Author + information
- Received January 29, 2013
- Revision received April 27, 2013
- Accepted April 29, 2013
- Published online August 6, 2013.
- Michael Poon, MD∗,†,‡∗ (, )
- Michael Cortegiano, BA∗,
- Alexander J. Abramowicz, BS∗,
- Margaret Hines, BA∗,
- Adam J. Singer, MD†,
- Mark C. Henry, MD†,
- Peter Viccellio, MD†,
- Jeffrey C. Hellinger, MD∗,
- Summer Ferraro, BA∗,
- Annie Poon, BA∗,
- Gilbert L. Raff, MD§,
- Szilard Voros, MD∗,‡,
- Michael E. Farkouh, MD, MSc⋮ and
- Pamela Noack, PhD¶
- ∗Department of Radiology, State University of New York (SUNY) Stony Brook University, Stony Brook Medicine, Stony Brook, New York
- †Department of Emergency Medicine, SUNY Stony Brook University, Stony Brook Medicine, Stony Brook, New York
- ‡Division of Cardiovascular Medicine, SUNY Stony Brook University, Stony Brook Medicine, Stony Brook, New York
- §Division of Cardiology, William Beaumont Hospital, Royal Oak, Michigan
- ⋮Peter Munk Cardiac Centre and Heart and Stroke Richard Lewar Centre of Excellence, University of Toronto, Toronto, Ontario, Canada
- ¶Department of Health Policy and Management, Stony Brook University School of Health Technology and Management, Stony Brook, New York
- ↵∗Reprint requests and correspondence:
Dr. Michael Poon, Advanced Cardiovascular Imaging Program, Department of Radiology, Stony Brook Medicine, HSC Level 4, Room 120, 101 Nicolls Road, Stony Brook, New York 11794–8460.
Objectives This study was designed to assess the effects on resource utilization of routine coronary computed tomographic angiography (CCTA) in triaging chest pain patients in the emergency department (ED).
Background The routine use of CCTA for ED evaluation of chest pain is feasible and safe.
Methods We conducted a retrospective multivariate analysis of data from two risk-matched cohorts of 894 ED patients presenting with chest pain to assess the impact of CCTA versus standard evaluation on admissions rate, length of stay, major adverse cardiovascular event rates, recidivism rates, and downstream resource utilization.
Results The overall admission rate was lower with CCTA (14% vs. 40%; p < 0.001). Standard evaluation was associated with a 5.5-fold greater risk for admission (odds ratio [OR]: 5.53; p < 0.001). Expected ED length of stay with standard evaluation was about 1.6 times longer (OR: 1.55; p < 0.001). There were no differences in the rates of death and acute myocardial infarction within 30 days of the index visit between the two groups. The likelihood of returning to the ED within 30 days for recurrent chest pain was 5 times greater with standard evaluation (OR: 5.06; p = 0.022). Standard evaluation was associated with a 7-fold greater likelihood of invasive coronary angiography without revascularization (OR: 7.17; p < 0.001), while neither group was significantly more likely to receive revascularization (OR: 2.06; p = 0.193). The median radiation dose with CCTA was 5.88 mSv (n = 1039; confidence interval: 5.2 to 6.4).
Conclusions The routine use of CCTA in ED evaluation of chest pain reduces healthcare resource utilization.
- admission rate
- chest pain
- coronary computed tomographic angiography
- invasive resource utilization
- length of stay
Chest pain is the most common chief complaint in patients ages 65 years and older and second most common in patients ages 15 to 65 years treated in emergency departments (EDs) in the United States. It accounts for over 6 million ED visits and costs more than $10 billion/year (1,2). The standard evaluation for chest pain diagnosis varies from ED to ED and across regions. Recently, ED observation or chest pain units have been established across the country in an attempt to homogenize the standard evaluation for chest pain diagnosis, as they have been shown to improve the clinical outcomes in patients with possible acute coronary syndromes (ACS) (3). However, only about one third of U.S. EDs currently operate observation units capable of performing timely functional testing (4). Thus, chest pain remains a major contributing cause of ED crowding and unnecessary hospital admissions.
Recently coronary computed tomographic angiography (CCTA) has emerged as a useful tool for ruling out the presence of significant obstructive coronary artery disease (CAD) in patients with stable symptoms (5). Additionally, in several small-scale, single-center studies and in three multicenter trials, CCTA was shown to be a safe and cost-efficient method of acute chest pain evaluation compared with the commonly used functional testing (6–9). However, these studies are limited in that CCTA availability was constrained to weekdays and office hours, whereas EDs are 24/7 operations. A recent prospective study showed that the implementation of a CCTA program that runs 12 h/day on weekdays only can safely discharge patients with negative or nonobstructive CCTA findings using a triage protocol (10). To date, there have been no studies evaluating effectiveness of daily availability of CCTA for at least 12 h/day continuously for chest pain triage in a busy ED. Such evaluations must be performed to assess the effectiveness of introducing a new diagnostic tool in clinical practice.
We have previously reported that routine CCTA in the evaluation of low- to intermediate-risk chest pain in a tertiary care, university-based, suburban, academic medical center ED was feasible and safe (11). Our current observational study sought to compare the overall impact on clinical outcomes and efficacy between CCTA and our local standard evaluation for the triage of chest pain in our ED when CCTA was available 12 h/day, 7 days/week.
We employed a retrospective, observational design, with all data abstracted from the hospital discharge and follow-up records. The ED, with an annual census of approximately 90,000, used the chest pain triage algorithm shown in Figure 1 before and after January 1, 2009, when routine CCTA was introduced. Following the initial clinical assessment, all chest pain patients with an admitting diagnosis of cardiac chest pain underwent an initial 12-lead electrocardiography (ECG) and blood draw for cardiac troponin I level. Patients without ST-segment elevation, ST-segment depression of ≥1 mm, or positive cardiac troponin I (>0.04 pg/ml) were assessed in the ED by standard evaluation before CCTA introduction and by either standard evaluation or CCTA after introduction on the basis of each ED physician's discretion. The study design was approved by the institutional review board.
From January 1, 2008, to April 30, 2010, the ED treated 9,308 patients with admitting diagnoses of chest pain (International Classification of Diseases-Ninth Revision-Clinical Modification codes 786.5X). Patients were excluded if they presented with ACS (positive troponin I or ischemic changes on ECG) (n = 24) or noncardiac chest pain (no ECG or cardiac biomarkers obtained during the ED visit) (n = 2,220), had a known history of CAD (n = 1,772), or were discharged to home with a length of stay in the ED shorter than 3 h (considered too short a duration to rule out ACS) (n = 601). After exclusion, 4,691 total cases remained in the study cohort. We then matched propensity scores for CCTA and standard evaluation to develop a matched sample of 894 patients for the two comparative groups (Fig. 2).
The standard evaluation included ED cardiac monitoring, in which serial ECGs were obtained for the detection of ischemic ECG changes, and serial blood draws for troponin I level were taken for ruling out ACS. Patients were discharged after evaluation in the ED or were admitted to rule out ACS on a medical floor. Inpatient stress testing was performed if deemed necessary by the evaluating cardiologist. The discharge plan included a follow-up evaluation with cardiology and outpatient stress testing within 72 h strongly recommended.
Coronary computed tomographic angiography
On January 1, 2009, CCTA was introduced as a new, alternative option for ED evaluation of non-ACS cardiac chest pain. The attending ED physician chose between standard evaluation or CCTA, depending on CCTA availability and clinical suitability. CCTA was offered from 8 am to 8 pm daily, including weekends and holidays. Patients with an estimated glomerular filtration rate >50 ml/min/1.73 m2 and for whom iodinated contrast was not contraindicated were considered for CCTA. All CCTA patients needed to be able to cooperate and to have a body mass index (BMI) of <50 mg/m2. Unless contraindicated, all CCTA patients received 50 mg of oral metoprolol if the heart rate was >60 beats/min, and 100 mg if the heart rate was >70 beats/min, as long as the mean blood pressure was >70 mm Hg on evaluation in the ED. Patients with active asthma for whom beta-blockers were contraindicated were given intravenous diltiazem 20 mg (0.25 mg/kg) over 2 min, which was repeated, if needed, after 15 min. Unless contraindicated, all patients received 0.4 mg of sublingual nitroglycerin about 5 min before the contrast study. All CCTA patients were scanned using a dedicated 64-slice CT scanner (GEVCT, GE Healthcare, Milwaukee, Wisconsin) located in the ED. Patients with a BMI <30 kg/m2 were scanned with 100 kV; patients with a BMI between 30 and 50 kg/m2 were scanned with 120 kV.
Patients with heart rates ≤65 beats/min were scanned prospectively. For heart rates >65 beats/min but <100 beats/min, retrospective gating with dose modulation was used. No patients were excluded based on a high or low heart rate as long as the cardiac rhythm was normal sinus and the heart rate was <100 beats/min. The volume of contrast Ultravist 370 (Bayer HealthCare, Montville, New Jersey) used was 75 ml for single rule-out and 110 ml for triple rule-out (i.e., ruling out CAD, acute aortic dissection, and acute pulmonary embolism). All CCTA studies were evaluated within 1 h of the scan by a level III ACC/ACR certified imaging expert. Nonobstructive CCTA was defined as <50% maximal diameter stenosis using images from the best cardiac phase and multiplanar reconstruction post-processing technique in a transverse coronary section across the narrowest segment compared with the nearest normal lumen. Normal CCTA was defined as 0% stenosis and 0 calcium score. Obstructive CCTA was defined as ≥50% stenosis (Fig. 3). The total effective radiation dose per CCTA study was recorded for each patient. Effective radiation dose was derived from the summed dose-length product multiplied by the standard conversion factor (0.014 mSv/[mGy·cm]).
The primary study outcome was the hospital admission rate, defined as the percentage of patients admitted to the hospital. The secondary study outcome was length of stay in the ED for patients discharged to home from the ED and length of stay in the hospital for patients admitted. Length of stay was obtained by subtracting the official arrival time to the ED from the discharge time as indicated in the hospital discharge record. Other study endpoints included 1-month follow-up and major adverse cardiovascular events (MACEs), including death, acute myocardial infarction (AMI), and return for pulmonary embolism; ED and hospital recidivism rates; downstream functional testing and invasive diagnostic and interventional procedures, that is, invasive coronary angiography, and percutaneous and surgical revascularization. To assess the radiation safety of CCTA, we tabulated the total effective radiation dose of each CCTA study.
Our null hypothesis was that no difference exists between the overall healthcare resource utilization and outcomes in patients receiving CCTA as opposed to standard evaluation. We assessed the demographics and medical risk of our population using independent t tests and concluded that standard-evaluation patients had a slightly higher clinical risk (Table 1). To control for this difference, we used logistic regression analysis to calculate propensity scores for receiving CCTA using the following risk factors routinely evaluated during ED workup as binomial variables: hypertension, hyperlipidemia, diabetes, patient current/historical smoking status, sex, BMI, renal status, heart failure, number of cardiac risk factors, and age (categorized by quintile) (12). We then calculated the area under the receiver-operating characteristic (ROC) curve to assess the overall predictability of our model. We then matched propensity scores for CCTA and standard evaluation to develop a matched sample, using nearest-neighbor matching that controlled for potential differences in severity of observable cardiac disease between the two groups, using StataPScore (software written by S. Becker and A. Ichino) and PSMatch2 (software written by E. Leuven and B. Sianesi). We theorized that as the time elapsed after CCTA implementation increased, the likelihood that the ED medical and clinical staff would use CCTA would also increase due to increased experience, and that the staff's knowledge of CCTA might be influenced by their work-shift assignments. Because these influences are not directly observable but may be picked up by our variable of interest, we tested for their impact by using a biprobit model. Following our evaluation of the unobserved bias using a biprobit model, we concluded that although unobserved biases were present, the effects of using CCTA still existed when we controlled for these biases. See the Online Appendix for further discussion of the biprobit model.
We performed multivariate logistic regression to evaluate the likelihood of being admitted and receiving downstream invasive evaluation, adjusting for additional risk factors including subacute ischemia, ischemia, atherosclerosis, aneurysm, stroke, angina, circulatory disorders, peripheral vascular disease, congestive heart failure, insurance status, race, number of diagnoses documented, and the time of arrival in the ED. We assessed ED length of stay using the maximum likelihood survival analysis, inverse Gaussian frailty model. Stata version 10.1 (StataCorp, College Station, Texas) was used for all data analyses.
Overall study endpoints
Figure 2 presents the study algorithm, showing each decision point and outcomes in the ED evaluation process. Our propensity score model provided only moderate predictive power in distinguishing CCTA and standard evaluation (ROC: 0.67). The results of the study endpoints using an independent t test comparison on our matched sample of standard evaluation and CCTA are presented in Table 2. Additionally, multivariate logistic analysis results, used to make further risk adjustments, are presented in Table 3.
Primary study endpoint
Overall Hospital Admission Rates
Admission rates for standard evaluation and CCTA were 40% and 14%, respectively (p < 0.0001). In our logistic analysis, the likelihood of being admitted was significantly greater in standard-evaluation patients and increased over time (all years, OR: 5.53 [95% CI: 3.8 to 8.0; p < 0.001]; for 2009, OR: 3.98 [95% CI: 2.5 to 6.3; p < 0.001]; for 2010, OR: 9.3 [95% CI: 4.7 to 18.5; p < 0.001]).
Length of Stay
We examined length of stay in patients discharged from the ED with a primary diagnosis of chest pain. In evaluating ED length of stay, we began with examining the change for the entire population before and after implementation of CCTA. The length of stay in the ED fell over the course of our study, from January 2008 to April 2010. However, the length of stay in patients arriving between 8 am and 12 pm was not significantly shorter (OR: 1.29; 95% CI: 0.79 to 2.10; p = 0.298). CCTA had the most significant impact on length of stay during the time period when the ED operated at peak volume and CCTA was open (from 12 pm to 8 am daily). The expected length of stay in the ED was 1.6 times higher with standard evaluation (OR: 1.55; 95% CI: 1.2 to 2.04; p < 0.001) (Table 3). Of the matched CCTA group, only 185 CCTA patients were evaluated when CCTA was closed, that is, from 8 pm to 8 am. The mean length of ED stay in the CCTA group was significantly shorter than in the standard-evaluation group (7.7 vs. 11.5 h). Figure 4 shows the cumulative distribution of patients remaining with length of stay data in the 2 groups of 432 cases each. Seventy-five percent of CCTA patients were expected to be discharged within 8 h compared with 14 h with standard evaluation.
Death Within 30 Days
We examined our medical center records and the Social Security Death Master File. There were no cardiac deaths in either group.
AMI Within 30 Days
We then evaluated the likelihood of experiencing an AMI from the time of the initial ED visit to within 30 days of the index visit. Six patients (1%) in the standard-evaluation group and 3 patients (<1%) in the CCTA group had AMI (p = 0.316). All AMIs occurred during the index visit. Of the 3 patients who had CCTA and AMI, all had AMI on admission. One patient had obstructive CCTA and the other 2 were nondiagnostic. Both patients with nondiagnostic CCTA had normal levels of initial cardiac biomarkers, and one was post intervention. Similarly, all 6 standard-evaluation cases had AMI during the index visit. Five had elevated levels of second or third cardiac biomarkers. One had a positive level post intervention. The risk for experiencing AMI was not significantly different between the standard-evaluation and CCTA cohorts (OR: 4.26; 95% CI: 0.3 to 71.4; p = 0.313).
ED or Hospital Recidivism Rate
We reviewed the number of patients who returned to the hospital within 30 days of the index admission: 3.6% of standard-evaluation patients returned compared with 1.3% of CCTA patients (p = 0.002). However, the likelihood of returning after complete risk adjustment was not significantly different between the two groups (OR: 8.53; 95% CI: 0.4 to 179.9; p = 0.168). We also explored reasons for return, including admission for pulmonary embolism, cardiac reasons (ICD-9 codes 390.∗∗ to 459.∗∗) excluding chest pain, and chest pain alone (Table 2). We found no significant differences in the likelihood of returning for cardiac reasons (OR: 2.30; p = 0.178) or for pulmonary embolism (OR: 9.23; p = 0.068). However, for chest pain alone, the likelihood of returning was significantly greater in the standard-evaluation patients (OR: 5.066; p = 0.022).
Downstream Resource Utilization of Diagnostic Procedures and Interventions
When patients were discharged from the ED after ACS rule-out, they were instructed to contact a cardiologist for possible stress testing within 72 h. We could validate only that 754 patients (21%) in the full standard-evaluation cohort received stress testing. Among admitted inpatients, 36.9% of standard-evaluation patients and 24.8% of CCTA patients had stress testing. Among discharged outpatients, 9.9% of standard-evaluation patients and 0.3% of CCTA patients had stress testing. The majority of the stress testing performed involved single-photon emission computed tomography, with which the total radiation dose for our typical 1-day stress–rest protocol was 12 mSv (3.5 mSv for rest and 8.5 mSv stress). Stress tests following CCTA were mainly used to determine the functional significance of nondiagnostic or obstructive CCTA results.
Invasive Coronary Angiography
Standard-evaluation patients were significantly more likely to receive invasive coronary angiography without subsequent revascularization (27 [3%] vs. 8 [1%]; p < 0.001). When risk-adjusted, the likelihood of standard-evaluation patients having invasive coronary angiography without subsequent revascularization was 7 times higher (OR: 7.17; 95% CI: 2.5 to 20.6; p < 0.001; Tables 2 and 3).
Coronary Revascularization (Percutaneous Coronary Intervention or Coronary Artery Bypass Grafting)
The difference in revascularization rates between CCTA and standard evaluation was not significant (23 [3%] vs. 19 [2%], respectively; p = 0.533]. The difference in the likelihood of standard-evaluation and CCTA patients undergoing coronary revascularization was not significant (OR: 2.06; 95% CI: 0.7 to 6.11; p = 0.193). No patients who had normal or nonobstructive CCTAs received invasive coronary angiography or coronary intervention.
CCTA-Associated Radiation Dose
After excluding patients who received additional noncardiac CT scanning on the same occasion of service and those with missing radiation-dose data (n = 22 [2%]), 1,039 patients remained. Nearly 32% of the CCTAs done were for triple rule-out. These cases typically carry exposure rates nearly 50% higher than those of cases ruling out CAD only (13). Our median radiation dose was 5.88 mSv (95% CI: 5.2 to 6.4)—16.22 mSv (95% CI: 15.0 to 17.4) for retrospective scans (n = 432 [42%]) and 3.61 mSv (95% CI: 3.4 to 3.8) for prospective scans (n = 605 [58%]).
We believe this is the largest single-center observational study to date comparing the routine use of CCTA to a standard-evaluation approach. Many EDs in the United States follow similar protocol for standard evaluation, in which most chest pain patients are discharged after serial sets of cardiac biomarkers and nondiagnostic ECGs (14).
The results of the primary endpoints on hospital admission rate and length of stay in the ED support those from the three completed randomized controlled trials to date: CCTA reduced unnecessary hospital admissions and ED length of stay (6,8,9). The current study adds to the literature as its retrospective, observational design begins to validate that these savings are possible in clinical practice. It addresses potential investigational bias resulting from experimental controls that cannot exist in an ED environment and may, therefore, render accurate efficiency analysis impossible. For example, the published randomized controlled trials used CCTA during business hours and weekdays only. This is not a reasonable expectation for a busy ED. Throughout our study, CCTA was continuously available for 12 h/day. As our implementation of CCTA progressed, ED physicians and clinical staff became more comfortable with, and aware of the advantages of, using CCTA, which resulted in a significant improvement in the OR for the overall hospital admission rates, from 3.98 in 2009 to 9.2 in 2010. The length-of-stay analysis was not biased by research protocols with dedicated personnel to conduct the study, yielding a truer assessment of how CCTA affects length of stay in clinical practice.
Many EDs in the United States consistently experience crowding that challenges providers' capability of efficiently caring for patients. A recent study documented that this crowding causes inefficiencies and can result in higher death rates (15). Additionally, it has been shown that ED crowding may be associated with post-traumatic stress disorder in patients who presented to the ED with ACS (16). The average length of stay in the ED, when inpatient treatment was not indicated, was 8.5 h. The potential savings of 2 to 3 h per patient through the use of CCTA was associated with no reduction in the quality of outcomes. This work is important because reducing extended ED stays contributes to improvement in overall quality and cost efficiency.
In our cohort of low-risk patients, we could document follow-up stress testing within 1 month in only 21% of standard-evaluation patients. Some patients may have left our network for follow-up, and that information may not have been available to us. This low stress-testing rate in the standard-evaluation patients was not unexpected, because only about one-third of EDs are equipped to perform stress testing, and significantly fewer are capable of timely management of the large volume of chest pain patients encountered each day (4). Accordingly, many patients are discharged without provocative testing or remain in the ED or hospital for observation for 10 h or more. A scientific statement from the American Heart Association on the testing of low-risk patients with chest pain in the ED recognizes the value of outpatient exercise stress testing within 72 h of the ED visit, as many hospitals lack the resources to properly complete inpatient stress testing (17). However, a recent study found that only 36% of patients referred to outpatient stress tests actually completed them, increasing patients' risk for subsequent ED visits and the potential for missed ACS (18). Because many EDs are limited in the ability to provide long-term follow-up of low-risk ACS patients, the longer-term risk to these patients is not fully understood and warrants further research.
When patients are admitted for chest pain, one of the most resource-intensive diagnostic procedures performed is invasive coronary angiography (19). As CCTA provides accurate diagnostic information concerning coronary obstruction, patients in the standard-evaluation cohort were 7 times more likely to have an invasive coronary angiography without subsequent revascularization. This is in sharp contrast to the findings in the recently published SPARC (Study of Myocardial Perfusion and Coronary Anatomy Imaging Roles in Coronary Artery Disease), in which invasive coronary angiography was more frequent after CCTA (20). This difference in invasive coronary angiography use was most likely due to differences in patient selection and use of stress testing. The current study consisted of mostly low-risk chest pain patients, whereas the SPARC patients were of intermediate to high risk. Additionally, because many hospitals across the country are not equipped to perform routine stress testing in the ED, physicians may rely on invasive coronary angiography as a primary diagnostic test in higher-risk patients. Our results demonstrate that significantly more potentially avoidable invasive coronary angiography procedures are performed after standard evaluation compared with CCTA. This results in increased risks related to invasive procedures and radiation exposure (21). The recently published ROMICAT-II (Rule Out Myocardial Infarction using Computer Assisted Tomography) trial (7) showed a higher rate of downstream invasive testing after CCTA compared with standard evaluation (11% vs. 7%, respectively). However, the standard-evaluation group in that trial had a functional testing rate of 78%—a rate that is not achievable in most U.S. EDs; therefore, the results for downstream testing rates in the studies are not comparable.
Controversy remains as to whether any testing is warranted in such low-risk patients following negative initial ECG and cardiac biomarkers because both CCTA and nuclear stress testing increase the risk for cancer due to radiation exposure. Our study in low-risk patients contained a total combined revascularization and AMI rate of 2.4% (95% CI: 1.6% to 3.7%). Applying this rate of risk for adverse incident to the general population of 6 million patients per year implies that doing nothing would place 144,000 patients annually at risk for inadequate care with serious outcomes. At an average age of 49 years, these patients would experience permanent detriment to quality of life and economic production. Yet, caring for emergent, low-risk acute chest pain patients represents a high cost to the system resulting from patient volume and diagnostic challenges. We estimated that ED chest pain patients receiving standard evaluation were 5.5 times more likely to be admitted. We have presented substantial evidence that CCTA improves the efficiency of triaging chest pain patients in the ED.
CCTA has been criticized for placing patients at increased risk for radiation exposure. In the ROMICAT-II trial (7), cumulative exposures were 13.9 mSv with CCTA and 4.7 mSv with standard evaluation. In the CT-STAT (Coronary Computed Tomographic Angiography for Systematic Triage of Acute Chest Pain Patients to Treatment) trial (6), the reported median radiation dose in the arm that received CCTA with 64-slice CT scanner technology was 12.8 mSv. Our patients' exposure levels were substantially lower (22,23). The median radiation exposure was 5.89 mSv, despite a 32% triple rule-out rate. Using our current pre-medication protocol of administering an oral beta-blocker 1 h prior to scanning, 58% of the patients achieved heart rate control adequate for prospective scanning. The mean effective radiation dose exposure with this dose-saving scanning technique was 3.62 mSv.
Although a retrospective and observational study design has limitations, it allows a fuller assessment of the in-practice value of introducing a new diagnostic modality in a busy clinical environment, now that safety has been confirmed. Observational studies always carry risk related to selection bias; however, we found our analytic results to be consistent and rigorous to many variations in methodology. The lack of availability of CCTA from 8 pm and 8 am was a limitation of the CCTA arm, but the incremental cost to perform CCTA around the clock is extremely high and probably not medically necessary, as all CCTA patients had negative initial cardiac biomarkers and ECG for ACS and most of these patients had normal or nonobstructive CCTA. Patients who came to the ED when CCTA was not available were either kept in the ED or admitted to the medical floor for further observation. In addition, our propensity model did not demonstrate a strong relationship between key medical history/conditions and the use of CCTA as a diagnostic tool (ROC: 0.67). We believe this was a result of the availability of CCTA over the course of the day and of the gradual increase in the adoption of CCTA by our staff. Our local standard evaluation lacks adequate resources to perform timely ED stress testing, which partially drives the superior results of our CCTA program. Nonetheless, our standard-evaluation practice represents the norm for many U.S. EDs that lack resources to perform timely stress testing. Additional evaluations will be necessary to confirm these results; however, as knowledge and capabilities of the use of CCTA progress, this technology may offer an alternative means of improving services for chest pain patients.
Implementation of a protocol for ruling out ACS in low-risk chest pain patients using CCTA will likely increase emergency physicians' ability to accurately and efficiently triage patients with this common presenting symptom. This may result in a reduced need for inpatient admissions, ED length of stay, chest pain recidivism rate, and downstream evaluation by invasive coronary angiography, and may enhance treatment efficacy.
The authors thank Elizabeth Vanner, PhD, and Debra Dwyer, PhD, for review and suggestion on statistical analysis.
Dr. Raff has received a research grant from Siemens. Dr. Voros has relationships with Global Genomics Group, LLC, Global Institute for Research, LLC, HDL, Inc., Innovations in Integrated Imaging, LLC, Integrated Cardiovascular Research Group, LLC, Merck & Co., Inc, Toshiba, and Vital Images; and has received financial support from Global Genomics Group, LLC, Global Institute for Research, LLC, HDL, Inc., Innovations in Integrated Imaging, LLC, Integrated Cardiovascular Research Group, LLC, Toshiba, and Vital Images. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- acute coronary syndrome(s)
- acute myocardial infarction
- coronary artery disease
- coronary computed tomographic angiography
- emergency department
- major adverse cardiovascular event(s)
- Received January 29, 2013.
- Revision received April 27, 2013.
- Accepted April 29, 2013.
- American College of Cardiology Foundation
- Niska R.,
- Bhuiya F.,
- Xu J.
- Priest V.L.,
- Scuffham P.A.,
- Hachamovitch R.,
- Marwick T.H.
- Goldstein J.A.,
- Chinnaiyan K.M.,
- Abidov A.,
- et al.
- Hoffmann U.,
- Bamberg F.,
- Chae C.U.,
- et al.
- Takakuwa K.M.,
- Halpern E.J.,
- Gingold E.L.,
- Levin D.C.,
- Shofer F.S.
- Amsterdam E.A.,
- Kirk J.D.,
- Bluemke D.A.,
- et al.
- Taylor A.J.,
- Cerqueira M.,
- Hodgson J.M.,
- et al.
- Winchester D.E.,
- Jois P.,
- Kraft S.M.,
- Wymer D.C.,
- Hill J.A.
- Chow B.J.,
- Abraham A.,
- Wells G.A.,
- et al.
- Hachamovitch R.,
- Nutter B.,
- Hlatky M.A.,
- et al.
- Tavakol M.,
- Ashraf S.,
- Brener S.J.
- Hausleiter J.,
- Martinoff S.,
- Hadamitzky M.,
- et al.