Author: Andrew Berg, MD (EM Resident Physician, PGY-1, NUEM) // Edited by: Victoria Weston, MD (EM Resident Physician, PGY-4, NUEM) // Expert Reviewer: Mike Gisondi, MD
Citation: [Peer-Reviewed, Web Publication] Berg A, Weston V (2016, April 12). Journal Club: Coronary CT Angiography Versus Traditional Care. [NUEM Blog. Expert Peer Review by Gisondi M]. Retrieved from http://www.nuemblog.com/blog/cta-for-chest-pain/
Introduction
Chest pain is the second most common reason for presenting to the emergency department, and accounts for over 8 million annual visits in the US [1]. It is estimated that admissions for chest pain cost over $3 billion annually, with only 10 to 15% of patients being ultimately diagnosed with acute coronary syndrome (ACS). Low to moderate risk chest pain patients present an especially complicated problem because not only do they make up a larger portion of chest pain presentations, but they have significantly lower pre-test probabilities which increases the rates of false positives on screening tests.
The current American College of Cardiology/American Heart Association guidelines recommend that exercise treadmill testing should be performed as the initial test in low- to intermediate-risk patients with ischemic type chest pain. Although the exercise treadmill test itself is often inexpensive, it has poor positive predictive value (in this population ranging from 8-57% [1]). This has the potential to lead to increased hospital admissions and unnecessary invasive testing.
Coronary CT Angiography (CCTA) has shown promise in three major randomized controlled trials evaluating its safety in risk-stratifying low- to intermediate-risk patients [2,3,4]. Additionally, CCTA-based screening was shown to increase rate of discharge directly from the ED, and decreased length of stay. However, these studies have only followed patients for up to 30 days after discharge from the emergency department. The study discussed in this Journal Club Review is a continuation study from one of these three studies in which it tracks the outcomes and resource utilization of these patients for one year from discharge. It contributes to the growing body of evidence that CCTA may soon have a more established role in screening patients with lower risk chest pain.
Study
Hollander JE, Gatsonis C, Greco EM, et al. Coronary Computed Tomography Angiography Versus Traditional Care: Comparison of One-Year Outcomes and Resource Use. Ann Emerg Med. 2015 http://www.ncbi.nlm.nih.gov/pubmed/26507904
Study Design
Population
The initial study was conducted at 5 separate hospitals in Pennsylvania during the time period from July 7, 2009 through November 3, 2011. They enrolled a total of 1392 patients, of which 1370 were randomized.
Patient Selection
In order to be randomized for the study, the patient had to meet the following 4 major inclusion criteria:
- Age >30
- Signs and symptoms that were consistent with ACS as to be determined by the study physician
- The ECG at presentation could not show evidence of acute ischemia
- TIMI risk score of 0 to 2
Additionally, patients were excluded for the following reasons:
- Had a co-existing condition that necessitated admission regardless of the ACS outcome
- Had normal findings on invasive angiography or CCTA in the prior year.
- Had any contraindications to CCTA (i.e. renal dysfunction or contrast allergy).
Interventional Protocol
The patients were randomization to either the CCTA treatment arm or the traditional care treatment arm.
- The CCTA treatment arm: Patients received CCTA per hospital protocol, which included the use of beta-blockers and nitroglycerin during image acquisition. The cut off for clinically significant coronary artery disease was defined as stenosis of 50% in any of the major vessels.
- The Traditional Care treatment arm: Patients either underwent graded exercise stress testing or pharmacologic stress testing.
Both groups of patients were contacted 30 days and again at 1 year after the ED visit. Additionally they conducted extensive medical record review of all hospitalizations and outpatient visits. If any patient could not be reached, they searched the Social Security Death Master File for vital status.
Outcome Measures
The primary outcome measure was safety. Given the infrequency of events, they could not sufficiently power the study to directly compare CCTA against the traditional care arm. The authors instead hypothesized that CCTA would have less than a 1% rate of adverse cardiac events over the one-year period.
The secondary outcome measure was related to resource utilization. They compared repeat ED visits, admissions, cardiologist visits, repeat testing, and medication use between the two groups.
Results
Interpretation
The major outcome of the study is that they were able to demonstrate that patients with negative CCTA screening had a 1-year event rate of less than 1%. Although the study was underpowered to directly compare CCTA vs. traditional care directly, there were no statistical differences in events between the two groups
The authors hypothesized that a negative CCTA screening would lead to overall decrease in resource utilization when compared to traditional care. The rationale behind this was that prior studies showed that negative stress tests did not significantly reduce further utilization, and that proving a patient had no coronary artery disease on CCTA should cut down on return visits for chest pain. However, the data showed no statistical differences between the two groups in ED visits, cardiologist office visits, hospital admissions, other testing or medication use.
Strengths
This study suggests that a negative CCTA predicts a less than 1% rate of serious cardiac events or death within 1-year. The randomization of such a large sample of patients with very few excluded patients or those lost to follow up, supports the internal validity of the study. Additionally, the outcome measures for safety that were selected (mortality and myocardial infarctions) were clinically relevant to an ED physician assessing a patient with ischemic type chest pain.
Weakness
A major limitation is that this study was inadequately powered to directly compare the two treatment arms in terms of safety and cost. Although there appeared to be no statistical difference between the two in either category, the sample size would not be adequate to demonstrate small differences between the groups. Although some clinicians are using CCTA as an alternative risk-stratification method, others feel that additional studies are needed to validate the test’s safety and effectiveness.
Although the trial was multi-centered with a large sampling of patients, all of the subjects were recruited within Pennsylvania. This limits the external validity of the study as it applies to different regions of the country.
Take Home Points
- This was a multi-center, randomized control trial comparing CCTA vs traditional care in low to moderate risk patients, assessing safety and resource utilization after one year
- The study randomized 1370 patients from five hospitals in Pennsylvania
- A negative CCTA screening predicted less than a 1% risk of serious cardiac events or death from cardiac causes
- Given the extremely low event rates in this population, the study was insufficiently powered to make a direct comparison of safety between the CCTA and Traditional Care treatment arms.
- Overall resource utilization (re-admissions, testing, medication use and catheterizations) between the two groups were not statistically different.
Expert Review
Dear Andrew,
Thanks for selecting an interesting topic and article to review this week. Great job in your summary, too. You raise some important points about the study design and our ability to extrapolate its findings to our patient population. There are several important points to discuss.
Let’s consider first, the patients: those with “low risk chest pain.” For my practice, that would include relatively younger patients (<60? <50?), without a history of coronary artery disease, with few or no major cardiac risk factors, chest pain that was not provoked by exercise and that resolved, a non-ischemic ECG, and a normal troponin-I. For the purposes of the Hollander et al study, ‘low risk’ must reflect a TIMI score between 0 and 2. [TIMI reminders: Remember that two episodes of chest pain buys your patient 1 TIMI point, as does aspirin use in the last week. Click here for a TIMI refresher.] We see these ‘low risk patients’ numerous times per shift in the emergency department and we generally admit the majority of them.
As for your concern about generalizing a population of patients in Philadelphia to a population of patients in Chicago…. I realize that you are from California and you have not yet become accustomed to the wonderful food of both cities.
Two pieces of Chicago deep dish pizza without toppings is 940 calories and 52g of fat. For comparison, a Philly cheesesteak is 650 calories and 21.5g of fat. In my opinion, the patient population cared for in the Hollander study is likely to be sufficiently similar to ours.
Next, consider your testing options. Yes, you have categorized these patients as ‘low risk’ (for coronary artery disease as the cause of their symptoms, to be clear – remember your other life threats!), but you are still sufficiently concerned enough that you want additional testing beyond just the ECG and single cardiac enzyme. What are your options? Ideally, we would have Tricorders that we could wave above a patient’s chest that would give us the answer. Alas, those do not yet exist. [But maybe soon?] In the real world, you could:
- (A) Obtain serial negative ECGs and enzymes and discharge the patient for an outpatient stress in the next couple of days
- (B) Admit the patient to an observation or clinical decision unit for serial ECGs, cardiac enzymes, and stress testing
- (C) Obtain a coronary CTA from the ED after a single negative ECG and troponin-I
Pros
Option C, the coronary CTA, is certainly a desirable option. The ‘pro’s’ to using CCTA are compelling. If normal, we can convincingly exclude CAD as the etiology of the patient’s chest pain. I like tests that convincingly exclude diseases, and in general practice most tests that we use on a daily basis can’t do that. The Hollander study suggests that patients with mild CAD (defined as < 50% stenosis in their methods) have very low rates of cardiac events at 1 year. That’s not the same as a ‘normal study convincingly excluding CAD,’ but it is still quite compelling. Add to the ‘pro’ column that CCTA helps to avoid an admission and likely reduces further cardiac testing (or, perhaps, entirely eliminates the need for such testing for many years to come.)
I’m sure that you have cared for a patient with chest pain who had a negative treadmill stress test in the previous several months… and you found yourself saying something to the effect of:
“Treadmill tests aren’t perfect, I think you should be admitted today for [fill in the blank… stress echo, nuclear stress, etc.]”
In sequence, the costs of such tests (and the necessary hospitalizations to do the tests) can quickly build up. A single negative CCTA may prevent the need for readmission and further testing. How about this: Have you seen a patient present with an acute MI just months after a normal treadmill stress test? How about a patient who suffered an acute MI in the parking lot immediately after a normal treadmill test in the hospital just hours before? (I have – twice.) So to my earlier point, the potential to convincingly exclude CAD is really enticing and makes CCTA that much more interesting to consider.
Cons
OK, the ‘cons’ of CCTA: Expense (yes it is pricey, but the debate is on as to whether CCTA lowers overall utilization and cost in populations of low risk chest pain patients); Radiation (CCTA is the equivalent of 600 chest radiographs, or 1 CT abdomen/pelvis, or the amount of radiation the average American is exposed to in the environment in about 5 years); and Availability (CCTA is usually limited to M-F, 9a-5p… due to the availability of radiology and cardiology consultants to interpret the exams… at institutions that offer CCTA from the ED at all.)
OK, last a little statistics. CCTA has high negative predictive value (see my comments about ‘convincingly exclude CAD’ above), but also a false positive rate to consider. The more we test, the more we will find patients with some degree of stenosis and find ourselves scratching our heads about causality. This is, however, the practice and the art of medicine. There are probably many other examples more common to your current practice in which similar issues of causality exist: how clinically significant is the small, subsegmental PE? What about the small aneurysm found on a CTA neck when you were looking for dissection? Is that aneurysm really the cause of the patient’s headache and neck pain? (Likely not.)
Therefore, you need to be thoughtful about how you will interpret CCTA results. I have a discussion with patients that goes something like this:
“I am looking for a yes/no answer with this test. If you have no evidence of plaque build-up and narrowing, you are going home with both of us reassured. Unfortunately though, some patients may have some narrowing of their vessels that creates an indeterminate result. If that is the finding on your test, I will want the input of a cardiologist and you may need to be admitted to the hospital for further testing.”
It’s a reasonable discussion and one to have with the patient before the test, so everyone understands how the result will dictate the plan.
Also, you raise this issue of the Hollander study having inadequate power to detect a significant difference between the two testing arms. True – but while the jury is still out to some degree, I think there is a growing body of evidence that CCTA is at least as good as traditional treatment algorithms (non-inferiority.) Interestingly, the Hollander study did not find a difference in overall resource utilization – one might infer that this finding occurred because of inadequate power too, as other studies have shown CCTA to reduce downstream testing and costs.
Further Discussion
We should also consider the HEART score in this discussion. Like TIMI, the HEART score is a prognostic tool that emergency providers can use to risk stratify patients with chest pain. There are some recent articles that suggest that HEART may in fact be superior to TIMI in the identification of patients in the ‘low risk’ category. A discussion of HEART vs TIMI is outside the scope of this blog post; however, if the standard moves away from the use of TIMI scores in favor of HEART scores, as it likely has, then it would be logical to use HEART in your selection of patients for CCTA. That being said, the Hollander paper uses TIMI scores, so keep that in mind as you consider how this particular paper applies to your current practice.
To conclude, I think CCTA offers us an important testing option when we care for patients who present with resolved chest pain that have an overall low risk for CAD. Patient selection is paramount – you want patients with a low pre-test probability for disease because these are the patients who are most likely to have a normal test and a discharge home from your ED. A TIMI score less than 2 will help you select those who are ‘low risk,’ but also consider if your patient is allergic to contrast dye and if they have normal renal function. Also, consider how the treatment algorithm changes if you send a D-Dimer that comes back positive… you will then likely choose to exclude PE with a CTA chest and that will use up your one angiogram of the day for this patient; their ‘r/o ACS’ will follow a traditional inpatient algorithm from there.
What do I do in my practice?
I like CCTA and I use it for all patients who meet the aforementioned criteria. The debate around this testing option is hot right now though, so I will be interested to see what our readers’ comments are on this topic. Thanks for this opportunity to comment on your post.
Mike Gisondi, MD, FACEP, FAAEM
Associate Professor of Emergency Medicine and Medical Education; Residency Program Director; Medical Education Scholarship Fellowship Director; Feinberg Academy of Medical Educators Director
Department on Emergency Medicine and Education, Northwestern University, Feinberg School of Medicine
References
- Amsterdam EA, Kirk JD, Bluemke DA, et al. 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.
- Litt HI, Gatsonis C, Snyder B, et al. CT angiography for safe discharge of patients with possible acute coronary syndromes. N Engl J Med. 2012;366(15):1393-403.
- Goldstein JA, Chinnaiyan KM, Abidov A, et al. The CT-STAT (Coronary Computed Tomographic Angiography for Systematic Triage of Acute Chest Pain Patients to Treatment) trial. J Am Coll Cardiol. 2011;58(14):1414-22.
- Hoffmann U, Truong QA, Schoenfeld DA, et al. Coronary CT angiography versus standard evaluation in acute chest pain. N Engl J Med. 2012;367(4):299-308.
- Hollander JE, Gatsonis C, Greco EM, et al. Coronary Computed Tomography Angiography Versus Traditional Care: Comparison of One-Year Outcomes and Resource Use. Ann Emerg Med. 2015