a b s t r a c t

Objective: The objective of the study is to evaluate cardiac risk factors and risk scores for prediction of coronary artery disease (CAD) and adverse outcomes in an emergency department (ED) population judged to be at low to intermediate risk for acute coronary syndrome.

Methods: Informed consent was obtained from consecutive ED patients who presented with chest pain and were evaluated with coronary computed tomography angiography . Cardiac risk factors, clinical presentation, electrocardiogram, and laboratory studies were recorded; the Thrombolysis in Myocardial Infarction and Global Registry of Acute coronary events (GRACE) scores were tabulated. Coronary computed tomography angiography findings were rated on a 6-level plaque burden scale and classified for significant CAD (stenosis >=50%). adverse cardiovascular outcomes were recorded at 30 days.

Results: Among 250 patients evaluated by cCTA, 143 (57%) had no CAD, 64 (26%) demonstrated minimal plaque (b 30% stenosis), 26 (10%) demonstrated mild plaque (b 50% stenosis), 9 (4%) demonstrated moderate single vessel disease (50%-70% stenosis), 2 (1%) demonstrated moderate multivessel disease, and 6 (2%) demonstrated severe disease (N 70% stenosis). Six patients developed adverse cardiovascular outcomes. Among Traditional cardiac risk factors, only age (older) and sex (male) were significant independent predictors of CAD. Correlation with CAD was poor for the TIMI (r = 0.12) and GRACE (r = 0.09-0.23) scores. The TIMI and GRACE scores were not useful to predict adverse outcomes. Coronary computed tomography angiography identified severe CAD in all subjects with adverse outcomes.

Conclusion: Among ED patients who present with chest pain judged to be at low to intermediate risk for acute coronary syndrome, Traditional risk factors are not useful to stratify risk for CAD and adverse outcomes. Coronary computed tomography angiography is an excellent predictor of CAD and outcome.

(C) 2013

1. Introduction

The National Ambulatory Medical Care Survey demonstrates an increasing number of patients presenting annually to the emergency department (ED) with acute chest pain [1], with more than 7 million ED chest pain visits in 2010 [2]. As admission of chest pain patients to a coronary care unit is not cost-effective [3], Observation Units have been created to improve care and to reduce hospital admissions [4].

Assessment of the ED patient with chest pain begins with the medical history, evaluation of cardiac risk factors, and assessment of symptoms. The electrocardiogram (ECG), physical examination, and cardiac biomarkers may identify higher risk patients. Risk scoring systems-including the Thrombolysis in Myocardial Infarc-

* Corresponding author. Thomas Jefferson University Hospital, 132 South 10th St, Philadelphia, PA 19107-5244. Tel.: +1 215 955 5345; fax: +1 215 955 8549.

E-mail address: [email protected] (E.J. Halpern).

tion (TIMI) [5] and Global Registry of acute coronary events (GRACE) [6] scores-have been validated to confer additional im- portant prognostic value [7]. Triage is often based upon myocardial perfusion imaging [8] or other forms of Stress testing [9]. A scientific statement of the American Heart Association supports expedited management of Low-risk chest pain patients by combining clinical and laboratory assessments with a confirmatory stress test as “safe, accurate and cost-effective” [10].

Notwithstanding the effort and costs expended on ED evaluation of chest pain patients, conventional ED care results in a documented miss rate of 2% to 5% for acute coronary syndrome (ACS) [11,12]. Missed ACS and subsequent myocardial infarctions (MIs) account for 20% to 39% of all ED malpractice judgments [13]. Several randomized trials have suggested that coronary computed tomography angiogra- phy (cCTA) may be more effective than stress testing for evaluation of low-risk chest pain in the ED [14,15]. Three randomized multicenter trials have recently confirmed cCTA as a safe and cost-effective diagnostic test to discharge low-risk chest pain patients [16-18].

0735-6757/$ - see front matter (C) 2013 http://dx.doi.org/10.1016/j.ajem.2013.08.001

The general availability of cCTA has unleashed the possibility of widespread cCTA testing for low-risk ED patients presenting with chest pain, although the utility of such testing depends upon the pretest probability of disease [19]. The purpose of the current study was to evaluate whether conventional cardiac risk factors as well 2 widely accepted risk scores-TIMI and GRACE-can be used for triage in an ED population judged to be at low to intermediate risk for ACS and to limit the number of ED patients who should be referred for cCTA.

1. Methods
   1. Study design and setting

The study was conducted over a 36-month period from August 2009 through July 2012 on a prospective cohort of consecutive patients who presented to the ED of a busy urban academic center with chest pain or similar symptoms that might represent an anginal equivalent and who were admitted to the observation unit and evaluated with cCTA. This health insurance portability and account- ability act-compliant clinical protocol was approved by the university institutional review board; written informed consent was obtained from each participant before enrollment.

Selection of participants

The study population included patients presenting to the ED with a chief complaint of chest pain; shortness of breath; syncope or near syncope; or pain radiating to the neck, shoulder, back, or arm and not appearing to be musculoskeletal in nature. Each patient underwent electrocardiography and initial myoglobin and Troponin I levels. Patients who were considered high risk for ACS based upon clinical presentation, ECG changes (with clear ST-segment elevation or depression), or biomarkers (elevated troponins) were admitted to the hospital or sent directly for cardiac catheterization and were excluded from the study. Patients with a history of a prior revascular- ization procedure (bypass surgery or angioplasty +- stent), patients whose pain was deemed noncardiac, and patients with an estimated glomerular filtration rate below 60 mL/min were excluded. The remaining low-to-intermediate-risk patients-who would otherwise be evaluated with a stress echocardiogram or nuclear perfusion stress test before discharge-were referred for cCTA and included in the study whenever dedicated cardiac time on the computed tomographic scanner was available within 12 hours of presentation.

Methods and measurements

To prospectively collect information on study participant risk factors, a coinvestigator who was not involved in performance of the cCTA and who was not involved in the care of the patient interviewed each patient with a standard structured questionnaire. This questionnaire was developed by consultation with our cardiologists and ED chest pain division to provide a complete, succinct description of the presenting complaint and cardiac risk factors that could be used for statistical analysis. Chest pain was described by location, character (sharp, aching, burning, tearing, crushing or pressure), radiation (to left arm, jaw/neck, back, right arm, or other), severity (1-10 scale), number of episodes, and duration of pain. Patients were asked if pain was exacerbated by exertion or emotional stress. Associated symptoms were recorded. Risk factor evaluation included history of hypertension, hyperlipid- emia, diabetes, smoking, prior myocardial infarct, history of known coronary disease, and family history of coronary disease. Medica- tions were recorded, and patients were specifically asked about aspirin usage in the previous 7 days. Patient laboratory studies obtained in the ED were reviewed for hyperlipidemia and elevation

of troponins. The 12-lead ECG obtained during ED presentation was reinterpreted for study purposes by a cardiology coinvestigator. The ED physical examination record was reviewed for body mass index, blood pressure, and Killip classification.

Cardiovascular risk scores

The TIMI score consists of 7 clinical variables with 1 point given for each condition including Age 65 years or older, the presence of at least 3 risk factors for coronary artery disease (CAD), a history of prior coronary stenosis greater than or equal to 50%, use of aspirin in the previous 7 days, severe angina defined as experiencing at least 2 episodes of angina within 24 hours, ST-segment deviation of at least

0.5 mm, and the presence of elevated serum cardiac markers. Details of the TIMI scoring system are available online at www.timi.org. As many of our patients complained of Atypical chest pain or continuous chest pain for an extended duration, we contacted the TIMI study group for clarification regarding the definition of severe anginal symptoms. The presence of continuous chest pain over a long interval was tabulated as multiple episodes for the purpose of the TIMI score (personal communication with Dr Elliott Antman).

The GRACE score is an alternative risk score to predict outcomes among patients presenting with ACS. The GRACE model uses various parameters including a patient’s age, Killip class, systolic blood pressure, heart rate, creatinine level, the presence of cardiac arrest on admission, ECG changes (definition is different than the TIMI criteria; ST-segment deviation must be at least 1 mm), and elevated serum cardiac markers to generate 4 probabilities: the Probability of death in the hospital and at 6 months from the time of hospital admission as well as the probability of death or MI in the hospital and at 6 months from the time of hospital admission. In contrast to the TIMI scores, the GRACE score is more complicated to compute and requires the use of a calculator. Details of the GRACE scoring system and a GRACE risk score calculator are available online at www. outcomes-umassmed.org/grace.

For the purpose of TIMI and GRACE scores, the ECG recorded in the ED at the time of presentation was reinterpreted by a cardiology coinvestigator using TIMI and GRACE criteria. Four sets of GRACE scores were calculated, including in-hospital and 6-month scores for death and for death/infarction. The research questionnaires as well as the TIMI and GRACE scores were not viewed by the ED physicians caring for the patients or by the physician who inter- preted the cCTA.

Computed tomographic scan protocol

Imaging was performed with a 256-MDCT scanner (Brilliance iCT; Philips Medical Systems, Cleveland, OH). Patients with initial heart rates greater than 60 beats per minute were treated with intravenous metoprolol (5-20 mg) to a target heart rate of 50 to 60 beats per minute. Sublingual nitroglycerin spray (800 ug) was administered 2 to 3 minutes before scanning. A biphasic injection protocol was used with 60 mL of ioversol (Optiray 350; Mallinckrodt Imaging, St. Louis, MO) followed by 30 mL of ioversol mixed with 30 mL of 0.9% saline solution, injected at 5 to 6 mL/s. To reduce radiation exposure, pros- pective ECG triggering with axial imaging was used in patients with a stable cardiac rhythm and heart rate 60 beats per minute or lower. Retrospective gating with helical imaging and tube current modula- tion was used in the remaining patients. Tube voltage was set to 120 kVp. For prospective axial scans, tube current was set at 60 to 100 mAs based upon patient size. Mean estimated effective biological dose for patients scanned with prospective ECG triggering was 4 mSv, assuming a 12-cm scan length and a normalization factor of 0.017 mSv x mGy^-1 x cm^-1 for the adult chest [20]. For helical scans, tube

current was generally set at 600 mAs per slice but was increased to 800 or 1000 mAs per slice for obese patients. Estimated mean bio- logical dose for patients scanned at 600 mAs per slice was 6 mSv, assuming a 12-cm scan length.

Computed tomographic image analysis

Reconstructed images of the coronary arteries were evaluated by 2 experienced cardiac radiologists. Both radiologists were aware of the clinical indication for each examination but were blinded to the detailed information obtained in the structured patient interviews as well as to the ECG findings and TIMI and GRACE scores. Disagree- ments between the 2 radiologists were resolved by consensus

discussion. Slab maximum intensity projection images were obtained with 5-mm slice thickness and rotated in various planes to best visualize any area of stenosis in all coronary segments. For any patient with evidence of stenosis, a tracked reconstruction of the coronary arteries was performed to allow evaluation with a curved multiplanar image and a straightened lumen view (comprehensive cardiac package software; Philips Brilliance workstation. See Fig.). Each coronary artery segment was rated as normal, minimal disease (b 30% stenosis), mild disease (b 50% stenosis), moderate disease (50%-70% stenosis), or severe disease (N 70% stenosis). Based upon these ratings, each cCTA study was classified on a 6-level plaque burden scale (0, no plaque; 1, minimal disease with b 30% stenosis; 2, mild disease with b 50% stenosis; 3, moderate single vessel disease

Fig. Coronary artery grading system for stenosis, demonstrating curved multiplanar and straightened lumen views of the left anterior descending artery in 6 patients. A, Normal left anterior descending artery. B, Minimal disease (arrow) with less than 30% stenosis. C, Mild disease (arrow) with less than 50% stenosis. D, Moderate disease (arrow) with 50% to 70% stenosis. E, Severe disease (arrow) with greater than 70% stenosis. Follow-up: treated with stent, symptoms resolved. F, Severe disease (arrow) with greater than 70% stenosis and more proximal disease with moderate 50% to 70% stenosis (arrowhead). Follow-up: Sent to catheterization laboratory for recurrent symptoms and diagnosed with ACS.

with 50%-70% stenosis; 4, moderate multivessel disease; and 5, severe disease with N 70% stenosis).

Outcomes

For the purpose of this study, a positive cardiac outcome was defined as confirmed ACS or MI, coronary revascularization, or cardiac-related death within 30 days. The diagnoses of ACS and MI were based upon clinical presentation and follow-up, ECG changes, and cardiac biomarkers. A coinvestigator attempted to obtain a telephone follow-up on each patient at 30 days after the cCTA. If there was no answer or the patient was unavailable, additional follow-up calls were made in an attempt to reach the patient for up to 1 year after the cCTA. For those patients who could not be reached by telephone, hospital medical records were searched for follow-up information including subsequent admissions to the hospital, cardiac events, or revascularization. Although the primary end point was to identify adverse outcomes at 30 days, in a minority of patients, these follow-up data were obtained from telephone conversations or medical records that were recorded up to 1 year after the cCTA.

Analysis

Based upon prior experience with cCTA at our institution, we expected to define significant CAD with greater than or equal to 50% stenosis in approximately 10% of our low-to-intermediate-risk ED patients, with adverse outcomes in no more than 2%. As the number of expected adverse outcomes was very small, the study was powered to identify risk factors that might improve the prediction of CAD and allow direct triage of these patients without cCTA. Based upon cost- effectiveness considerations, further noninvasive testing for CAD is unlikely to be cost-effective when the probability of significant CAD drops below 1% or rises close to 50% [19]. Power calculations were performed assuming a 2-sample comparison of proportions using a 2-sided test with an ? of .05. For a risk factor present in 50% of the study population, a sample size of 250 is calculated to provide 80% power to detect a reduction in CAD rate from 10% to 1% and 85% power to detect an increase in CAD rate above 25%. For a risk factor that is present in only 20% of the study population, a sample size of 250 is calculated to provide a power of 70% to detect an increase in CAD rate above 25% and 88% power to detect an increase in CAD rate above 30%. Statistical analysis was performed by using Stata software (version 12.0; StatCorp, College Station, TX). Risk factors included in the analysis are enumerated in Table 1. Two sets of dependent variables were used in the analysis, the presence of coronary disease on cCTA, and the clinical follow-up outcome data. For the purpose of evaluating the predictive value of risk factors on the presence of coronary disease, significant coronary disease was defined as greater than or equal to 50% stenosis in at least 1 vessel (a score >=3 in our grading system). To determine whether cCTA was predictive of adverse cardiac outcomes, we cross-tabulated the outcome data with the

presence of disease on cCTA.

Univariate analysis

Pearson Correlation coefficients were computed to evaluate correlation of TIMI and GRACE scores with the presence of CAD as demonstrated by cCTA. A ?² test was performed to test the association of binary risk factors (sex, hypertension, hyperlipidemia, diabetes, smoking, etc) with the presence of coronary disease or an adverse cardiac outcome. Logistic regression was used to calculate the odds ratio for the association of binary and continuous risk factors (age, body mass index, severity of pain, TIMI score, GRACE score) with the presence of significant CAD. When multiple categories of a single risk factor were present, as is the case for character of chest pain and radiation of chest pain, an odds ratio was computed for each individual category (ie, chest pressure, tightness, etc).

Multivariate analysis

For those risk factors that were found to be significant in the univariate analysis, multivariate logistic regression was performed to determine which of these risk factors provided independent signif- icant Predictive power for CAD. The TIMI and GRACE scores were included in a further multivariate analysis to determine whether they provided additional predictive information. P b .05 was considered statistically significant for all statistical analyses.

1. Results
   1. Characteristics of study subjects

Our cohort consisted of 250 patients evaluated with cCTA, with follow-up clinical data on 221 of 250 patients. Final analysis for the prediction of CAD was based on all 250 patients, whereas final analysis for prediction of outcome was based upon the 221 patients with clinical follow-up.

Demographics of the study population, tabulation of risk factors, and odds ratios for prediction of CAD are reported in Table 1. More than half the patients sent for cCTA had no visible coronary plaque (n = 143; 57% of subjects). Minimal plaque was present in 64 patients (26%); mild plaque, in 26 patients (10%); moderate Single-vessel disease was present in 9 patients (4%); moderate multivessel disease was present in 2 patients (1%); and severe plaque, in 6 patients (2%). For the purposes of our analysis, 17 patients were classified as having significant CAD (stenosis >=50%).

1. Main results

Among traditional risk factors, only age (older), sex (male), and hypercholesterolemia were significant predictors of CAD in the univariate analysis (Table 1). The mean age of patients with significant CAD (58.7 years) was significantly greater than the mean

Table 1

Demographics and risk factors

Demographics	Mean (range)	SD	Odds ratio to predict CAD >=50%
Age (y)	50.9 (22-81)	+-11	1.07 (P = .004)
Height (in)	66.6 (56-76)	+-4	1.08 (P N .2)
Weight (lb)	190 (84-330)	+-49	1.00 (P N .4)
Body mass index	30.1 (14-60)	+-7.4	0.96 (P N .2)
	Number	Percentage
White	92	37%
Black	131	52%
Hispanic	15	6%
Asian	8	3%
Other	4	2%
	Risk factor present	Frequency
Male sex	109	44%	3.4 (P = .02)
Hypertension	151	60%	1.6 (P = .37)
Hypertensive on meds	47	19%	1.9 (P = .24)
Hypercholesterolemia	93	37%	3.3 (P b .02)
Smoker	74	30%	1.3 (P = .61)
Diabetes	51	20%	1.7 (P = .35)
Prior infarction	7	3%	0.0 (P = .47)
Presentation of pain
Pressure	112	45%	0.6 (P = .28)
Sharp	80	32%	0.8 (P = .66)
Aching	43	17%	0.3 (P = .20)
Crushing	29	12%	0.4 (P = .41)
Burning	21	8%	0.6 (P = .64)
Tearing	4	2%	4.4 (P = .21)
Exertional	84	34%	2.4 (P = .10)
Radiating pain	140	56%	1.35 (P = .69)
Severity (1-10)	Mean, 6.1	SD, 2.9	0.9 (P = .29)
ECG
Abnl ECG (TIMI)	17	7%	0.9 (P = .88)
Abnl ECG (GRACE)	7	3%	2.4 (P = .44)

Table 2

Prevalence of significant CAD as a function of TIMI score

TIMI score	CAD score b 3 (b 50% stenosis)	Significant CAD (>=50% stenosis)
0	37	0
1	103	7
2	61	9
3	22	1
4	7	0
5	1	0

age of patients without significant CAD (50.4 years). Other factors including diabetes, hypertension, hypertension on medication, family history, smoking history, history of prior infarction, character of chest pain, radiation of chest pain (to the arm, neck/jaw or back), and severity of chest pain were not significant predictors of CAD. On multivariate analysis, hypercholesterolemia was correlated with age but was not a significant independent predictor of CAD. Age and sex continued to be significant predictors on multivariate analysis. Of note, 5 of the 7 patients who claimed to have a history of prior MI demonstrated no evidence of CAD, and the other 2 had CAD with less than 50% stenosis on cCTA. Although reinterpretation of the present- ing ECG studies by our cardiology coinvestigator demonstrated an abnormal result in 17 patients based upon TIMI criteria and 7 patients based upon GRACE criteria (Table 1), only 1 patient with an Abnormal ECG result demonstrated significant CAD.

Table 2 tabulates the association of TIMI scores, which varied from 0 to 5, with the presence of significant CAD on cCTA. Of note, 16 of 17 patients with coronary stenosis greater than or equal to 50% had a TIMI score less than or equal to 2, whereas 30 of 31 patients with a TIMI score greater than or equal to 3 had no significant coronary stenosis. Table 3 presents the Pearson correlation coefficients between risk scores (TIMI and GRACE) and the cCTA grade of CAD, demonstrating only a weak correlation between risk score and plaque burden. Although TIMI and GRACE scores were significantly correlat- ed with each other, the r values for the correlation of TIMI with the various GRACE scores were in the range of 0.24 to 0.31 (P b .001).

The data in Tables 1, 2, and 3 demonstrate that many patients in our study with significant coronary disease on cCTA were not identified by traditional risk factor and/or risk score assessment. Furthermore, most patients with elevated TIMI and GRACE scores in this study were, in fact, free of coronary disease. Logistic regression of TIMI scores for prediction of significant CAD with stenosis greater than or equal to 50% failed to demonstrate a significant association. Logistic regression of GRACE scores for prediction of CAD with stenosis greater than or equal to 50% did demonstrate a significant association, but the odds ratio was in the range of 1.01 to 1.03, suggesting weak Predictive ability. Furthermore, when any of the GRACE scores was combined with patient age and sex in a multivariate logistic analysis, the GRACE score did not provide additional predictive power for significant CAD (odds ratios, 0.98- 0.99; P >= .4).

Follow-up clinical data were obtained in 221 of 250 patients, demonstrating 6 adverse cardiovascular events within 30 days. The 6

Table 3

Correlation of risk scores with the presence of significant CAD

Risk score Correlation with cCTA CAD rating Associated P value

TIMI score 0.12 .07 GRACE score

In-hospital death	0.16	.01
6-mo death	0.23	b.01
In-hospital death or MI	0.09	.17
6-mo death or MI	0.22	b.01

events included 2 patients with confirmed ACS, 2 patients with MI, and 2 patients who required revascularization (1 with coronary bypass surgery and 1 with a stent). All 6 adverse cardiovascular events occurred in patients with severe CAD identified by cCTA (cCTA grade, 5). Although 5 of 6 patients with adverse events were either hypertensive or diabetic and 2 of 6 did complain of exacerbation of chest pain with exertion, all 6 adverse cardiovascular events occurred in patients with a TIMI score of 2 or less (Table 4). Among the 31 patients with TIMI scores greater than or equal to 3, none experienced an adverse cardiovascular event on 30-day follow-up. Although all patients who experienced adverse cardiovascular events did have at least 1 cardiac risk factor, these risk factors were equally frequent among patients who did not suffer adverse events.

1. Discussion

Our study confirms the ability of cCTA to predict risk for adverse cardiovascular events, as all patients who had an adverse event by 30 days demonstrated the presence of severe CAD on cCTA. Our data also support the conclusion that it is safe to discharge a patient with a normal cCTA or a cCTA showing less than 50% stenosis, as no adverse outcomes were documented in the 233 patients with less than 50% coronary stenosis. Among conventional criteria used to risk stratify coronary disease, only increasing age and male sex were independent predictors of significant CAD. This finding is concordant with the landmark publication on the probability of CAD by Diamond and Forrester [21], which defined the pretest likelihood of disease based upon age, sex, and presentation of symptoms. Finally, our data suggest that other risk factors are poor predictors for the presence of CAD in ED patients and should not be used to triage low-to-intermediate-risk patients away from cCTA. These findings are concordant with those of the ROMICAT trial subanalyses that demonstrate only “modest” discriminatory capacity of the Goldman, Sanchis, and TIMI risk scores for the diagnosis of ACS, as compared with “good” discriminatory capacity for plaque burden on cCTA [22,23].

Numerous studies have validated TIMI and GRACE risk scores to predict patient outcome in the setting of ACS [24]. For patients presenting to the ED with potential ACS, TIMI and GRACE are the most commonly used scores for risk stratification, with receiver operating characteristic curve (ROC) areas of 0.757 and 0.728, respectively, for the prediction of 30-day event rates [25]. A comparison of TIMI, GRACE and PURSUIT scores demonstrated “good predictive accuracy for death or MI at 1 year, and enabled the identification of high-risk subsets of patients who will benefit most from myocardial revascu- larization performed during initial hospital stay [26].”

So why did these risk scores fare poorly in our study of patients who presented to our ED with potential ACS? Firstly, we should note that the ROC areas of 0.757 and 0.728 cited above do not suggest excellent predictive value. Furthermore, the comparison of TIMI, GRACE, and PURSUIT cited above with “good predictive accuracy” reports ROC areas of 0.551 to 0.585 for TIMI and 0.672 to 0.715 for GRACE.

An additional explanation for the poor performance of TIMI and GRACE scores in our study is likely related to the patient population. For patients with confirmed ACS, TIMI and GRACE scores are validated to predict outcome, but no risk score has been validated for identification of ACS in the ED setting [27]. The low-to-intermedi- ate-risk ED chest pain population is very different from the clinical trial and registry settings of ACS patients in which the TIMI and GRACE scores were developed. Furthermore, although 3-vessel and left main diseases are more common in ACS patients with TIMI scores of 3 to 4 as compared with patients with TIMI scores of 0 to 2, significantly more single-vessel disease may be present in ACS patients with TIMI scores of 0 to 2 as compared with those of scores of 5 to 7 [28]. It is likely that the ACS patients who are most commonly seen in the ED are those with single-vessel disease and lower risk scores rather than

those with extensive CAD and higher risk scores. Our data underscore the importance of risk model validation in an appropriate target population rather than a clinical trial population to establish its generalizability before integration into clinical practice [29].

Adverse outcome

MI ACS ACS CABG

Stent MI

Abbreviations: F, female; M, male; B, black; W, white; HTN, hypertension; N, no; Y, yes; Lt, left; RCA, right coronary artery; LAD, left anterior descending artery; LCx, left circumflex coronary artery; CABG, coronary artery bypass graft.

How can we best expedite management of the low-to-intermedi- ate-risk patient who presents to the ED with chest pain? Patients who present with chest pain that can be attributed to a noncardiac cause should require no further cardiac testing. Patients who appear to be at high risk for ACS based upon typical anginal symptoms, ECG findings, and biomarkers should be referred for cardiac catheterization. However, the remaining large group of low-to-intermediate-risk patients with atypical chest pain cannot be easily stratified by risk factors and scores. Randomized trials have demonstrated excellent negative predictive value for cCTA in evaluation of patients presenting to the ED with chest pain and have suggested that cCTA may be more cost-effective than traditional stress testing for rapid discharge of these patients [14-18]. Another significant advantage of cCTA over stress testing is the ability to assess simultaneously for Noncardiac etiologies of chest pain in the mediastinum, lungs, and chest wall. In appropriate clinical situations where both cardiac and noncardiac vascular causes of chest pain are suspected, a “triple rule-out” computed tomography angiography study can evaluate the aorta, coronary arteries, and pulmonary arteries and expedite discharge of up to 75% of patients presenting to the ED with atypical chest pain [30,31].

GRACE-in-hospital death/MI

109

76

98

73

70

91

GRACE-6-mo

death

114

66

90

73

71

81

GRACE-6-mo

death/MI

125

77

89

79

73

89

cCTA: findings of severe disease

RCA LAD

LCx LAD

3 vessels RCA

A recent editorial in the New England Journal of Medicine critiqued the use of cCTA for low-risk ED patients citing the “choosing wisely” campaign for diagnostic testing, and declared: “The underlying assumption…is that some diagnostic test must be performed before discharging these low-to-intermediate-risk patients from the emer- gency department. This assumption is unproven and probably unwarranted” [32]. This editorial appears to oppose the recommenda- tions of the 2010 American Heart Association scientific statement that supports expedited management of low-risk ED chest pain patients by combining clinical assessment with a confirmatory diagnostic test to exclude ischemia as “safe, accurate and cost-effective” [10]. Our study suggests that it would be unwise to discharge low-to-intermediate-risk ED patients based on risk factors and scores without confirmatory testing, as cCTA did identify a small percentage of patients with significant CAD (17/250, 6.8%) and 30-day adverse events (6/221, 2.7%) who might not otherwise be detected based upon clinical presentation to the ED. Furthermore, it is unlikely that ED physicians would be willing to discharge low-to-intermediate-risk patients when missed CAD/ACS and subsequent MIs account for 20% to 39% of all ED malpractice judgments [13]. A recent analysis by the Physician Insurers Association of America states “Between 1985 and 2012, the third most common patient condition for which claims were filed against emergency physicians were acute myocardial infarctions (AMIs). Claims involving AMIs resulted with an indemnity payment 52% of the time with an average payment of $261856″ [33].

Hypercholesterolemia

Family history

N Y N Y N Y

Smoker

Radiation of chest pain

N N N

Lt arm Lt arm N

Pain worse with exertion

Y N Y N N N

TIMI

GRACE-in-hospital death

130

88

133

87

84

105

N N N N Y N

2

1

2

1

2

2

Several limitations of our study design should be noted. The

Table 4

Risk factors and scores in patients with adverse outcomes

accuracy of risk factor assessment depends upon accurate recording of risk factors, and patients are often uncertain of their own history or unable to adequately explain their symptoms. This uncertainty is illustrated by the 7 patients who claimed to have had a prior MI but who had no significant coronary disease on cCTA. Furthermore, clinical history and risk factor assessment are often quite variable in the ED setting. For the current study, we attempted to address this issue by prospectively interviewing all patients with a standardized history and risk factor questionnaire that included all standard risk factors as well as all data needed to compute TIMI and GRACE risk scores. When there was uncertainty regarding a risk factor, we checked back with the medical records and ED chart. With respect to follow-up data, we initially wanted to collect all follow-up data at 30 days. However, as we were unable to contact many patients within this time interval, the telephone follow-up period was expanded to increase the follow-up rate, and additional follow-up was obtained from medical records.

Age

Sex

Race

HTN

Diabetes

81

54

61

64

64

59

F F M M M F

B B B W W W

N Y N Y Y N

Y N Y N Y N

Y N N Y Y N

Nonetheless, 30-day follow-up data could not be obtained for 29 patients, representing 11.6% of the patients involved in the study. Furthermore, although 30-day follow-up data should be sufficient to demonstrate the safety of ED discharge, this short-term follow-up is not sufficient to demonstrate the long-term prognostic utility of cCTA. Finally, as the focus of this study was upon CAD and cardiovascular outcomes, the 30-day follow-up was focused upon cardiac events, but we did not tabulate the noncardiac pathology detected by cCTA, nor did we obtain follow-up for noncardiac adverse events.

The relatively small study population and low rate of adverse outcomes among ED patients is a limitation for statistical analysis of the risk factors and might interfere with detection of statistically significant risk factors. Considering those variables with odds ratios greater than 1.5 in Table 1, it is possible that a larger study population would have found statistical significance in the risk factors of hypertension, diabetes, a description of tearing chest pain, and exacerbation of chest pain with exertion. Nonetheless, our outcome results demonstrate that these risk factors did not have good discriminating ability for adverse cardiovascular outcomes in our study population. To improve our statistical power to demonstrate significant risk factors, we chose to analyze the predictive value of traditional cardiac risk factors and risk scores for both CAD and clinical outcomes. Some degree of CAD was present in 43% of the study population, and significant CAD with greater than 50% stenosis was present in 7% of the study population. Because the development of ACS requires the presence of underlying coronary disease, CAD is reasonable as a proxy marker for potential future development of an adverse cardiovascular event.

In summary, our data demonstrate that among patients who present to the ED with chest pain and are assessed to be at low-to- intermediate-risk of ACS, traditional risk factors and TIMI and GRACE scores do not predict plaque burden or adverse cardiovascular outcomes. Coronary computed tomography angiography is an effective strategy to triage these patients, but risk factors cannot be used to decide which patients should have cCTA. Specifically, cCTA identified many patients with significant CAD who would be missed by evaluation of risk factors and TIMI and GRACE scores. Coronary computed tomography angiography expedited the discharge of more than 75% of patients-even those with higher risk scores-based upon normal or minimal disease in the coronary arteries. In conclusion, for a patient who presents to the ED with chest pain and is assessed to be at low to intermediate risk for ACS, cCTA is an appropriate diagnostic study and is superior to clinical assessment by risk factors and risk scores for patient triage.

References

1. Niska R, Bhuiya F, Xu J. National Hospital Ambulatory Medical Care Survey: 2007 emergency department summary. National Health Statistics Report #26, 8/6/2010.
2. National Hospital Ambulatory Medical Care Survey: 2010 Emergency Department Summary Tables. See http://www.cdc.gov/nchs/data/ahcd/nhamcs_emergency/ 2010_ed_web_tables.pdf.
3. Tosteson AN, Goldman L, Udvarhelyi IS, Lee TH. Cost-effectiveness of a coronary care unit versus an intermediate care unit for emergency department patients with chest pain. Circulation 1996;94:143-50.
4. Venkatesh AK, Geisler BP, Gibson Chambers JJ, Baugh CW, Bohan JS, Schuur JD. Use of observation care in US emergency departments, 2001 to 2008. PLoS One 2011;6(9):e24326. doi:10.1371/annotation/e7927237-fcce-4738-9564- e9f5d85341cb Epub 2011, Erratum in: PLoS One. 2012;7(1).
5. Antman EM, Cohen M, Bernink PJ, McCabe CH, Horacek T, Papuchis G, et al. The TIMI risk score for unstable angina/non-ST elevation MI: a method for prognostication and therapeutic decision making. JAMA 2000;284(7):835-42.
6. The G.R.A.C.E. Investigators. Rationale and design of the GRACE (Global Registry of Acute Coronary Events) project: a multinational Registry of patients hospitalized with acute coronary syndromes. Am Heart J 2001;141:190-9.
7. Yan AT, Yan RT, Tan M, Casanova A, Labinaz M, Sridhar K, et al. Risk scores for risk stratification in acute coronary syndromes: useful but simpler is not necessarily better. Eur Heart J 2007;28(9):1072-8.
8. Kontos MC, Schmidt KL, McCue M, Rossiter LF, Jurgensen M, Nicholson CS, et al. A comprehensive strategy for the evaluation and triage of the chest pain patient: a cost comparison study. J Nucl Cardiol 2003;10(3):284-90.
9. Chandra A, Rudraiah L, Zalenski RJ. Stress testing for risk stratification of patients with low to moderate probability of Acute cardiac ischemia. Emerg Med Clin North Am 2001;19(1):87-103.
10. Amsterdam EA, Kirk JD, Bluemke DA, Diercks D, Farkouh ME, Garvey JL, et al. American Heart Association Exercise, Cardiac Rehabilitation, and Prevention Committee of the Council on Clinical Cardiology, Council on Cardiovascular Nursing, and Interdisciplinary Council on Quality of Care and Outcomes Research. Testing of low-risk patients presenting to the emergency department with chest pain: a scientific statement from the American Heart Association. Circulation 2010;122(17):1756-76.
11. Pope JH, Aufderheide TP, Ruthazer R, Woolard RH, Feldman JA, Beshansky JR, et al. missed diagnoses of acute cardiac ischemia in the emergency department. N Engl J Med 2000;342:1163-70.
12. Schull M, Vermeulen MJ, Stukel TA. The risk of missed diagnosis of acute myocardial infarction associated with emergency department volume. Ann Emerg Med 2006;48:647-55.
13. Storrow AB, Gibler WB. Chest pain centers: diagnosis of acute coronary syndromes. Ann Emerg Med 2000;35:449-61.
14. Goldstein JA, Gallagher MJ, O’Neill WW, Ross MA, O’Neil BJ, Raff GL. A randomized controlled trial of multi-slice coronary computed tomography for evaluation of acute chest pain. J Am Coll Cardiol 2007;49(8):863-71.
15. Miller AH, Pepe PE, Peshock R, Bhore R, Yancy CC, Xuan L, et al. Is coronary computed tomography angiography a resource sparing strategy in the risk stratification and evaluation of acute chest pain? Results of a randomized controlled trial. Acad Emerg Med 2011;18(5):458-67.
16. Goldstein JA, Chinnaiyan KM, Abidov A, Achenbach S, Berman DS, Hayes SW, et al, CT-STAT Investigators. The CT-STAT (Coronary Computed Tomographic Angiog- raphy for Systematic Triage of Acute Chest Pain Patients to Treatment) trial. J Am Coll Cardiol 2011 Sep 27;58(14):1414-22.
17. Litt HI, Gatsonis C, Snyder B, Singh H, Miller CD, Entrikin DW, et al. CT angiography for safe discharge of patients with possible acute coronary syndromes. N Engl J Med 2012;366(15):1393-403.
18. Hoffmann U, Truong QA, Schoenfeld DA, Chou ET, Woodard PK, Nagurney JT, et al, ROMICAT-II Investigators. Coronary CT angiography versus standard evaluation in acute chest pain. N Engl J Med 2012;367(4):299-308.
19. Halpern EJ, Fischman D, Savage M, Koka AR, DeCaro M, Levin DC. Decision analytic model for evaluation of suspected coronary disease with stress testing and coronary CT angiography. Acad Radiol 2010;17(5):577-86.
20. McCollough C, Cody D, Edyvean S, et al. The measurement, reporting, and management of radiation dose in CT. College Park, MD: American Association of Physicists in Medicine; 2008 . AAPM report no. 96.
21. Diamond GA, Forrester JS. Analysis of probability as an aid in the clinical diagnosis of coronary-artery disease. N Engl J Med 1979;300(24):1350-8.
22. Manini AF, Dannemann N, Brown DF, Butler J, Bamberg F, Nagurney JT, et al. Rule- Out Myocardial Infarction using Coronary Artery Tomography (ROMICAT) Study Investigators. Limitations of risk score models in patients with acute chest pain. Am J Emerg Med 2009;27(1):43-8.
23. Ferencik M, Schlett CL, Bamberg F, Truong QA, Nichols JH, Pena AJ, et al. Comparison of traditional cardiovascular risk models and coronary atherosclerotic plaque as detected by computed tomography for prediction of acute coronary syndrome in patients with acute chest pain. Acad Emerg Med 2012;19(8):934-42.
24. D’Ascenzo F, Biondi-Zoccai G, Moretti C, Bollati M, Omede P, Sciuto F, et al. TIMI, GRACE and alternative risk scores in acute coronary syndromes: a meta-analysis of 40 derivation studies on 216,552 patients and of 42 validation studies on 31,625 patients. Contemp Clin Trials 2012;33(3):507-14.
25. Lee B, Chang AM, Matsuura AC, Marcoon S, Hollander JE. Comparison of cardiac risk scores in ED patients with potential acute coronary syndrome. Crit Pathw Cardiol 2011;10(2):64-8.
26. de Araujo Goncalves P, Ferreira J, Aguiar C, Seabra-Gomes R. TIMI, PURSUIT, and GRACE risk scores: sustained prognostic value and interaction with revascular- ization in NSTE-ACS. Eur Heart J 2005;26(9):865-72.
27. Backus BE, Six AJ, Kelder JH, Gibler WB, Moll FL, Doevendans PA. Risk scores for patients with chest pain: evaluation in the emergency department. Curr Cardiol Rev 2011;7(1):2-8.
28. Garcia S, Canoniero M, Peter A, de Marchena E, Ferreira A. Correlation of TIMI risk score with angiographic severity and extent of coronary artery disease in patient with non-ST-elevation acute coronary syndromes. Am J Cardiol 2004;93(7): 813-6.
29. Yan AT, Jong P, Yan RT, Tan M, Fitchett D, Chow CM, et al. Clinical trial-derived risk model may not generalize to real-world patients with acute coronary syndrome. Am Heart J 2004;148:1020-7.
30. Halpern EJ. Triple-rule-out CT, angiography for evaluation of acute chest pain and possible acute coronary syndrome. Radiology 2009;252(2):332-45.
31. Takakuwa KM, Halpern EJ, Shofer FS. A time and imaging cost analysis of low-risk ED observation patients: a conservative 64-section Computed tomography coronary angiography “triple rule-out” compared to nuclear stress test strategy. Am J Emerg Med 2011;29(2):187-95.
32. Redberg RF. Coronary CT, angiography for acute chest pain. N Engl J Med 2012;367(4):375-6.
33. 2012 Risk Management Review-Emergency Medicine. Physician Insurers Association of America (PIAA).