 trauma center and Academic Medical Center serving a mostly affluent suburban population. We also have a full-service cardiac catheterization laboratory and are an American College of Cardiology Society of Cardiovascular Patient Care accredited chest pain center. Measurements and Outcomes Patient information obtained in the ED as well as all information obtained during hospital admission were gathered using a standardized data collection form following recommendations for studies including patients with acute coronary syndromes (ACS). 8 Pertinent information included chief complaint, history of present illness, vital signs, past medical history, comorbidities, cardiac risk factors, laboratory results, and ECG findings. The data for all variables were based on all current and prior electronic medical records, including those from any outpatient visits. Since all study patients had elevated ctn, it was very rare that the presence or absence of these data was not clearly stated in the current or prior medical records. In the rare cases of missing data regarding specific cardiac risk factors, these were classified as absent. With regard to smoking, patients were classified as either currently or not currently smoking. The main outcomes were type of MI and inpatient mortality. Determination of Cause of Elevated ctn The universal definition of MI was used to categorize the subtypes of AMI as types 1 and 2. 4 We also determined the presence of non-mi causes of myocardial cell necrosis, when obvious ischemia or imbalance between oxygen supply and demand was not apparent following the methods described by Saaby et al. 5 At the beginning of the study all adjudicators (two emergency physicians and a cardiologist) met to review and discuss the definitions set forth in the third universal definitions and jointly reviewed the first 20 cases. Thereafter, a single adjudicator reviewed all charts and interobserver agreement for the primary outcome and ECG classification was determined on a randomly selected sample of 20 patients. In addition, periodic meetings were held with all of the adjudicators to review any cases that were not obvious to any of the adjudicators and in these cases adjudication was by consensus of all three members. On a subset of 20 randomly selected subjects, agreement on MI classification of cause of elevated ctn and ECG was 100% for both. ECG Interpretation. ECG findings were categorized into six groups: normal, nonspecific, abnormal but nondiagnostic, ischemic (old, unchanged), ischemic (new), and consistent with AMI following universally accepted definitions. 9 On the subset of 20 randomly selected patients, agreement was also 100% between the two independent observers. Troponin Assay At the time of the study, testing of ctn in ED patients was performed by laboratory technicians in a stat lab (located in a dedicated area of the central laboratory) using the i-stat ctni test (Abbott Point of Care). With this assay the 99th percentile upper reference limit of the normal reference population is 0.04 ng/ml with a

3 ACADEMIC EMERGENCY MEDICINE November 2016, Vol. 23, No coefficient variation of 10% at that level. If at any point the POC ctni was elevated, the sample was retested in the central laboratory using the ADVIA Centaur TnI- Ultra assay (Siemens USA), which has a similar 99th percentile upper reference limit of 0.04 ng/ml and a coefficient variation of 10% at that level. Patients with elevated ctn then had serial measurements performed 3 and 6 hours later whenever possible. Data Analyses Categorical data were summarized as counts and percentages of the frequency of occurrence and compared between groups with chi-square and Fisher s exact tests. Continuous data were summarized as means and standard deviations (SD) or medians and interquartile ranges for normally distributed and nonnormally distributed data, respectively. Student t-test and Mann- Whitney U-test were used to compare normally distributed and skewed data, respectively. Comparisons were made in baseline characteristics and outcomes between T1M1, T2M1, multifactorial, or indeterminate causes, and non-mi myocardial cell necrosis. We also compared T1MI with all non-t1mi combined. RESULTS General Characteristics and Etiologies of Elevated ctn During the study period there were 96,612 patient visits to the ED. A POC ctn was ordered in 13,502 (14.0%) patients of whom 1,310 (9.7%, 95% confidence interval [CI], 9.2 to 10.2) were above the 99% percentile of the normal reference population. All patients with an elevated POC ctn also had an elevated central laboratory ctn assay. In all these patients, another central laboratory ctn assay was performed followed by serial ctn 3 and 6 hours later (whenever possible) using this latter assay. Overall, 27 patients were excluded as not being ED relevant. Of the 1,283 study patients with elevated ctn, the mean (SD) patient age was 71 (15) years and 36% were female; 340 (26.5%, 95% CI = 24.2 to 29.0) had a T1MI and 943 (73.5%, 95% CI = 71.0 to 75.8) had troponin elevations secondary to a non-t1mi. In the non-t1mi cohort, 452 were adjudicated to be due to T2MI (35.2%, 95% CI = 32.7 to 37.9), 33 to nonischemic myocardial cell necrosis (2.5%, 95% CI = 1.8 to 3.5), and 458 due to multifactorial or indeterminate causes of myocardial injury (35.7%, 95% CI = 33.1 to 38.4). A full breakdown of the causes of elevated ctn is presented in Table 1. Comparison of T1MI and Non-T1M1 Patients Chest pain was present in slightly more than half of patients with T1MI (59%) but in only 9% of those with T2MI, 27% of those with nonischemic injury, and 13% of patients with multifactorial or indeterminate causes of elevated ctn. Non-T1MI patients were slightly older, more likely to be female, and had higher blood urea nitrogen (BUN) and creatinine (Table 2). In general, comorbidities were more common in non-t1mi, while cardiac risk factors were more common in T1MI. However, the incidence of prior CAD, MI, and CABG was similar among the various types of MI (Table 2). Non- Table 1 Etiology of Elevated ctn Primary Etiology No. of Patients %(n = 1,283) Injury related to primary myocardial ischemia (T1MI) Injury related to supply/demand imbalance of myocardial ischemia (T2MI) Tachy-/bradyarrythmias Aortic dissection or severe 27 2 aortic valve disease Hypertrophic cardiomyopathy 1 <1 Cardiogenic, hypovolemic, or septic shock Severe respiratory failure Severe anemia 29 2 Coronary spasm 16 1 Severe hypertension 62 5 Coronary embolism or vasculitis 1 <1 Coronary endothelial dysfunction 0 0 without significant CAD Injury not related to myocardial 33 3 ischemia Cardiac contusion, surgery, 9 1 ablation, pacing, or defibrillator shocks Rhabdomyolysis with cardiac 12 1 involvement Myocarditis 11 1 Cardiotoxic agents 1 <1 Multifactorial or indeterminate myocardial injury Heart failure Stress (Takostubo) cardiomyopathy 10 1 Severe pulmonary embolism 28 2 or pulmonary hypertension Sepsis and critically ill patients 4 1 Renal failure Severe acute neurological 45 4 disease, e.g., stroke, SAH Infiltrative diseases, e.g., 2 <1 amyloidosis, sarcoidosis Strenuous exercise 0 0 CAD = coronary artery disease; ctn = cardiac troponin; SAH = subarachnoid hemorrhage; T1MI = type 1 myocardial infarction; T2MI = type 2 myocardial infarction. T1MI patients were less likely to have diagnostic ECGs and had lower initial and subsequent ctn levels than patients with T1MI (Table 3). Patients with non-t1mi had higher BUN, creatinine, and BNP than those with T1MI, reflective of the common occurrence of renal and heart failure in those with non-t1mi. A more detailed breakdown of baseline characteristics by cause of elevated ctn is presented in Table 4. More patients with T1MI were admitted to an intensive care unit (ICU) than non-t1mi (88.7% vs. 56.5%, p < 0.001) and ED deaths occurred more commonly in T1MI versus non-t1mi patients (2.4% vs. 0.3%, p < 0.001). In contrast, in-hospital mortality rates were similar between patients with T1MI and non-t1mi (11% vs. 10%, p = 0.48). However, when broken down into the various subtypes of non-t1mi, significant differences were noted between patients with an ischemic etiology and those with nonischemic or multifactorial etiologies. Mortality rates by etiology were T1MI 11%,

4 1270 Meigher et al. CAUSES OF ELEVATED TROPONIN Table 2 Comparison of Baseline Characteristics, ECG, and Labs Between T1MI and non-t1mi T1MI (n = 340) Non-T1MI (n = 943) p-value Age (y), mean (SD) 69 (14) 72 (16) % Male Cardiac risk factors, % (95% CI) Hypertension 66 (60 71) 65 (62 68) 0.90 Diabetes mellitus 37 (32 42) 30 (27 33) 0.03 Elevated cholesterol 53 (48 59) 43 (40 47) Family history of MI 6 (4 9) 3 (2 5) 0.04 Tobacco use 38 (33 43) 27 (25 30) <0.001 Comorbidities, % (95% CI) CHF 16 (13 21) 33 (30 37) <0.001 CAD 42 (37 48) 42 (39 45) 0.91 MI 15 (11 19) 13 (11 16) 0.50 PCI 21 (17 26) 16 (14 18) <0.001 CABG 21 (17 25) 21 (18 24) 0.91 Renal failure 16 (13 21) 28 (25 31) <0.001 ECG findings, % (95% CI) <0.001 Consistent with MI 48 (43 54) 5 (4 6) Ischemia not known to be old 18 (14 23) 15 (13 18) Abnormal but nondiagnostic 13 (13 17) 30 (27 33) Old ischemia 3 (1 5) 4 (3 6) Nonspecific 8 (5 11) 20 (18 23) Early repolarization 0 (0 1) 1 (0.5 2) Normal 10 (7 14) 25 (22 28) Laboratory findings BNP (pg/ml), median (IQR) 334 ( ) 533 ( ) Mean BUN (mg/dl), mean (SD) 29.0 (19.2) 35.0 (22.5) <0.001 Creatinine (mg/dl), mean (SD) 1.7 (1.9) 2.0 (2.1) 0.02 BNP = B-type natriuretic peptide; BUN = blood urea nitrogen; CABG = coronary artery bypass graft; CHF = congestive heart failure; IQR = interquartile range; MI = myocardial infarction; PCI = percutaneous coronary intervention. Table 3 Comparison of ctn Levels, ng/ml Type Median (IQR) N (%) Median (IQR) N (%) Median (IQR) N (%) Type 1 (n = 340) 050 ( ) 335 (98) 1.96 ( ) 315 (93) 3.04 ( ) 301 (89) Non-Type 1 (n = 943) 0.14 ( ) 928 (98) 0.15 ( ) 866 (92) 0.16 ( ) 796 (84) p <0.001 <0.001 <0.001 Type 1 (n = 340) 050 ( ) 335 (98) 1.96 ( ) 315 (93) 3.04 ( ) 301 (89) Type 2 (n = 452) 0.14 ( ) 441 (98) 0.17 ( ) 409 (90) 0.17 ( ) 373 (83) Non-ischemic injury (n = 33) 0.21 ( ) 33 (100) 0.25 ( ) 32 (97) 0.35 ( ) 30 (91) Multifactorial (n = 458) 0.13 ( ) 454 (99) 0.13 ( ) 425 (93) 0.14 ( ) 393 (86) p <0.001 <0.001 <0.001 ctn = cardiac troponin. 1st troponin 2nd troponin 3rd troponin T2MI 12%, nonischemic myocardial necrosis 6%, and multifactorial or indeterminate 7% (p = 0.04). In contrast, overall mortality rate for patients with a normal initial and serial ctn was 1.3% during the same time period. DISCUSSION Our results demonstrate that non-t1mi is approximately three times more common than T1MI in patients presenting to a large academic ED and with an elevated ctn. Patients with T1MI were more likely to have chest pain and traditional cardiovascular risk factors, while patients with non-t1mi were more likely to have comorbidities like renal or heart failure. Non-T1MI were also slightly older, more likely female, and less likely to have diagnostic ECGs. Initial and subsequent ctn levels were higher in patients with T1MI versus non-t1mi. Most importantly, mortality was equally high in patients with T1MI and non-t1mi. Our finding that most patients with elevated ctn do not have T1MI has important clinical implications since the emergent evaluation and management of patients with T1MI and non-t1mi differ significantly. Of equal importance is our finding that in-hospital mortality is similar in both groups. Thus, the presence of any myocardial cell necrosis (as indicated by an elevated ctn) must be taken seriously, regardless of etiology, and these patients require further evaluation and management.

5 ACADEMIC EMERGENCY MEDICINE November 2016, Vol. 23, No Table 4 Comparison of Baseline Characteristics, ECG, and Labs Between Various Causes of Elevated ctn T1MI T2M1 Nonischemic Multifactorial p-value Age (y), mean (SD) 69 (14) 72 (16) 61 (25) 73 (15) <0.001 % Male Cardiac risk factors, % (95% CI) Hypertension 66 (60 71) 64 (59 68) 55 (35 71) 67 (63 72) 0.37 Diabetes mellitus 37 (32 42) 27 (23 31) 21 (10 39) 35 (30 39) Elevated cholesterol 53 (48 59) 43 (38 47) 33 (19 52) 45 (40 49) Family history of MI 6 (4 9) 4 (2 6) 6 (1 22) 3 (1 5) 0.12 Tobacco use 38 (33 43) 28 (24 33) 24 (12 43) 27 (23 31) Comorbidities, % (95% CI) CHF 16 (13 21) 26 (22 31) 9 (2 25) 42 (38 47) <0.001 CAD 42 (37 48) 37 (32 41) 18 (8 36) 49 (44 54) <0.001 MI 15 (11 19) 13 (10 16) 9 (2 25) 14 (11 18) 0.70 PCI 21 (17 26) 14 (11 18) 3 (.2 18) 19 (15 22) CABG 21 (17 25) 18 (15 22) 12 (4 29) 24 (20 28) 0.08 Renal failure 16 (13 21) 19 (16 23) 9 (2 25) 38 (33 42) <0.001 ECG findings, % (95% CI) Suggestive of MI 48 (43 54) 3 (2 5) 9 (2 25) 6 (4 9) <0.001 Ischemia not known to be old 18 (14 23) 18 (15 22) 21 (10 39) 12 (10 16) Abnormal but nondiagnostic 13 (13 17) 29 (25 34) 21 (10 39) 30 (26 35) Old ischemia 3 (1 5) 4 (2 6) 3 (0.2 18) 5 (3 7) Nonspecific 8 (5 11) 20 (17 24) 15 (6 33) 21 (17 25) Early repolarization 0 (0 1) 2 (0.7 3) 3 (0.2 18) 0.4 (0.1 2) Normal 10 (7 14) 24 (20 29) 27 (14 46) 25 (21 29) Laboratory findings BNP (pg/ml), median (IQR) 334 (83 1,159) 426 ( ) 344 (66 957) 692 (302 1,444) <0.001 BUN (mg/dl), mean (SD) 29.0 (19.2) 32.3 (21.1) 23.7 (16.1) 38.5 (23.6) <0.001 Creatinine (mg/dl), mean (SD) 1.7 (1.9) 1.6 (1.4) 1.1 (0.4) 2.4 (2.6) <0.001 BNP = B-type natriuretic peptide; BUN = blood urea nitrogen; CABG = coronary artery bypass graft; CHF = congestive heart failure; ctn = cardiac troponin; IQR = interquartile range; MI = myocardial infarction; PCI = percutaneous coronary intervention. With the advent of electronic medical records and the implementation of computerized physician order entry (often with lists of prepopulated laboratory test orders, such as ctn) more and more patients are found to have an elevated ctn, even when MI is not a diagnostic consideration. A major dilemma for the emergency physician is deciding how to proceed when ctn is elevated, especially when a T1MI is very unlikely. Indeed, our results suggest that an elevated ctn is much more likely to be due to non-t1mi or other nonischemic causes. Therefore, the management of these patients will depend on the performance of a risk assessment using a risk stratification tool, such as the TIMI risk score or the GRACE risk score. In some of these patients less emergent cardiac catheterization may be necessary. Thus a major unanswered question is whether ordering of ctn in patients deemed unlikely to have ACS (e.g., pneumonia or sepsis) is warranted or helpful. Regardless of the underlying etiology of cardiac cell damage, multiple studies, including ours, have confirmed that the presence of an elevated troponin is associated with worse short-term clinical outcomes, including death. For example, others note ctn is elevated in 43% of noncardiac ICU patients 10 who then suffer a 250% mortality increase (odds ratio = 2.5, 95% CI = 1.9 to 3.4) compared to patients without elevated ctn. 11 Up to 10% of acute stroke patients have elevated ctn, which is then associated with increased mortality. 12 Troponin elevations not due to coronary flow limiting etiologies are also seen in up to half of all patients with severe sepsis or septic shock and are associated with significant increases in mortality. 13,14 In addition to helping predict prognosis, according to the National Academy of Clinical Biochemistry Guidelines, ctn may also help determine the extent of left ventricular dysfunction and need for inotropic support in septic patients. 15 However, the therapeutic implications of elevated troponins in critically ill patients are not clear. The universal definitions of MI have not been widely adopted among U.S. emergency physicians. 16 The reluctance in using these definitions may be in part due to the ambiguous criteria and overlapping findings between the five subtypes, especially types 1 and 2. Both types may include seriously ill patients with chest pain, elevated troponin levels, and ischemic ECG changes. Yet, their respective etiologies differ and generally require different treatment strategies. While T1MI more classically presents with ST-segment elevation myocardial infarction and identifiable CAD, T2MI may also, at times, present with ST-elevations and CAD unrelated to primary cardiac injury. 17,18 In addition, not all causes of T2MI are easily explained by a mismatch in cardiac oxygen supply and demand or multifactorial injury, and a universal consensus on the diagnostic criteria and features of T2MI has yet to be reached. 19 Our results demonstrating that T1MI is responsible for a minority of ctn elevations are in agreement with a recent single-center, prospective study in Denmark that included 1,961 patients with elevated ctn (>0.03 lg/l) over a 1-year period. 5 Of these patients, only 553 (28.2%) were thought to be secondary to a MI and 1,408 (71.8%) were thought to be secondary to nonischemic

6 1272 Meigher et al. CAUSES OF ELEVATED TROPONIN myocardial necrosis. Among those with true ischemic MI, 397 were type 1, 144 were type 2, and the remainder were due to types 4 and 5. Thus, only one in five patients with elevated ctn had T1MI and would likely require emergency coronary artery intervention. Unsurprisingly, almost half of all T1MI patients were diagnosed with ECG ST-segment elevations consistent with acute MI. While this is not a novel finding, the 3% prevalence of ST-elevation on the ECG in T2MI patients is noteworthy and is supported by recent research, 17 further adding to the diagnostic confusion. Most non- T1MI were caused by multifactorial disease states with heart failure and renal failure accounting for the majority of cases. Regardless of etiology, mortality within both cohorts was high and previous research suggests that T2MI carries an even higher short-term mortality burden. 20 It is well established that chest pain is often absent in T2MI patients. 18 In non-acs patients presenting to the ED with moderate to severe physiologic dysfunction or any T2MI-related disease states, ctn should be tested to aid in diagnosis and, more importantly, to help with risk stratification. Conversely, selected low-risk patients with serial troponin values below the 99th percentile may be safely discharged from the ED LIMITATIONS Our study has several limitations. First, we conducted a retrospective analysis of medical charts introducing all of the biases inherent to this type of study design. However, we mitigated the potential for selection bias by reviewing all consecutive ED patients with elevated ctn. Second, our study is limited to a single, large academic ED that might not be representative of other settings. Third, while adjudication of the cause of elevated ctn was structured and interobserver agreement was high, it may not be accurate or agreed upon by other reviewers. Fourth, while very rare, when data regarding the presence or absence of cardiac risk factors and comorbidities were absent, we made the assumption that the condition was absent. This may have introduced some bias. However, all present and prior medical records were reviewed for completeness, and since all patients had elevated ctn it was very unlikely that these elements were missing. Another major limitation of this study is the potential for classification bias and the extent of this bias cannot be ascertained. For example, it is very likely that patients without documented chest pain would be more likely to be classified as having a T2MI. Therefore, it comes as no surprise that patients with T2MI were commonly found to not have complained of chest pain. Thus some of the differences between the groups may have been overestimated. Finally, the chart abstractors were not blinded to study objectives. CONCLUSION In conclusion, we have shown that the majority of ED patients with elevated cardiac troponins do not have type 1 myocardial infarction and that most are caused by type 2 myocardial infarction or nonischemic myocardial injury or are multifactorial or of indeterminate etiology. Importantly, mortality was equally high among patients with and without type 1 myocardial infarction, reemphasizing the prognostic importance of cardiac troponins. References 1. Amsterdam EA, Wenger NK, Brindis RG, et al AHA/ACC Guideline for the Management of Patients with Non-ST-Elevation Acute Coronary Syndromes: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 2014;64: e Segraves JM, Frishman WH. Highly sensitive cardiac troponin assays: a comprehensive review of their clinical utility. Cardiol Rev 2015;23: Thygesen K, Alpert JS, White HD. Universal definition of myocardial infarction. Eur Heart J 2007;28: Thygesen K, Alpert JS, Jaffe AS, et al. Third universal definition of myocardial infarction. J Am Coll Cardiol 2012;60: Saaby L, Poulsen TS, Hosbond S, et al. Classification of myocardial infarction: frequency and features of type 2 myocardial infarction. Am J Med 2013;126: Gilbert EH, Lowenstein SR, Koziol-McLain J, Barta DC, Steiner J. Chart reviews in emergency medicine research: where are the methods? Ann Emerg Med 1996;27: Kaji AH, Schriger D, Green S. Looking through the retroscope: reducing bias in emergency medicine chart review studies. Ann Emerg Med 2014;64: Hollander JE, Blomkalns AL, Brogan GX, et al. Standardized reporting guidelines for studies evaluating risk stratification of emergency department patients with potential acute coronary syndromes. Ann Emerg Med 2004;44: Cullen L, Than M, Brown AF, et al. Comprehensive standardized data definitions for acute coronary syndrome research in emergency departments in Australasia. Emerg Med Australas 2010;22: Lim W, Qushmaq I, Devereaux PJ, et al. Elevated cardiac troponin measurements in critically ill patients. Arch Intern Med 166: Lim W, Cook DJ, Griffith LE, Crowther MA, Devereaue PJ. Elevated cardiac troponin levels in critically ill patients: prevalence, incidence, and outcomes. Am J Crit Care 2006;15: Jensen JK, Kristensen SR, Bak S, Atar D, Hoilund- Carlsen PF, Mickey H. Frequency and significance of troponin T elevation in acute ischemic stroke. Am J Cardiol 2007;99: Maeder M, Fehr T, Rickli H, Amman P. Sepsisassociated myocardial dysfunction: diagnostic and prognostic impact of cardiac troponins and natriuretic peptides. Chest 2006;129: Mehta NJ, Kahn IA, Gupta V, Jani K, Gowda RM, Smith PR. Cardiac troponin I predicts myocardial dysfunction and adverse outcomes in septic shock. Int J Cardiol 2004;95:13 17.

7 ACADEMIC EMERGENCY MEDICINE November 2016, Vol. 23, No Wu I, Yu F, Chou J, Lin T, Chen H, Lee C. Predictive risk factors for upper gastrointestinal bleeding with simultaneous myocardial injury. Kaohsiung J Med Sci 2007;23: Pierpoint GL, McFalls EO. Interpreting troponin elevations: do we need multiple diagnoses? Eur Heart J 2009;30: Sandoval Y, Smith SW, Thordsen SE, Apple FS. Supply/demand type 2 myocardial infarction: should we be paying more attention? J Am Coll Cardiol 2014;63: Alpert JS, Thygesen KA, White HD, Jaffe AS. Diagnostic and therapeutic implications of type 2 myocardial infarction: review and commentary. Am J Med 2014;127: Smilowitz NR, Naoulou, Sedlis SP. Diagnosis and management of type II myocardial infarction: increased demand for a limited supply of evidence. Curr Atheroscler Rep 2015;17: Shah AS, McAllister DA, Mills R, et al. Sensitive troponin assay and the classification of myocardial infarction. Am J Med 2015;128: e Than M, Aldous S, Lord SJ, et al. A 2-hour diagnostic protocol for possible cardiac chest pain in the emergency department: a randomized clinical trial. JAMA Intern Med 2014;174: Cullen L, Mueller C, Parsonage WA, et al. Validation of high-sensitivity troponin I in a 2-hour diagnostic strategy to assess 30-day outcomes in emergency department patients with possible acute coronary syndrome. J Am Coll Cardiol 2013;62: Than M, Cullen L, Reid CM, et al. A 2-h diagnostic protocol to assess patients with chest pain symptoms in the Asia-Pacific region (ASPECT): a prospective observational validation study. Lancet 2011;377: The National Library of Medicine is offering a free online TOXNET class this fall TOXNET is a web-based system of databases covering hazardous chemicals, environmental health, toxic releases, chemical nomenclature, poisoning, risk assessment and regulations, and occupational safety and health. The independent modules cover TOXLINE, ChemIDplus, TRI, TOXMAP, Hazardous Substances Data Bank, IRIS, Has-Map, LactMed, WISER, CHEMM, REMM, LiverTox and more. You ll learn about the resources through videos, guided tutorials, and discovery exercises.