- Email: [email protected]
Death or revascularization among nonadmitted ED patients with low-positive vs negative troponin T results
Death or revascularization among non-admitted emergency department patients with low-positive versus negative troponin T results Christopher W. Baugh MD, MBA, Joshua M. Kosowsky MD, David A. Morrow MD, Jonathan D. Sonis MD, Allen G. Gold, Clare E. Ronan, Daniel J. Pallin MD, MPH PII: DOI: Reference:
S0735-6757(14)00332-5 doi: 10.1016/j.ajem.2014.05.013 YAJEM 54296
To appear in:
American Journal of Emergency Medicine
Received date: Revised date: Accepted date:
12 February 2014 10 April 2014 12 May 2014
Please cite this article as: Baugh Christopher W., Kosowsky Joshua M., Morrow David A., Sonis Jonathan D., Gold Allen G., Ronan Clare E., Pallin Daniel J., Death or revascularization among non-admitted emergency department patients with low-positive versus negative troponin T results, American Journal of Emergency Medicine (2014), doi: 10.1016/j.ajem.2014.05.013
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT TITLE:
Death or revascularization among non-admitted emergency department patients with low-positive versus negative troponin T results
RI P
Authors:
T
Running head: Low-positive troponin outcomes
Christopher W. Baugh, MD, MBA; Joshua M. Kosowsky, MD; David A. Morrow, MD;
SC
Jonathan D. Sonis, MD; Allen G. Gold; Clare E. Ronan; Daniel J. Pallin, MD, MPH Institutional Affiliations:
NU
Department of Emergency Medicine, Brigham and Women’s Hospital, Boston, MA (CB, JK, JS, CR,
MA
DP); Department of Cardiology, Brigham and Women's Hospital, Boston, MA (DM); Harvard Medical School (CB, JK, DM, DP); Harvard-Affiliated Emergency Medicine Residency (JS), and New York
ED
Institute of Technology College of Osteopathic Medicine, New York, NY (AG).
PT
Corresponding author (reprints not available):
Key Words:
CE
Christopher W. Baugh MD, MBA, 75 Francis St., Boston, MA 02115, [email protected] emergency department, troponin, biomarkers, myocardial infarction, outcomes,
AC
observation
Funding and support: This research was made possible by support from the Eleanor and Miles Shore Fellowship Program for Scholars in Medicine, administrated by the Department of Emergency Medicine at Brigham and Women's Hospital. Abbreviations: ED – Emergency Department; EM – Emergency Medicine; CI – Confidence Interval; RD – Risk Difference; MI – Myocardial Infarction; ECG – Electrocardiogram; ICU – Intensive Care Unit; SSDI – Social Security Death Index
ACCEPTED MANUSCRIPT Abstract Study Objective: Compare outcomes among emergency department (ED) patients with low-positive
T
(0.01-0.02 ng/mL) versus negative troponin T.
RI P
Methods: Retrospective cohort study of non-admitted ED patients with troponin testing at a tertiary-care hospital. Trained research assistants used a structured tool to review charts from all non-admitted ED
SC
patients with troponin testing, 12/1/2009─11/30/2010. Outcomes of death and coronary revascularization
NU
were assessed at 30 days and 6 months via medical record review, Social Security Death Index searches, and patient contact.
MA
Results: There were 57,596 ED visits; with 33,388 (58%) discharged immediately, 6,410 (11%) assigned to the observation unit, and 17,798 (31%) admitted or other. Troponin was measured in 2,684 (6.7%) of
ED
the non-admitted cases. Troponin was negative in 2,523 (94.0%), low-positive in 78 (2.9%), and positive (≥0.03 ng/mL) in 83 (3.1%). Of troponin-negative cases, 0.8% (95%CI 0.4%-1.1%) died or were
PT
revascularized by 30 days, versus 2.8% (95%CI 0.0%-6.7%) of low-positive cases (risk difference [RD]
CE
2.0%, 95%CI -1.8%-5.9%). At 6 months the rates were 1.7% (95%CI 1.1%-2.2%) and 12.9% (95%CI 5.0%-20.7%) (RD 11%, 95%CI 3.3%-19.1%,). Death alone at 30 days occurred in 0.4% (95%CI 0.1%-
AC
0.6%) versus 1.3% (95%CI 0.0%-3.8%) (RD 0.9%, 95%CI -1.6%-3.4%). Death at 6 months occurred in 1.2% (95%CI 0.8%-1.6%) versus 11.7% (95%CI 4.5%-18.9%) (RD 10%, 95%CI 3.3%-17.7%). Conclusion: Among patients not initially admitted, rates of death and coronary revascularization differed insignificantly at 30 days but significantly at 6 months. Detailed inspection of our results reveals that the bulk of the added risk at 6 months was due to non-cardiac mortality.
ACCEPTED MANUSCRIPT 1. INTRODUCTION Cardiac troponin I and troponin T are the most specific biochemical markers for the diagnosis of
T
myocardial injury [1]. Initial studies and consensus recommendations led to the acceptance of the 99th
RI P
percentile among apparently normal people as the cutoff for “normal” [2]. This of course implies that 1% of patients without acute myocardial injury have a result above this threshold.
SC
Initial enthusiasm about the specificity of troponin assays has given way to the realization that troponin can be elevated in conditions other than acute myocardial injury, including chronic renal failure,
NU
chronic heart disease, and skeletal muscle disease [1,3]. In such cases, a low-positive troponin result may
MA
be an important prognostic indicator for long-term serious outcomes due to causes other than coronary occlusion, but the relevance of this result during a "rule-out myocardial infarction" evaluation is poorly understood [4]. Additional data regarding the short and intermediate-term prognosis among such patients
ED
are needed to support evidence-based decision-making.
PT
Chest pain accounts for 5-7% of all ED visits in the United States, with >6,000,000 visits annually [5,6]. History, physical exam, ECG and cardiac biomarkers are not sufficiently sensitive to rule
CE
out acute coronary occlusion during the initial ED evaluation [5]. Prior studies suggest that between 2-8% of ED patients with acute myocardial infarction and unstable angina were discharged home and missed
AC
myocardial infarction accounted for about 20% of malpractice claims against emergency physicians [7,8]. ED observation units were created to improve detection of acute ischemia by providing the time needed for ongoing evaluation, without having to commit a large number of low-risk patients to costly full inpatient admission [9,10]. In this study we seek to quantify the rate of serious health outcomes among patients with lowpositive initial troponin values who are not initially admitted to an inpatient setting. We compare outcomes in the “low positive” group to outcomes in the “negative” group.
ACCEPTED MANUSCRIPT 2. METHODS 2.1 Study Design and Setting
T
We performed a retrospective cohort study at an urban, academic, adult-only, tertiary-care ED.
RI P
Board-certified or board-eligible emergency physicians supervise residents and physician assistants, the usual first-line providers. All troponin testing was ordered by a physician or physician assistant; no
SC
standing orders permitted troponin ordering directly by a nurse. Our cohort was seen in the ED from 12/1/2009-11/30/2010. This date range was chosen because 12/1/2009 was the first day a new troponin
NU
assay was used. Prior to this date, our institution used the Siemens TnI-Ultra, which reported all results <0.03 ng/mL as “less than assay.” From this date forward, our institution used a 4th-generation
MA
conventional troponin T assay (Roche Diagnostics), which reports all results ≥0.01 ng/mL, and designates all non-zero results as abnormal. Our clinicians were confused because there were now patients with
ED
troponin values of 0.01-0.02 ng/mL, when previously all such values had been reported as “less than
PT
assay.” We refer to results in this range (0.01-0.02 ng/mL) as “low-positive.” Providers were notified of the change in an email, but no other formal education or direction was offered. Prior to the change in
CE
assay, our ED’s usual practice was to admit all patients with a newly positive troponin to the hospital. After the introduction of the new assay, there was confusion about how to manage clinically low-risk
AC
patients with troponin results below the reporting limit of the prior assay. At all times, as per the standard of care in emergency medicine, all patients with worrisome ECG findings or with moderate or high gestalt clinical probability for acute coronary occlusion, were admitted to an inpatient service, regardless of troponin results.
2.2 Selection of Participants Our goal in this investigation was to examine outcomes among patients with low-positive initial troponin results who were not admitted to an inpatient setting. We used our computerized ED order entry system to generate a list of all patients with an order for troponin testing during the study period. We then excluded all whose initial disposition was inpatient admission. We also excluded trauma patients who had
ACCEPTED MANUSCRIPT troponin testing to rule out myocardial contusion. Finally, we excluded patients discharged against
T
medical advice and those who died in the ED.
RI P
2.3 Data Collection and Processing
Trained research assistants collected the study data from our electronic medical record, which is
SC
shared by all of the ambulatory clinics affiliated with the ED and by seven area hospitals, using a standardized data collection form. A physician investigator reviewed all cases with an outcome (defined
NU
below). We sent all English and Spanish-speaking patients with a domestic mailing address who had
MA
incomplete 30-day and/or 6-month follow up data a letter asking them about outcomes and requesting permission to obtain outside hospital records. We made follow up calls to those patients who did not respond to the letter after 6 weeks. Patients who lived abroad or did not speak English or Spanish were
ED
analyzed as lost to follow up (see below).
PT
While our main outcome was death or revascularization, we knew that we would be unable to contact some patients. Thus, we used all-cause mortality as an important secondary outcome measure. We
CE
ascertained mortality by searching the Social Security Death Index (SSDI) [11]. We waited at least one year to check the index to ensure adequate time had passed so that deaths occurring during the follow up
AC
window had been reported. We accessed a current version of the SSDI Master File via our institution’s Research Patient Data Registry, a centralized data warehouse that linked patients in our study to deaths listed in the Master File within the study timeframe. We categorized revascularization events as "urgent" or "non-urgent" according to the operative report.
2.4 Outcome Measures Our main outcome measure was the risk difference for the patient-centered composite outcome of death or coronary revascularization at 30 days, among negative versus low-positive troponin cases. We also evaluated the composite outcome at 6 months; and death alone at 30 days and 6 months. We also report the number of subsequent inpatient hospitalizations with a primary discharge diagnosis of
ACCEPTED MANUSCRIPT myocardial infarction or other serious cardiovascular cause (pulmonary embolus, aortic dissection and stroke) by 30 days and 6 months, following prior investigations [12,13]. We did not consider further
RI P
T
diagnostic testing to represent an outcome.
2.5 Data Analysis
SC
We calculated risk differences and their 95% confidence intervals (95%CI) for each outcome. Our prospective sample size calculations indicated that 1,980 visits would be needed to achieve 80%
NU
power to detect an absolute difference of 0.5% versus 2% for the main outcome, with alpha=0.05.
MA
We constructed multivariable models to identify and control for confounding of the relationship of low-positive troponin to outcomes. A priori, we considered the following covariates to be of potential relevance: gender, age, race, prior history of myocardial infarction, coronary revascularization, high
ED
cholesterol, hypertension, congestive heart failure, end-stage renal disease, diabetes mellitus, tobacco use,
PT
and recent cocaine use. An example of how confounding could occur would be if diabetics were more likely to be revascularized for various reasons, and were also more likely to have a mild troponin T
CE
elevation because they are more likely to have renal insufficiency. We used the method of Bursac to identify confounders [14]. We found no evidence of confounding and the results of this analysis are not
AC
shown (but are available in the Technical Appendix). We were also concerned about the possibility that low-positive troponin results might have different implications in some subgroups. Therefore, we sought evidence of effect modification. We constructed logistic regression models with each of the outcomes mentioned above as dependent variables, and with the following predictors: troponin, age, end-stage renal disease, and presence or absence of history of congestive heart failure, as well as 2-way interaction terms of each of these with troponin. We identified no interactions (i.e. no evidence of different effects at different levels of these covariates), and the results of this analysis are not shown (but are available in the Technical Appendix). We performed all analyses using SAS 9.1 (Cary, NC). This study was approved by our Institutional Review Board.
ACCEPTED MANUSCRIPT
3. RESULTS
T
3.1 Characteristics of Study Subjects
RI P
Figure 1 is a participant flow chart, and Table 1 describes the cohort. During the study period, there were 57,596 ED visits, with 16,096 (27.9%) admitted, 33,388 (58.0%) discharged home, and 6,410
SC
(11.1%) assigned to the ED observation unit. The remainder had alternative dispositions such as death, left without being seen, inter-facility transfer, or other. Of visits with an initial disposition of home or ED
NU
observation, troponin was measured in 2,684 (6.7%). Results were negative in 2,523 (94.0%), low-
MA
positive in 78 (2.9%), and positive (≥0.03 ng/mL) in 83 (3.1%). At 30 days, data on coronary revascularization and hospitalization were available for 86% and 90% of those with initially negative and low-positive troponin, respectively. At 6 months, data on
ED
coronary revascularization and hospitalization were available for 83% and 90% of those with initially
PT
negative and low-positive troponin, respectively. The SSDI could capture all deaths except in unusual circumstances, such as foreign nationals. We had no reason to suspect differential misclassification
AC
3.2 Main Results
CE
according to follow-up data availability.
The main outcome occurred in 17, or 0.8% (95%CI 0.4%-1.1%) of troponin-negative cases, versus 2 or 2.8% (95%CI 0.0%-6.7%) of low-positive cases (risk difference 2.0%, 95%CI -1.8%-5.9%) (Table 2). At 6 months, the composite outcome occurred in 35, or 1.7% (95%CI 1.1%-2.2%) and 9, or 12.9% (95%CI 5.0%-20.7%) (risk difference 11%, 95%CI 3.3%-19.1%,). Death at 30 days occurred in 9, or 0.4% (95%CI 0.1%-0.6%) versus 1, or 1.3% (95%CI 0.0%-3.8%) (risk difference 0.9%, 95%CI -1.6%3.4%). Death at 6 months occurred in 30, or 1.2% (95%CI 0.8%-1.6%) versus 9, or 11.7% (95%CI 4.5%18.9%) (risk difference 10%, 95%CI 3.3%-17.7%). At 30 days there were nine revascularization events. Of these, three were urgent (two from the negative troponin group and one from the low-positive troponin group). Between 30 days and 6 months
ACCEPTED MANUSCRIPT there were seven additional revascularizations, four of which were urgent (all four from the negative troponin group).
T
We assessed the rate of hospitalization with a primary discharge diagnosis of myocardial
RI P
infarction, pulmonary embolus, aortic dissection, or stroke by 30 days and 6 months in all three groups. In the negative troponin group there were 9, or 0.4% (95%CI 0.1%-0.7%) at 30 days and 13, or 0.6%
SC
(95%CI 0.3%-1.0%) at 6 months. In the low positive troponin group there were 0, or 0.0% at 30 days and 1, or 1.5% (95%CI 0.0%-4.5%) at 6 months. The risk difference at 6 months was 0.9% (95%CI 2.1%-
MA
NU
3.9%).
4. LIMITATIONS
Loss to follow-up is a fundamental challenge in any follow-up study of a rare event, and this
ED
study is no exception. Even with a prospective design, some loss to follow-up is inevitable. For our
PT
primary outcome, we had definitive data on whether revascularization was performed on only 86% of our troponin negative and 90% of our low-positive participants. The concern is that we might have missed
CE
coronary revascularization outcomes among the low-positive initial troponin cases, and if we were more
AC
likely to miss outcomes in the low-positive group than in the negative group, our results would be skewed. Our cohort had 78 cases with low-positive initial troponin results. We had definitive follow-up data for 70 of them. In this group of 70, there was one case of revascularization (1.4%). Given this, had we missed one additional case of revascularization among those eight visits and none among the 349 visits with incomplete follow up in the troponin negative group, our primary outcome would be different. However, we were fortunate to be able to rely on the SSDI for mortality. In retrospect, it might have been wise to frame mortality as our main outcome measure, which is a framework previously used to follow patients after an ED visit for conditions such as syncope [15]. It also bears mentioning that the SSDI is not 100% perfect. For example, foreigners visiting the United States and entering our cohort would be missed if they died overseas [11].
ACCEPTED MANUSCRIPT Because this was a single-center study, generalizability is a concern. If clinicians at other centers were less cautious in their initial decision to admit or not admit, or to test or not test, measured outcomes
T
using this design could be worse, in troponin-negative or low-positive cases, or both. Our cohort was
RI P
derived from an institution that had recently switched from the Siemens TnI-Ultra to the Roche 4thgeneration troponin T assay, and this could have led our clinicians to lower or raise their threshold for
SC
admission of low-risk patients. The width of our main outcome's 95% confidence interval is discussed in the Discussion section proper.
NU
We focused only on patients who were clinically low risk (as evidenced by initial disposition) and
MA
had a negative or low-positive initial troponin result. It was beyond the scope of this investigation to evaluate changes in troponin results over time. Many of our patients had only a single troponin assay. Very few had changes in the results (data not shown), reflecting the low-risk status of the cohort; this
PT
ED
study was not powered to evaluate that group.
5. DISCUSSION
CE
To evaluate the significance of low-positive initial troponin results among ED patients, we compared outcomes among patients with low-positive results to outcomes among patients with negative
AC
results. We found no significant difference in mortality or coronary revascularization at 30 days. This suggests that patients with low-positive troponin values using the Roche 4th-generation troponin T assay without high-risk clinical indicators of acute coronary occlusion may be candidates for outpatient management. This conclusion has two caveats. First, our measured risk difference for revascularization or death, 2.0%, and its 95%CI of −1.8% to 5.8%, leave open the possibility that a much larger study could find significantly increased 30-day risk in the low-positive group. However, as an outcome, mortality has greater validity, because catheterization is subject to diagnostic workup bias. Patients with detectible troponin values are more likely to have additional testing, and medical testing drives medical procedures. It is impossible to know, in retrospect, whether a cardiologist’s decision to dilate or stent a stenotic vessel
ACCEPTED MANUSCRIPT is a good indicator of that patient’s “need” for that procedure. Measurement of fractional flow reserve lends some objectivity to the decision to intervene, but this technique was not in routine use at our
T
institution during the study period [16]. Of note, the vulnerability of our revascularization outcome to
RI P
workup bias renders our analysis conservative, because its effect would be unidirectional, away from the null. Thus, the fact that the risk difference for mortality was lower, and its 95%CI narrower, are
SC
somewhat reassuring. Nevertheless, the upper bound of our mortality outcome's 95% confidence interval was 3.4%, emphasizing the need for further study in larger samples.
NU
The second caveat is that the clear difference in outcomes by 6 months suggests that patients with
MA
an initial low-positive troponin values are indeed at higher risk than those with negative results. But it appears that they are at risk for death due to causes other than acute coronary occlusion. Inspection of our results reveals that the bulk of the added risk at 6 months was due to non-cardiac mortality (Table 3).
ED
Comparing outcomes at 6 months versus 30 days also suggests that mortality, but not coronary
PT
revascularization, increased over time. This suggests that the patients with initially low-positive troponin results were, in fact, at higher risk, but not because of their coronary arteries, and not in the short term.
CE
(Differential loss to follow-up in the low-positive versus negative groups with respect to revascularization would undermine these conclusions, but we have no reason to suspect that the patients seen in our ED
AC
with low-positive troponin results would be more likely to be catheterized outside our system than those with negative troponin results.) These results are important because the standard of care for low-risk chest pain patients continues to evolve. In the past, all “rule-out MI” patients were admitted to an inpatient setting. Now, management of low-risk chest pain patients in ED observation units is commonplace. At one time, protocols required observation stays of up to twenty-four hours, and stress testing prior to discharge was nearly universal. Over the years, the time frame for repeat troponin testing has contracted to 6 hours or even less, and the need for provocative testing has been questioned [17-20]. Taken together, these trends and observations suggest that in the future, patients deemed by clinician gestalt or via validated risk stratification tools to be at low risk for acute coronary syndrome might safely be discharged after negative or low-positive
ACCEPTED MANUSCRIPT troponin results, with outpatient follow-up. Moreover, cycling of a “delta troponin” in as little as one to two hours is showing promise as a new strategy to rapidly exclude the diagnosis of acute myocardial
T
ischemia [18,21]. Pressure to treat patients as outpatients or "observation" patients, rather than
RI P
"inpatients," continues to mount. Understanding the meaning of low-positive troponin results is crucial to appropriate management. These patients may be well served by avoiding short heart-focused admissions,
SC
in favor of more intensive outpatient care to address other causes of mortality. Any positive troponin should alert clinicians to a chance of mortality greater than 12% in 6 months.
NU
The fact that troponin indicates non-coronary mortality is of historical interest. When troponin
MA
assays were first commercialized, clinicians celebrated their specificity for myocardial infarction. Initial confidence soon gave way to more circumspection, as it became clear that troponin could be elevated in many circumstances other than acute myocardial infarction. Then, non-coronary troponin elevations were
ED
found to predict mortality in the intensive-care setting, where short-term mortality data are easiest to
PT
collect [3,22,23]. Other studies examining the general population have also shown a connection between troponin elevation and mortality over longer periods of time [24]. Troponin is increasingly viewed as a
CE
non-specific marker of mortality [25]. The present study adds to these prior findings by providing data on mortality after measurement of troponin in the ED setting.
AC
A further complication is found in the fact that, in recent years, the proliferation of troponin assays has introduced variability in cutoffs [26]. As institutions adopt different assays, providers may be confused about how to interpret values close to the “normal” range [27]. For example, the Siemens TnIUltra assay has a cutoff for “normal” of 0.03 ng/mL, and reports lower results as “less than assay.” In contrast, the Roche 4th generation troponin T assay reports results as low as 0.01 ng/mL, and all non-zero values are considered “abnormal.” Clinicians switching between these assays could be confused about the interpretation of results between 0 and 0.03 ng/mL [27]. This happens when an institution changes its assay or when data from one institution are interpreted by clinicians at another. Given the imminent introduction of a new class of “highly-sensitive” troponin assays to the United States, this issue will take on even more importance in the near future [28]. Thus, two factors confound clinicians’ daily
ACCEPTED MANUSCRIPT interpretation of positive, but very low troponin results: uncertainty as to their biological significance, and keeping track of the subtle differences in test characteristics from one assay to the next.
T
Normalization, as was done with the International Normalized Ratio (INR) might be one solution.
RI P
For example, instead of reporting the absolute value of the troponin concentration, laboratories could report the cumulative distribution function (i.e. the percentile) for that result according to the distribution
SC
of results observed by the manufacturer during elucidation of test characteristics leading to FDA approval. However, this goal is more complicated than it seems on its surface, and efforts to derive
NU
methods for normalization have been underway for some time [29].
MA
Clinical judgment is the cornerstone of risk stratification in patients who are evaluated for acute coronary occlusion but do not have ECG changes or other clear clinical indicators of its presence. Risk stratification tools, such as the Thrombolysis In Myocardial Infarction (TIMI) score, the Global Registry
ED
of Acute Coronary Events (GRACE) score, the Goldman rule, and the Acute Cardiac Ischemia-Time
PT
Insensitive Predictive Instrument (ACI-TIPI) hold promise for the differentiation of high-risk from lowrisk patients [30-33]. However, none of these tools has proven sufficiently accurate to identify patients
CE
who are safe for discharge from the ED without serial cardiac markers or provocative testing. Combinations of several risk scores may improve predictive performance, but may also be so
AC
cumbersome as to be difficult to implement in daily practice [34]. Perhaps more importantly, the patients included in the TIMI, GRACE, and ACI-TIPI studies were selected by experts in the acute coronary syndrome; and we can rest assured that practicing clinicians will continue to obtain troponin measurements in patients more diverse than those included in such studies. Thus, the significance of lowpositive results will remain important.
6. CONCLUSION A low-positive troponin result in a patient deemed to be at low risk for acute myocardial infarction is a harbinger of other causes of mortality in the succeeding few months, but may not indicate risk of serious outcomes due to acute coronary occlusion. Providers should consider alternative potential
ACCEPTED MANUSCRIPT causes of near-team mortality in these patients. Additional studies are warranted, in other ED populations
AC
CE
PT
ED
MA
NU
SC
RI P
T
and with larger samples.
ACCEPTED MANUSCRIPT REFERENCES [1] Panteghini M. Acute coronary syndrome: biochemical strategies in the troponin era. Chest 2002;122:1428-
RI P
T
35.
[2] Thygesen K. Alpert JS. White HD. Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of
SC
Myocardial Infarction. Jaffe AS. Apple FS. Galvani M. Katus HA. Newby LK. Ravkilde J. Chaitman B.
NU
Clemmensen PM. Dellborg M. Hod H. Porela P. Underwood R. Bax JJ. Beller GA. Bonow R. Van der Wall
MA
EE. Bassand JP. Wijns W. Ferguson TB. Steg PG. Uretsky BF. Williams DO. Armstrong PW. Antman EM.
ED
Fox KA. Hamm CW. Ohman EM. Simoons ML. Poole-Wilson PA. Gurfinkel EP. Lopez-Sendon JL. Pais P.
PT
Mendis S. Zhu JR. Wallentin LC. Fernandez-Aviles F. Fox KM. Parkhomenko AN. Priori SG.
CE
Tendera M.
Voipio-Pulkki LM. Vahanian A. Camm AJ. De Caterina R. Dean V. Dickstein K. Filippatos G.
AC
Funck-
Brentano C. Hellemans I. Kristensen SD. McGregor K. Sechtem U. Silber S. Tendera M. Widimsky P.
Zamorano JL. Morais J. Brener S. Harrington R. Morrow D. Lim M. Martinez-Rios MA. Steinhubl S. Levine GN. Gibler WB. Goff D. Tubaro M. Dudek D. Al-Attar N. Universal definition of myocardial infarction. Circulation 2007;116:2634-53. [3] Babuin L, Vasile VC, Rio Perez JA, et al. Elevated cardiac troponin is an independent risk factor for short-
ACCEPTED MANUSCRIPT and long-term mortality in medical intensive care unit patients. Crit Care Med 2008;36:759-65. [4] Daniels LB, Laughlin GA, Clopton P, Maisel AS, Barrett-Connor E. Minimally elevated cardiac
T
troponin T
the
SC
Rancho Bernardo Study. J Am Coll Cardiol 2008;52:450-9.
RI P
and elevated N-terminal pro-B-type natriuretic peptide predict mortality in older adults: results from
[5] Hess EP, Brison RJ, Perry JJ, et al. Development of a clinical prediction rule for 30-day cardiac
NU
events in
MA
emergency department patients with chest pain and possible acute coronary syndrome. Ann Emerg Med 2012;59:115,25.e1.
ED
[6] Chandra A, Rudraiah L, Zalenski RJ. Stress testing for risk stratification of patients with low to
PT
moderate
probability of acute cardiac ischemia. Emerg Med Clin North Am 2001;19:87-103.
associated
CE
[7] Schull MJ, Vermeulen MJ, Stukel TA. The risk of missed diagnosis of acute myocardial infarction
AC
with emergency department volume. Ann Emerg Med 2006;48:647-55. [8] Pope JH, Aufderheide TP, Ruthazer R, et al. Missed diagnoses of acute cardiac ischemia in the emergency department. N Engl J Med 2000;342:1163-70. [9] Graff LG, Dallara J, Ross MA, et al. Impact on the care of the emergency department chest pain patient from the chest pain evaluation registry (CHEPER) study. Am J Cardiol 1997;80:563-8. [10] Jaffe AS, Babuin L, Apple FS. Biomarkers in Acute Cardiac Disease: The Present and the Future. J Am Coll Cardiol 2006;48:1-11.
ACCEPTED MANUSCRIPT [11] Schisterman EF, Whitcomb BW. Use of the Social Security Administration Death Master File for ascertainment of mortality status Population Health Metrics 2004;2:1-6.
T
[12] Poon M, Cortegiano M, Abramowicz AJ, et al. Associations between routine coronary computed
RI P
tomographic angiography and reduced unnecessary hospital admissions, length of stay, recidivism rates, and
SC
invasive coronary angiography in the emergency department triage of chest pain. J Am Coll Cardiol \ 2013;62:543-52.
NU
[13] Lipinski MJ, McVey CM, Berger JS, Kramer CM, Salerno M. Prognostic value of stress cardiac
MA
magnetic
resonance imaging in patients with known or suspected coronary artery disease: a systematic review and
ED
meta-analysis. J Am Coll Cardiol 2013;62:826-38.
PT
[14] Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression.
CE
Source Code for Biology and Medicine 2008;3:1-8.
after a
AC
[15] Derose SF, Gabayan GZ, Chiu VY, Sun BC. Patterns and preexisting risk factors of 30-day mortality
primary discharge diagnosis of syncope or near syncope. Acad Emerg Med 2012;19:488-96. [16] De Bruyne B, Pijls NH, Kalesan B, et al. Fractional flow reserve-guided PCI versus medical therapy in stable coronary disease. N Engl J Med 2012;367:991-1001. [17] Prasad V, Cheung M, Cifu A. Chest pain in the emergency department: the case against our current practice of routine noninvasive testing. Arch Intern Med 2012;172:1506-9. [18] Than M, Cullen L, Aldous S, et al. 2-Hour accelerated diagnostic protocol to assess patients with chest pain
ACCEPTED MANUSCRIPT symptoms using contemporary troponins as the only biomarker: the ADAPT trial. J Am Coll Cardiol 2012;59:2091-8.
T
[19] Redberg RF. Coronary CT angiography for acute chest pain. N Engl J Med 2012;367:375-6.
RI P
[20] Scheuermeyer FX, Innes G, Grafstein E, et al. Safety and efficiency of a chest pain diagnostic algorithm with
SC
selective outpatient stress testing for emergency department patients with potential ischemic chest pain. Ann
NU
Emerg Med 2012;59:256,264.e3.
MA
[21] Reichlin T, Schindler C, Drexler B, et al. One-hour rule-out and rule-in of acute myocardial infarction using
high-sensitivity cardiac troponin T. Arch Intern Med 2012;172:1211-8.
ED
[22] Ammann P, Maggiorini M, Bertel O, et al. Troponin as a risk factor for mortality in critically ill
PT
patients
without acute coronary syndromes. J Am Coll Cardiol 2003;41:2004-9.
CE
[23] Wu TT, Yuan A, Chen CY, et al. Cardiac troponin I levels are a risk factor for mortality and multiple organ
outcome
AC
failure in noncardiac critically ill patients and have an additive effect to the APACHE II score in
prediction. Shock 2004;22:95-101. [24] de Lemos JA, Drazner MH, Omland T, et al. Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA 2010;304:2503-12. [25] Mahajan VS, Jarolim P. How to interpret elevated cardiac troponin levels. Circulation 2011;124:2350-4. [26] Apple FS, Collinson PO, IFCC Task Force on Clinical Applications of Cardiac,Biomarkers. Analytical
ACCEPTED MANUSCRIPT characteristics of high-sensitivity cardiac troponin assays. Clin Chem 2012;58:54-61. [27] Apple FS. A new season for cardiac troponin assays: it's time to keep a scorecard. Clin Chem
T
2009;55:1303-
RI P
6.
[28] Storrow AB, Lardaro TA, Alexander PT, Apple FS. How Low Can We Go? The High-Sensitivity
SC
Cardiac Troponin Debate. Ann Emerg Med 2013;.
NU
[29] Christenson RH, Bunk DM, Schimmel H, Tate JR, IFCC Working Group on Standardization of
MA
Troponin,I.
Point: Put simply, standardization of cardiac troponin I is complicated. Clin Chem 2012;58:165-8. [30] Goldman L, Weinberg M, Weisberg M, et al. A computer-derived protocol to aid in the diagnosis of
ED
emergency room patients with acute chest pain. N Engl J Med 1982;307:588-96.
PT
[31] Antman EM, Cohen M, Bernink PJ, et al. The TIMI risk score for unstable angina/non-ST elevation MI: A
CE
method for prognostication and therapeutic decision making. JAMA 2000;284:835-42.
predictive
AC
[32] Selker HP, Beshansky JR, Griffith JL, et al. Use of the acute cardiac ischemia time-insensitive
instrument (ACI-TIPI) to assist with triage of patients with chest pain or other symptoms suggestive of acute cardiac ischemia. A multicenter, controlled clinical trial. Ann Intern Med 1998;129:845-55. [33] Eagle KA, Lim MJ, Dabbous OH, et al. A validated prediction model for all forms of acute coronary syndrome: estimating the risk of 6-month postdischarge death in an international registry. JAMA 2004;291:2727-33. [34] Marcoon S, Chang AM, Lee B, Salhi R, Hollander JE. HEART score to further risk stratify patients with low TIMI scores. Crit pathw cardiol 2013;12:1-5.
ACCEPTED MANUSCRIPT Figure 1: Patient Selection Strategy and Primary Outcome
RI P
T
57,596 adult ED visits between 12/1/2009 and 11/30/2010
CE
9/2,523 deaths in troponin negative group
AC
1/78 deaths in troponin 0.01-0.02 ng/mL group
PT
2,601 visits with 30day mortality follow up completed via chart review and SSDI
NU
83 visits with initial troponin ≥0.03mng/ mL
ED
2,684 visits with initial troponin negative or 0.010.02 ng/mL
245 visits excluded as trauma patients
MA
2,929 visits with a disposition of home or ED observation and a troponin resulted in the ED
SC
39,798 visits with an initial disposition of home or ED observation
2,244 visits (86%) with 30-day revascularization follow up completed via chart review, letter and/or call
8/2,174 revascularizations in troponin negative group
17/2,174 deaths or revascularizations in troponin negative group
2/70 deaths or revascularizations in troponin 0.01-0.02 ng/mL group
1/70 revascularizations in troponin 0.01-0.02 ng/mL group
ACCEPTED MANUSCRIPT Table 1: Patient Characteristics of Study Group Initial Troponin T Low Positive Negative
RI P
T
Patient Data # of visits
2,523
78
56.0
68.0
41.3%
56.4%
47.6% 28.9% 20.2% 1.4% 1.7%
55.1% 30.8% 11.5% 1.3% 0%
47.9%
66.7%
3.1%
3.8%
15.2%
17.9%
28.2%
50.0%
Prior history of myocardial infarction
17.7%
48.7%
Prior coronary revascularization
14.5%
39.7%
History of end-stage renal disease
0.7%
7.7%
History of congestive heart failure
11.6%
46.2%
ED length of stay (no observation used)*
4.5 hours (3.56.0 hours)
5.5 hours (3.57.5 hours)
ED length of stay (subsequent observation used)*
3.1 hours (2.34.0 hours)
3.2 hours (2.14.5 hours)
Observation length of stay, if used*
8.0 hours (5.713.7 hours)
7.8 hours (5.313.1 hours)
Age
SC
Male sex Race
MA
NU
White African American Hispanic Asian Other History of high cholesterol History of recent cocaine use
ED
History of tobacco use
AC
CE
PT
History of diabetes mellitus
*
0.01-0.02 ng/mL
Reported as the median (interquartile range)
ACCEPTED MANUSCRIPT Table 2: Main Results
Initial Troponin T Low-positive
Risk Difference
(
(0.01-0.02 ng/mL)
(95% confidence interval)
2.8% (95%CI 0.0%-6.7%)
2.0% (95%CI 1.8%-5.9%)
RI P
T
Negative
30 day death or revascularization
0.8% (95%CI 0.4%-1.1%)
Other Outcomes*
SC
Primary Outcome*
30 day revascularization
0.3% (95%CI 0.1%-0.5%)
1.4% (95%CI 0.0%-3.8%)
1.1% (95%CI 1.7%-3.9%)
6 month death or revascularization
1.7% (95%CI 1.1%-2.2%)
12.9% (95%CI 5.0%-20.7%)
11% (95%CI 3.3%-19%)
6 month death
1.2% (95%CI 0.8%-1.6%)
11.7% (95%CI 4.5%-18.9%)
10% (95%CI 3.3%-18%)
6 month revascularization
0.3% (95%CI 0.1%-0.6%)
0%
0.3%
30 day hospitalization†
0.4% (95%CI 0.1%-0.7%)
0%
0.4%
0.6% (95%CI 0.3%-1.0%)
1.5% (95%CI 0.0%-4.5%)
0.9% (95%CI 2.1%-3.9%)
ED
MA
NU
0.4% (95%CI 0.1%-0.6%)
CE
The denominators for the death and revascularization calculations are different because we found outcomes via different strategies; as a result, the composite outcome is not simply a sum of the individual outcomes Hospitalizations with primary discharge diagnosis of myocardial infarction, pulmonary embolism, stoke, or aortic dissection
AC
†
PT
6 month hospitalization†
*
1.3% (95%CI 0.0%-3.8%)
0.9% (95%CI 1.6%-3.4%)
30 day death
ACCEPTED MANUSCRIPT Table 3: Cause of Death of Study Participants* via Search of Death Records Number of Patients
Cancer, not specified
2
Lung cancer
2
Stroke
1
Pancreatic cancer
1
Multiple myeloma
1
Congestive heart failure
1
Drug overdose
1
Chronic obstructive pulmonary disease
1
Lung cancer
5
Congestive heart failure
4
ED
Cardiopulmonary arrest Brain cancer
PT
Ovarian carcinoma Cancer, not specified
Colon cancer
AC
Subdural hematoma
CE
Chronic obstructive pulmonary disease Gastric cancer
4 2 2 1 1 1 1 1
Pulmonary embolism
1
Common bile duct carcinoma
1
Renal cell carcinoma
1
Squamous cell carcinoma of the trachea
1
Ischemic colitis
1
Drug overdose
1
Sepsis
1
Alzheimer’s disease
1
Unknown/not available
13
Total
43
*
RI P SC
NU 10
MA
Total
30-day to 6 month Cause of Death
T
30-day Cause of Death
Troponin negative and low-positive patients only
S0735-6757(14)00332-5 doi: 10.1016/j.ajem.2014.05.013 YAJEM 54296
To appear in:
American Journal of Emergency Medicine
Received date: Revised date: Accepted date:
12 February 2014 10 April 2014 12 May 2014
Please cite this article as: Baugh Christopher W., Kosowsky Joshua M., Morrow David A., Sonis Jonathan D., Gold Allen G., Ronan Clare E., Pallin Daniel J., Death or revascularization among non-admitted emergency department patients with low-positive versus negative troponin T results, American Journal of Emergency Medicine (2014), doi: 10.1016/j.ajem.2014.05.013
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT TITLE:
Death or revascularization among non-admitted emergency department patients with low-positive versus negative troponin T results
RI P
Authors:
T
Running head: Low-positive troponin outcomes
Christopher W. Baugh, MD, MBA; Joshua M. Kosowsky, MD; David A. Morrow, MD;
SC
Jonathan D. Sonis, MD; Allen G. Gold; Clare E. Ronan; Daniel J. Pallin, MD, MPH Institutional Affiliations:
NU
Department of Emergency Medicine, Brigham and Women’s Hospital, Boston, MA (CB, JK, JS, CR,
MA
DP); Department of Cardiology, Brigham and Women's Hospital, Boston, MA (DM); Harvard Medical School (CB, JK, DM, DP); Harvard-Affiliated Emergency Medicine Residency (JS), and New York
ED
Institute of Technology College of Osteopathic Medicine, New York, NY (AG).
PT
Corresponding author (reprints not available):
Key Words:
CE
Christopher W. Baugh MD, MBA, 75 Francis St., Boston, MA 02115, [email protected] emergency department, troponin, biomarkers, myocardial infarction, outcomes,
AC
observation
Funding and support: This research was made possible by support from the Eleanor and Miles Shore Fellowship Program for Scholars in Medicine, administrated by the Department of Emergency Medicine at Brigham and Women's Hospital. Abbreviations: ED – Emergency Department; EM – Emergency Medicine; CI – Confidence Interval; RD – Risk Difference; MI – Myocardial Infarction; ECG – Electrocardiogram; ICU – Intensive Care Unit; SSDI – Social Security Death Index
ACCEPTED MANUSCRIPT Abstract Study Objective: Compare outcomes among emergency department (ED) patients with low-positive
T
(0.01-0.02 ng/mL) versus negative troponin T.
RI P
Methods: Retrospective cohort study of non-admitted ED patients with troponin testing at a tertiary-care hospital. Trained research assistants used a structured tool to review charts from all non-admitted ED
SC
patients with troponin testing, 12/1/2009─11/30/2010. Outcomes of death and coronary revascularization
NU
were assessed at 30 days and 6 months via medical record review, Social Security Death Index searches, and patient contact.
MA
Results: There were 57,596 ED visits; with 33,388 (58%) discharged immediately, 6,410 (11%) assigned to the observation unit, and 17,798 (31%) admitted or other. Troponin was measured in 2,684 (6.7%) of
ED
the non-admitted cases. Troponin was negative in 2,523 (94.0%), low-positive in 78 (2.9%), and positive (≥0.03 ng/mL) in 83 (3.1%). Of troponin-negative cases, 0.8% (95%CI 0.4%-1.1%) died or were
PT
revascularized by 30 days, versus 2.8% (95%CI 0.0%-6.7%) of low-positive cases (risk difference [RD]
CE
2.0%, 95%CI -1.8%-5.9%). At 6 months the rates were 1.7% (95%CI 1.1%-2.2%) and 12.9% (95%CI 5.0%-20.7%) (RD 11%, 95%CI 3.3%-19.1%,). Death alone at 30 days occurred in 0.4% (95%CI 0.1%-
AC
0.6%) versus 1.3% (95%CI 0.0%-3.8%) (RD 0.9%, 95%CI -1.6%-3.4%). Death at 6 months occurred in 1.2% (95%CI 0.8%-1.6%) versus 11.7% (95%CI 4.5%-18.9%) (RD 10%, 95%CI 3.3%-17.7%). Conclusion: Among patients not initially admitted, rates of death and coronary revascularization differed insignificantly at 30 days but significantly at 6 months. Detailed inspection of our results reveals that the bulk of the added risk at 6 months was due to non-cardiac mortality.
ACCEPTED MANUSCRIPT 1. INTRODUCTION Cardiac troponin I and troponin T are the most specific biochemical markers for the diagnosis of
T
myocardial injury [1]. Initial studies and consensus recommendations led to the acceptance of the 99th
RI P
percentile among apparently normal people as the cutoff for “normal” [2]. This of course implies that 1% of patients without acute myocardial injury have a result above this threshold.
SC
Initial enthusiasm about the specificity of troponin assays has given way to the realization that troponin can be elevated in conditions other than acute myocardial injury, including chronic renal failure,
NU
chronic heart disease, and skeletal muscle disease [1,3]. In such cases, a low-positive troponin result may
MA
be an important prognostic indicator for long-term serious outcomes due to causes other than coronary occlusion, but the relevance of this result during a "rule-out myocardial infarction" evaluation is poorly understood [4]. Additional data regarding the short and intermediate-term prognosis among such patients
ED
are needed to support evidence-based decision-making.
PT
Chest pain accounts for 5-7% of all ED visits in the United States, with >6,000,000 visits annually [5,6]. History, physical exam, ECG and cardiac biomarkers are not sufficiently sensitive to rule
CE
out acute coronary occlusion during the initial ED evaluation [5]. Prior studies suggest that between 2-8% of ED patients with acute myocardial infarction and unstable angina were discharged home and missed
AC
myocardial infarction accounted for about 20% of malpractice claims against emergency physicians [7,8]. ED observation units were created to improve detection of acute ischemia by providing the time needed for ongoing evaluation, without having to commit a large number of low-risk patients to costly full inpatient admission [9,10]. In this study we seek to quantify the rate of serious health outcomes among patients with lowpositive initial troponin values who are not initially admitted to an inpatient setting. We compare outcomes in the “low positive” group to outcomes in the “negative” group.
ACCEPTED MANUSCRIPT 2. METHODS 2.1 Study Design and Setting
T
We performed a retrospective cohort study at an urban, academic, adult-only, tertiary-care ED.
RI P
Board-certified or board-eligible emergency physicians supervise residents and physician assistants, the usual first-line providers. All troponin testing was ordered by a physician or physician assistant; no
SC
standing orders permitted troponin ordering directly by a nurse. Our cohort was seen in the ED from 12/1/2009-11/30/2010. This date range was chosen because 12/1/2009 was the first day a new troponin
NU
assay was used. Prior to this date, our institution used the Siemens TnI-Ultra, which reported all results <0.03 ng/mL as “less than assay.” From this date forward, our institution used a 4th-generation
MA
conventional troponin T assay (Roche Diagnostics), which reports all results ≥0.01 ng/mL, and designates all non-zero results as abnormal. Our clinicians were confused because there were now patients with
ED
troponin values of 0.01-0.02 ng/mL, when previously all such values had been reported as “less than
PT
assay.” We refer to results in this range (0.01-0.02 ng/mL) as “low-positive.” Providers were notified of the change in an email, but no other formal education or direction was offered. Prior to the change in
CE
assay, our ED’s usual practice was to admit all patients with a newly positive troponin to the hospital. After the introduction of the new assay, there was confusion about how to manage clinically low-risk
AC
patients with troponin results below the reporting limit of the prior assay. At all times, as per the standard of care in emergency medicine, all patients with worrisome ECG findings or with moderate or high gestalt clinical probability for acute coronary occlusion, were admitted to an inpatient service, regardless of troponin results.
2.2 Selection of Participants Our goal in this investigation was to examine outcomes among patients with low-positive initial troponin results who were not admitted to an inpatient setting. We used our computerized ED order entry system to generate a list of all patients with an order for troponin testing during the study period. We then excluded all whose initial disposition was inpatient admission. We also excluded trauma patients who had
ACCEPTED MANUSCRIPT troponin testing to rule out myocardial contusion. Finally, we excluded patients discharged against
T
medical advice and those who died in the ED.
RI P
2.3 Data Collection and Processing
Trained research assistants collected the study data from our electronic medical record, which is
SC
shared by all of the ambulatory clinics affiliated with the ED and by seven area hospitals, using a standardized data collection form. A physician investigator reviewed all cases with an outcome (defined
NU
below). We sent all English and Spanish-speaking patients with a domestic mailing address who had
MA
incomplete 30-day and/or 6-month follow up data a letter asking them about outcomes and requesting permission to obtain outside hospital records. We made follow up calls to those patients who did not respond to the letter after 6 weeks. Patients who lived abroad or did not speak English or Spanish were
ED
analyzed as lost to follow up (see below).
PT
While our main outcome was death or revascularization, we knew that we would be unable to contact some patients. Thus, we used all-cause mortality as an important secondary outcome measure. We
CE
ascertained mortality by searching the Social Security Death Index (SSDI) [11]. We waited at least one year to check the index to ensure adequate time had passed so that deaths occurring during the follow up
AC
window had been reported. We accessed a current version of the SSDI Master File via our institution’s Research Patient Data Registry, a centralized data warehouse that linked patients in our study to deaths listed in the Master File within the study timeframe. We categorized revascularization events as "urgent" or "non-urgent" according to the operative report.
2.4 Outcome Measures Our main outcome measure was the risk difference for the patient-centered composite outcome of death or coronary revascularization at 30 days, among negative versus low-positive troponin cases. We also evaluated the composite outcome at 6 months; and death alone at 30 days and 6 months. We also report the number of subsequent inpatient hospitalizations with a primary discharge diagnosis of
ACCEPTED MANUSCRIPT myocardial infarction or other serious cardiovascular cause (pulmonary embolus, aortic dissection and stroke) by 30 days and 6 months, following prior investigations [12,13]. We did not consider further
RI P
T
diagnostic testing to represent an outcome.
2.5 Data Analysis
SC
We calculated risk differences and their 95% confidence intervals (95%CI) for each outcome. Our prospective sample size calculations indicated that 1,980 visits would be needed to achieve 80%
NU
power to detect an absolute difference of 0.5% versus 2% for the main outcome, with alpha=0.05.
MA
We constructed multivariable models to identify and control for confounding of the relationship of low-positive troponin to outcomes. A priori, we considered the following covariates to be of potential relevance: gender, age, race, prior history of myocardial infarction, coronary revascularization, high
ED
cholesterol, hypertension, congestive heart failure, end-stage renal disease, diabetes mellitus, tobacco use,
PT
and recent cocaine use. An example of how confounding could occur would be if diabetics were more likely to be revascularized for various reasons, and were also more likely to have a mild troponin T
CE
elevation because they are more likely to have renal insufficiency. We used the method of Bursac to identify confounders [14]. We found no evidence of confounding and the results of this analysis are not
AC
shown (but are available in the Technical Appendix). We were also concerned about the possibility that low-positive troponin results might have different implications in some subgroups. Therefore, we sought evidence of effect modification. We constructed logistic regression models with each of the outcomes mentioned above as dependent variables, and with the following predictors: troponin, age, end-stage renal disease, and presence or absence of history of congestive heart failure, as well as 2-way interaction terms of each of these with troponin. We identified no interactions (i.e. no evidence of different effects at different levels of these covariates), and the results of this analysis are not shown (but are available in the Technical Appendix). We performed all analyses using SAS 9.1 (Cary, NC). This study was approved by our Institutional Review Board.
ACCEPTED MANUSCRIPT
3. RESULTS
T
3.1 Characteristics of Study Subjects
RI P
Figure 1 is a participant flow chart, and Table 1 describes the cohort. During the study period, there were 57,596 ED visits, with 16,096 (27.9%) admitted, 33,388 (58.0%) discharged home, and 6,410
SC
(11.1%) assigned to the ED observation unit. The remainder had alternative dispositions such as death, left without being seen, inter-facility transfer, or other. Of visits with an initial disposition of home or ED
NU
observation, troponin was measured in 2,684 (6.7%). Results were negative in 2,523 (94.0%), low-
MA
positive in 78 (2.9%), and positive (≥0.03 ng/mL) in 83 (3.1%). At 30 days, data on coronary revascularization and hospitalization were available for 86% and 90% of those with initially negative and low-positive troponin, respectively. At 6 months, data on
ED
coronary revascularization and hospitalization were available for 83% and 90% of those with initially
PT
negative and low-positive troponin, respectively. The SSDI could capture all deaths except in unusual circumstances, such as foreign nationals. We had no reason to suspect differential misclassification
AC
3.2 Main Results
CE
according to follow-up data availability.
The main outcome occurred in 17, or 0.8% (95%CI 0.4%-1.1%) of troponin-negative cases, versus 2 or 2.8% (95%CI 0.0%-6.7%) of low-positive cases (risk difference 2.0%, 95%CI -1.8%-5.9%) (Table 2). At 6 months, the composite outcome occurred in 35, or 1.7% (95%CI 1.1%-2.2%) and 9, or 12.9% (95%CI 5.0%-20.7%) (risk difference 11%, 95%CI 3.3%-19.1%,). Death at 30 days occurred in 9, or 0.4% (95%CI 0.1%-0.6%) versus 1, or 1.3% (95%CI 0.0%-3.8%) (risk difference 0.9%, 95%CI -1.6%3.4%). Death at 6 months occurred in 30, or 1.2% (95%CI 0.8%-1.6%) versus 9, or 11.7% (95%CI 4.5%18.9%) (risk difference 10%, 95%CI 3.3%-17.7%). At 30 days there were nine revascularization events. Of these, three were urgent (two from the negative troponin group and one from the low-positive troponin group). Between 30 days and 6 months
ACCEPTED MANUSCRIPT there were seven additional revascularizations, four of which were urgent (all four from the negative troponin group).
T
We assessed the rate of hospitalization with a primary discharge diagnosis of myocardial
RI P
infarction, pulmonary embolus, aortic dissection, or stroke by 30 days and 6 months in all three groups. In the negative troponin group there were 9, or 0.4% (95%CI 0.1%-0.7%) at 30 days and 13, or 0.6%
SC
(95%CI 0.3%-1.0%) at 6 months. In the low positive troponin group there were 0, or 0.0% at 30 days and 1, or 1.5% (95%CI 0.0%-4.5%) at 6 months. The risk difference at 6 months was 0.9% (95%CI 2.1%-
MA
NU
3.9%).
4. LIMITATIONS
Loss to follow-up is a fundamental challenge in any follow-up study of a rare event, and this
ED
study is no exception. Even with a prospective design, some loss to follow-up is inevitable. For our
PT
primary outcome, we had definitive data on whether revascularization was performed on only 86% of our troponin negative and 90% of our low-positive participants. The concern is that we might have missed
CE
coronary revascularization outcomes among the low-positive initial troponin cases, and if we were more
AC
likely to miss outcomes in the low-positive group than in the negative group, our results would be skewed. Our cohort had 78 cases with low-positive initial troponin results. We had definitive follow-up data for 70 of them. In this group of 70, there was one case of revascularization (1.4%). Given this, had we missed one additional case of revascularization among those eight visits and none among the 349 visits with incomplete follow up in the troponin negative group, our primary outcome would be different. However, we were fortunate to be able to rely on the SSDI for mortality. In retrospect, it might have been wise to frame mortality as our main outcome measure, which is a framework previously used to follow patients after an ED visit for conditions such as syncope [15]. It also bears mentioning that the SSDI is not 100% perfect. For example, foreigners visiting the United States and entering our cohort would be missed if they died overseas [11].
ACCEPTED MANUSCRIPT Because this was a single-center study, generalizability is a concern. If clinicians at other centers were less cautious in their initial decision to admit or not admit, or to test or not test, measured outcomes
T
using this design could be worse, in troponin-negative or low-positive cases, or both. Our cohort was
RI P
derived from an institution that had recently switched from the Siemens TnI-Ultra to the Roche 4thgeneration troponin T assay, and this could have led our clinicians to lower or raise their threshold for
SC
admission of low-risk patients. The width of our main outcome's 95% confidence interval is discussed in the Discussion section proper.
NU
We focused only on patients who were clinically low risk (as evidenced by initial disposition) and
MA
had a negative or low-positive initial troponin result. It was beyond the scope of this investigation to evaluate changes in troponin results over time. Many of our patients had only a single troponin assay. Very few had changes in the results (data not shown), reflecting the low-risk status of the cohort; this
PT
ED
study was not powered to evaluate that group.
5. DISCUSSION
CE
To evaluate the significance of low-positive initial troponin results among ED patients, we compared outcomes among patients with low-positive results to outcomes among patients with negative
AC
results. We found no significant difference in mortality or coronary revascularization at 30 days. This suggests that patients with low-positive troponin values using the Roche 4th-generation troponin T assay without high-risk clinical indicators of acute coronary occlusion may be candidates for outpatient management. This conclusion has two caveats. First, our measured risk difference for revascularization or death, 2.0%, and its 95%CI of −1.8% to 5.8%, leave open the possibility that a much larger study could find significantly increased 30-day risk in the low-positive group. However, as an outcome, mortality has greater validity, because catheterization is subject to diagnostic workup bias. Patients with detectible troponin values are more likely to have additional testing, and medical testing drives medical procedures. It is impossible to know, in retrospect, whether a cardiologist’s decision to dilate or stent a stenotic vessel
ACCEPTED MANUSCRIPT is a good indicator of that patient’s “need” for that procedure. Measurement of fractional flow reserve lends some objectivity to the decision to intervene, but this technique was not in routine use at our
T
institution during the study period [16]. Of note, the vulnerability of our revascularization outcome to
RI P
workup bias renders our analysis conservative, because its effect would be unidirectional, away from the null. Thus, the fact that the risk difference for mortality was lower, and its 95%CI narrower, are
SC
somewhat reassuring. Nevertheless, the upper bound of our mortality outcome's 95% confidence interval was 3.4%, emphasizing the need for further study in larger samples.
NU
The second caveat is that the clear difference in outcomes by 6 months suggests that patients with
MA
an initial low-positive troponin values are indeed at higher risk than those with negative results. But it appears that they are at risk for death due to causes other than acute coronary occlusion. Inspection of our results reveals that the bulk of the added risk at 6 months was due to non-cardiac mortality (Table 3).
ED
Comparing outcomes at 6 months versus 30 days also suggests that mortality, but not coronary
PT
revascularization, increased over time. This suggests that the patients with initially low-positive troponin results were, in fact, at higher risk, but not because of their coronary arteries, and not in the short term.
CE
(Differential loss to follow-up in the low-positive versus negative groups with respect to revascularization would undermine these conclusions, but we have no reason to suspect that the patients seen in our ED
AC
with low-positive troponin results would be more likely to be catheterized outside our system than those with negative troponin results.) These results are important because the standard of care for low-risk chest pain patients continues to evolve. In the past, all “rule-out MI” patients were admitted to an inpatient setting. Now, management of low-risk chest pain patients in ED observation units is commonplace. At one time, protocols required observation stays of up to twenty-four hours, and stress testing prior to discharge was nearly universal. Over the years, the time frame for repeat troponin testing has contracted to 6 hours or even less, and the need for provocative testing has been questioned [17-20]. Taken together, these trends and observations suggest that in the future, patients deemed by clinician gestalt or via validated risk stratification tools to be at low risk for acute coronary syndrome might safely be discharged after negative or low-positive
ACCEPTED MANUSCRIPT troponin results, with outpatient follow-up. Moreover, cycling of a “delta troponin” in as little as one to two hours is showing promise as a new strategy to rapidly exclude the diagnosis of acute myocardial
T
ischemia [18,21]. Pressure to treat patients as outpatients or "observation" patients, rather than
RI P
"inpatients," continues to mount. Understanding the meaning of low-positive troponin results is crucial to appropriate management. These patients may be well served by avoiding short heart-focused admissions,
SC
in favor of more intensive outpatient care to address other causes of mortality. Any positive troponin should alert clinicians to a chance of mortality greater than 12% in 6 months.
NU
The fact that troponin indicates non-coronary mortality is of historical interest. When troponin
MA
assays were first commercialized, clinicians celebrated their specificity for myocardial infarction. Initial confidence soon gave way to more circumspection, as it became clear that troponin could be elevated in many circumstances other than acute myocardial infarction. Then, non-coronary troponin elevations were
ED
found to predict mortality in the intensive-care setting, where short-term mortality data are easiest to
PT
collect [3,22,23]. Other studies examining the general population have also shown a connection between troponin elevation and mortality over longer periods of time [24]. Troponin is increasingly viewed as a
CE
non-specific marker of mortality [25]. The present study adds to these prior findings by providing data on mortality after measurement of troponin in the ED setting.
AC
A further complication is found in the fact that, in recent years, the proliferation of troponin assays has introduced variability in cutoffs [26]. As institutions adopt different assays, providers may be confused about how to interpret values close to the “normal” range [27]. For example, the Siemens TnIUltra assay has a cutoff for “normal” of 0.03 ng/mL, and reports lower results as “less than assay.” In contrast, the Roche 4th generation troponin T assay reports results as low as 0.01 ng/mL, and all non-zero values are considered “abnormal.” Clinicians switching between these assays could be confused about the interpretation of results between 0 and 0.03 ng/mL [27]. This happens when an institution changes its assay or when data from one institution are interpreted by clinicians at another. Given the imminent introduction of a new class of “highly-sensitive” troponin assays to the United States, this issue will take on even more importance in the near future [28]. Thus, two factors confound clinicians’ daily
ACCEPTED MANUSCRIPT interpretation of positive, but very low troponin results: uncertainty as to their biological significance, and keeping track of the subtle differences in test characteristics from one assay to the next.
T
Normalization, as was done with the International Normalized Ratio (INR) might be one solution.
RI P
For example, instead of reporting the absolute value of the troponin concentration, laboratories could report the cumulative distribution function (i.e. the percentile) for that result according to the distribution
SC
of results observed by the manufacturer during elucidation of test characteristics leading to FDA approval. However, this goal is more complicated than it seems on its surface, and efforts to derive
NU
methods for normalization have been underway for some time [29].
MA
Clinical judgment is the cornerstone of risk stratification in patients who are evaluated for acute coronary occlusion but do not have ECG changes or other clear clinical indicators of its presence. Risk stratification tools, such as the Thrombolysis In Myocardial Infarction (TIMI) score, the Global Registry
ED
of Acute Coronary Events (GRACE) score, the Goldman rule, and the Acute Cardiac Ischemia-Time
PT
Insensitive Predictive Instrument (ACI-TIPI) hold promise for the differentiation of high-risk from lowrisk patients [30-33]. However, none of these tools has proven sufficiently accurate to identify patients
CE
who are safe for discharge from the ED without serial cardiac markers or provocative testing. Combinations of several risk scores may improve predictive performance, but may also be so
AC
cumbersome as to be difficult to implement in daily practice [34]. Perhaps more importantly, the patients included in the TIMI, GRACE, and ACI-TIPI studies were selected by experts in the acute coronary syndrome; and we can rest assured that practicing clinicians will continue to obtain troponin measurements in patients more diverse than those included in such studies. Thus, the significance of lowpositive results will remain important.
6. CONCLUSION A low-positive troponin result in a patient deemed to be at low risk for acute myocardial infarction is a harbinger of other causes of mortality in the succeeding few months, but may not indicate risk of serious outcomes due to acute coronary occlusion. Providers should consider alternative potential
ACCEPTED MANUSCRIPT causes of near-team mortality in these patients. Additional studies are warranted, in other ED populations
AC
CE
PT
ED
MA
NU
SC
RI P
T
and with larger samples.
ACCEPTED MANUSCRIPT REFERENCES [1] Panteghini M. Acute coronary syndrome: biochemical strategies in the troponin era. Chest 2002;122:1428-
RI P
T
35.
[2] Thygesen K. Alpert JS. White HD. Joint ESC/ACCF/AHA/WHF Task Force for the Redefinition of
SC
Myocardial Infarction. Jaffe AS. Apple FS. Galvani M. Katus HA. Newby LK. Ravkilde J. Chaitman B.
NU
Clemmensen PM. Dellborg M. Hod H. Porela P. Underwood R. Bax JJ. Beller GA. Bonow R. Van der Wall
MA
EE. Bassand JP. Wijns W. Ferguson TB. Steg PG. Uretsky BF. Williams DO. Armstrong PW. Antman EM.
ED
Fox KA. Hamm CW. Ohman EM. Simoons ML. Poole-Wilson PA. Gurfinkel EP. Lopez-Sendon JL. Pais P.
PT
Mendis S. Zhu JR. Wallentin LC. Fernandez-Aviles F. Fox KM. Parkhomenko AN. Priori SG.
CE
Tendera M.
Voipio-Pulkki LM. Vahanian A. Camm AJ. De Caterina R. Dean V. Dickstein K. Filippatos G.
AC
Funck-
Brentano C. Hellemans I. Kristensen SD. McGregor K. Sechtem U. Silber S. Tendera M. Widimsky P.
Zamorano JL. Morais J. Brener S. Harrington R. Morrow D. Lim M. Martinez-Rios MA. Steinhubl S. Levine GN. Gibler WB. Goff D. Tubaro M. Dudek D. Al-Attar N. Universal definition of myocardial infarction. Circulation 2007;116:2634-53. [3] Babuin L, Vasile VC, Rio Perez JA, et al. Elevated cardiac troponin is an independent risk factor for short-
ACCEPTED MANUSCRIPT and long-term mortality in medical intensive care unit patients. Crit Care Med 2008;36:759-65. [4] Daniels LB, Laughlin GA, Clopton P, Maisel AS, Barrett-Connor E. Minimally elevated cardiac
T
troponin T
the
SC
Rancho Bernardo Study. J Am Coll Cardiol 2008;52:450-9.
RI P
and elevated N-terminal pro-B-type natriuretic peptide predict mortality in older adults: results from
[5] Hess EP, Brison RJ, Perry JJ, et al. Development of a clinical prediction rule for 30-day cardiac
NU
events in
MA
emergency department patients with chest pain and possible acute coronary syndrome. Ann Emerg Med 2012;59:115,25.e1.
ED
[6] Chandra A, Rudraiah L, Zalenski RJ. Stress testing for risk stratification of patients with low to
PT
moderate
probability of acute cardiac ischemia. Emerg Med Clin North Am 2001;19:87-103.
associated
CE
[7] Schull MJ, Vermeulen MJ, Stukel TA. The risk of missed diagnosis of acute myocardial infarction
AC
with emergency department volume. Ann Emerg Med 2006;48:647-55. [8] Pope JH, Aufderheide TP, Ruthazer R, et al. Missed diagnoses of acute cardiac ischemia in the emergency department. N Engl J Med 2000;342:1163-70. [9] Graff LG, Dallara J, Ross MA, et al. Impact on the care of the emergency department chest pain patient from the chest pain evaluation registry (CHEPER) study. Am J Cardiol 1997;80:563-8. [10] Jaffe AS, Babuin L, Apple FS. Biomarkers in Acute Cardiac Disease: The Present and the Future. J Am Coll Cardiol 2006;48:1-11.
ACCEPTED MANUSCRIPT [11] Schisterman EF, Whitcomb BW. Use of the Social Security Administration Death Master File for ascertainment of mortality status Population Health Metrics 2004;2:1-6.
T
[12] Poon M, Cortegiano M, Abramowicz AJ, et al. Associations between routine coronary computed
RI P
tomographic angiography and reduced unnecessary hospital admissions, length of stay, recidivism rates, and
SC
invasive coronary angiography in the emergency department triage of chest pain. J Am Coll Cardiol \ 2013;62:543-52.
NU
[13] Lipinski MJ, McVey CM, Berger JS, Kramer CM, Salerno M. Prognostic value of stress cardiac
MA
magnetic
resonance imaging in patients with known or suspected coronary artery disease: a systematic review and
ED
meta-analysis. J Am Coll Cardiol 2013;62:826-38.
PT
[14] Bursac Z, Gauss CH, Williams DK, Hosmer DW. Purposeful selection of variables in logistic regression.
CE
Source Code for Biology and Medicine 2008;3:1-8.
after a
AC
[15] Derose SF, Gabayan GZ, Chiu VY, Sun BC. Patterns and preexisting risk factors of 30-day mortality
primary discharge diagnosis of syncope or near syncope. Acad Emerg Med 2012;19:488-96. [16] De Bruyne B, Pijls NH, Kalesan B, et al. Fractional flow reserve-guided PCI versus medical therapy in stable coronary disease. N Engl J Med 2012;367:991-1001. [17] Prasad V, Cheung M, Cifu A. Chest pain in the emergency department: the case against our current practice of routine noninvasive testing. Arch Intern Med 2012;172:1506-9. [18] Than M, Cullen L, Aldous S, et al. 2-Hour accelerated diagnostic protocol to assess patients with chest pain
ACCEPTED MANUSCRIPT symptoms using contemporary troponins as the only biomarker: the ADAPT trial. J Am Coll Cardiol 2012;59:2091-8.
T
[19] Redberg RF. Coronary CT angiography for acute chest pain. N Engl J Med 2012;367:375-6.
RI P
[20] Scheuermeyer FX, Innes G, Grafstein E, et al. Safety and efficiency of a chest pain diagnostic algorithm with
SC
selective outpatient stress testing for emergency department patients with potential ischemic chest pain. Ann
NU
Emerg Med 2012;59:256,264.e3.
MA
[21] Reichlin T, Schindler C, Drexler B, et al. One-hour rule-out and rule-in of acute myocardial infarction using
high-sensitivity cardiac troponin T. Arch Intern Med 2012;172:1211-8.
ED
[22] Ammann P, Maggiorini M, Bertel O, et al. Troponin as a risk factor for mortality in critically ill
PT
patients
without acute coronary syndromes. J Am Coll Cardiol 2003;41:2004-9.
CE
[23] Wu TT, Yuan A, Chen CY, et al. Cardiac troponin I levels are a risk factor for mortality and multiple organ
outcome
AC
failure in noncardiac critically ill patients and have an additive effect to the APACHE II score in
prediction. Shock 2004;22:95-101. [24] de Lemos JA, Drazner MH, Omland T, et al. Association of troponin T detected with a highly sensitive assay and cardiac structure and mortality risk in the general population. JAMA 2010;304:2503-12. [25] Mahajan VS, Jarolim P. How to interpret elevated cardiac troponin levels. Circulation 2011;124:2350-4. [26] Apple FS, Collinson PO, IFCC Task Force on Clinical Applications of Cardiac,Biomarkers. Analytical
ACCEPTED MANUSCRIPT characteristics of high-sensitivity cardiac troponin assays. Clin Chem 2012;58:54-61. [27] Apple FS. A new season for cardiac troponin assays: it's time to keep a scorecard. Clin Chem
T
2009;55:1303-
RI P
6.
[28] Storrow AB, Lardaro TA, Alexander PT, Apple FS. How Low Can We Go? The High-Sensitivity
SC
Cardiac Troponin Debate. Ann Emerg Med 2013;.
NU
[29] Christenson RH, Bunk DM, Schimmel H, Tate JR, IFCC Working Group on Standardization of
MA
Troponin,I.
Point: Put simply, standardization of cardiac troponin I is complicated. Clin Chem 2012;58:165-8. [30] Goldman L, Weinberg M, Weisberg M, et al. A computer-derived protocol to aid in the diagnosis of
ED
emergency room patients with acute chest pain. N Engl J Med 1982;307:588-96.
PT
[31] Antman EM, Cohen M, Bernink PJ, et al. The TIMI risk score for unstable angina/non-ST elevation MI: A
CE
method for prognostication and therapeutic decision making. JAMA 2000;284:835-42.
predictive
AC
[32] Selker HP, Beshansky JR, Griffith JL, et al. Use of the acute cardiac ischemia time-insensitive
instrument (ACI-TIPI) to assist with triage of patients with chest pain or other symptoms suggestive of acute cardiac ischemia. A multicenter, controlled clinical trial. Ann Intern Med 1998;129:845-55. [33] Eagle KA, Lim MJ, Dabbous OH, et al. A validated prediction model for all forms of acute coronary syndrome: estimating the risk of 6-month postdischarge death in an international registry. JAMA 2004;291:2727-33. [34] Marcoon S, Chang AM, Lee B, Salhi R, Hollander JE. HEART score to further risk stratify patients with low TIMI scores. Crit pathw cardiol 2013;12:1-5.
ACCEPTED MANUSCRIPT Figure 1: Patient Selection Strategy and Primary Outcome
RI P
T
57,596 adult ED visits between 12/1/2009 and 11/30/2010
CE
9/2,523 deaths in troponin negative group
AC
1/78 deaths in troponin 0.01-0.02 ng/mL group
PT
2,601 visits with 30day mortality follow up completed via chart review and SSDI
NU
83 visits with initial troponin ≥0.03mng/ mL
ED
2,684 visits with initial troponin negative or 0.010.02 ng/mL
245 visits excluded as trauma patients
MA
2,929 visits with a disposition of home or ED observation and a troponin resulted in the ED
SC
39,798 visits with an initial disposition of home or ED observation
2,244 visits (86%) with 30-day revascularization follow up completed via chart review, letter and/or call
8/2,174 revascularizations in troponin negative group
17/2,174 deaths or revascularizations in troponin negative group
2/70 deaths or revascularizations in troponin 0.01-0.02 ng/mL group
1/70 revascularizations in troponin 0.01-0.02 ng/mL group
ACCEPTED MANUSCRIPT Table 1: Patient Characteristics of Study Group Initial Troponin T Low Positive Negative
RI P
T
Patient Data # of visits
2,523
78
56.0
68.0
41.3%
56.4%
47.6% 28.9% 20.2% 1.4% 1.7%
55.1% 30.8% 11.5% 1.3% 0%
47.9%
66.7%
3.1%
3.8%
15.2%
17.9%
28.2%
50.0%
Prior history of myocardial infarction
17.7%
48.7%
Prior coronary revascularization
14.5%
39.7%
History of end-stage renal disease
0.7%
7.7%
History of congestive heart failure
11.6%
46.2%
ED length of stay (no observation used)*
4.5 hours (3.56.0 hours)
5.5 hours (3.57.5 hours)
ED length of stay (subsequent observation used)*
3.1 hours (2.34.0 hours)
3.2 hours (2.14.5 hours)
Observation length of stay, if used*
8.0 hours (5.713.7 hours)
7.8 hours (5.313.1 hours)
Age
SC
Male sex Race
MA
NU
White African American Hispanic Asian Other History of high cholesterol History of recent cocaine use
ED
History of tobacco use
AC
CE
PT
History of diabetes mellitus
*
0.01-0.02 ng/mL
Reported as the median (interquartile range)
ACCEPTED MANUSCRIPT Table 2: Main Results
Initial Troponin T Low-positive
Risk Difference
(
(0.01-0.02 ng/mL)
(95% confidence interval)
2.8% (95%CI 0.0%-6.7%)
2.0% (95%CI 1.8%-5.9%)
RI P
T
Negative
30 day death or revascularization
0.8% (95%CI 0.4%-1.1%)
Other Outcomes*
SC
Primary Outcome*
30 day revascularization
0.3% (95%CI 0.1%-0.5%)
1.4% (95%CI 0.0%-3.8%)
1.1% (95%CI 1.7%-3.9%)
6 month death or revascularization
1.7% (95%CI 1.1%-2.2%)
12.9% (95%CI 5.0%-20.7%)
11% (95%CI 3.3%-19%)
6 month death
1.2% (95%CI 0.8%-1.6%)
11.7% (95%CI 4.5%-18.9%)
10% (95%CI 3.3%-18%)
6 month revascularization
0.3% (95%CI 0.1%-0.6%)
0%
0.3%
30 day hospitalization†
0.4% (95%CI 0.1%-0.7%)
0%
0.4%
0.6% (95%CI 0.3%-1.0%)
1.5% (95%CI 0.0%-4.5%)
0.9% (95%CI 2.1%-3.9%)
ED
MA
NU
0.4% (95%CI 0.1%-0.6%)
CE
The denominators for the death and revascularization calculations are different because we found outcomes via different strategies; as a result, the composite outcome is not simply a sum of the individual outcomes Hospitalizations with primary discharge diagnosis of myocardial infarction, pulmonary embolism, stoke, or aortic dissection
AC
†
PT
6 month hospitalization†
*
1.3% (95%CI 0.0%-3.8%)
0.9% (95%CI 1.6%-3.4%)
30 day death
ACCEPTED MANUSCRIPT Table 3: Cause of Death of Study Participants* via Search of Death Records Number of Patients
Cancer, not specified
2
Lung cancer
2
Stroke
1
Pancreatic cancer
1
Multiple myeloma
1
Congestive heart failure
1
Drug overdose
1
Chronic obstructive pulmonary disease
1
Lung cancer
5
Congestive heart failure
4
ED
Cardiopulmonary arrest Brain cancer
PT
Ovarian carcinoma Cancer, not specified
Colon cancer
AC
Subdural hematoma
CE
Chronic obstructive pulmonary disease Gastric cancer
4 2 2 1 1 1 1 1
Pulmonary embolism
1
Common bile duct carcinoma
1
Renal cell carcinoma
1
Squamous cell carcinoma of the trachea
1
Ischemic colitis
1
Drug overdose
1
Sepsis
1
Alzheimer’s disease
1
Unknown/not available
13
Total
43
*
RI P SC
NU 10
MA
Total
30-day to 6 month Cause of Death
T
30-day Cause of Death
Troponin negative and low-positive patients only