If Maslow were to generate a hierarchy of needs to assess the utility of the coronary CT angiogram (CCTA) in the Emergency Department for the evaluation of chest pain it would look something like this:
We have fairly robust data demonstrating the use of CCTA does not improve mortality, nor does it decrease the rate of missed MIs in patients evaluated in the Emergency Department 1,2,3,4,5. It is also clear the use of CCTA is associated with an increase in the number of patients exposed to downstream invasive tests and medical therapies6. A trial published in the April 2016 issue of Annals of Emergency Medicine by Hollander et al found that the use of CCTA did not reduce the rate Emergency Department revisits (36% in the CCTA group vs 38% in the standard care group)7.
While CCTA has consistently proven itself to be generally unhelpful in the management of patients presenting to the ED with chest pain, true believers argue that CCTA allows for a more focused cohort of patients who have CAD. These refinements in patient selection will lead to improvement in long-term cardiovascular outcomes, not readily apparent when using the short-term outcomes utilized by the studies in question. With the positive results reported in recently published SCOT-HEART trial’s long term follow up paper, it would initially appear that proponents of the CCTA were justified in their support of this diagnostic approach. But a closer examination may cast a less favorable light on these results.
Published in the NEJM, Newby et al randomized 4146 patients with stable chest pain, referred for cardiology evaluation to either a standard work up, or a standard work up with the addition of CCTA8. The initial publication examining the short-term outcomes of the SCOT-HEART cohort were examined in detail in a previous post, but in short, patients randomized to CCTA were more often given the diagnosis of CAD and were more likely to be treated with medical therapies and invasive procedures than the patients in the standard care group9. Despite this, the rate of cardiovascular death and myocardial infarction during the follow up period (1.7 years) was 1.3 vs 2.0, a 0.7% non-statistical difference. The overall mortality was 0.8% vs 1.0%, respectively.
The secondary analysis by Newby et al intended to evaluate the downstream consequences of a CCTA strategy. The authors continued to follow the original cohort of patients after enrollment for a median of 4.8 years (3-7 years). Similar to their initial findings, during follow-up, patients assigned to CCTA more frequently initiated preventive therapies (19.4% vs. 14.7%) and anti-anginal therapies (13.2% vs. 10.7%). While the CCTA group had higher rates of invasive coronary angiograpy in the initial paper, this difference was not observed over a longer follow-up period (23.6% in the CTA group and 24.2% in the standard-care group). Nor did the authors observe a difference in need for emergent revascularization, (13.5% in the CTA group and 12.9% in the standard-care group).
The authors reported a statistically significant decrease in the rate of their primary long-term outcome, death from coronary heart disease or nonfatal myocardial infarction (2.3% in the CTA group vs. 3.9% in the standard-care group; P=0.004). This difference was driven completely by a difference in the number of non-fatal myocardial infarctions observed in the respective groups (2.1% in the CCTA group and 3.5% in the standard care group). There was no difference in either cardiovascular mortality (0.2% vs 0.6%), or overall mortality (2.1% in both groups). And while many have argued that these results support the incorporation of CCTA into the workup of chest pain patients, I think they reveal more about our dependence on medical technology and desire for clinical certainty.
It is important to note that the trial authors were not blinded to group allocation or to the results of the CCTA. In fact, we know from the previous publication that the results of the CCTA led to higher rate of perceived diagnostic certainty and comfort in the clinicians treating the patients. It also increased the number of patients diagnosed with CAD (23% in the CCTA arm vs 11% in the standard care group). Keep in mind, there was no formal event adjudication, and all endpoints were classified primarily on the basis of diagnostic codes. This flaw in the trial’s design produces results that are at significant risk for diagnostic review bias. Diagnostic review bias is a source of confounding that occurs when those who are responsible for the determination of endpoints are aware of the results of the index test10. In this case, the diagnosis of myocardial infarction was made by the treating clinicians. It is not hard to imagine a scenario where you have two identical patients, both presenting with an elevation in troponin. If the patient already underwent a CCTA which found no evidence of CAD, then it is much more likely for the treating clinician to classify the troponin elevation to be due to an alternative process. Whereas if the patient had not received a CCTA, the very same troponin elevation will likely be classified as a MI. The decreased rate of non-fatal myocardial infarctions may very well have nothing to do with CCTA’s ability to focus our therapeutic tools on the patients who stand to benefit, but rather due to a diagnostic shell game.
The sense of diagnostic certainty provided by CCTA is alluring. But study after study demonstrates that these radiographic reassurances fail to offer clinical benefits to patients. Even the small decrease in the rate of MI observed in the follow up data of SCOT-HEART is very likely to be a result of confounding rather than true clinical benefit. In fact, the only benefit consistently demonstrated by CCTA is the diagnostic comfort it provides to the treating clinician. At some point, we will have to be completely honest and accept that our endorsement of a CCTA-guided diagnostic strategy has very little to do with our patients and far more to do with ourselves.
Sources Cited:
- EM Nerd-The Case of the Partial Cohort - May 24, 2020
- EM Nerd: The Case of the Sour Remedy Continues - January 20, 2020
- EM Nerd-The Case of the Adjacent Contradictions - December 23, 2019
Rory
If the authors were using diagnostic codes (not subjective calls) post ER discharge months or years down the line, are you suggesting that they were disregarding the final coding in the charts and superimposing their judgment, i.e., your note of no “event adjudication” of a final diagnosis?
Is that diagnostic review bias or fraud?
Thanks
Brad
No nothing so nefarious. Diagnostic review bias is simply the confounding introduced into any trial when the outcome of interested could potentially be influenced by the diagnostic test in question. In this case the CCTA. Those who were determining the diagnoses that were later being used as the primary outcome of the trial were not blinded to the results of the CCTA. The knowledge of the CCTA results could have easily influenced how the treating physician interprets a troponin elevation. Is it considered an MI or due to some other process? The way to protect against diagnostic review bias is… Read more »
Rory excellent. there are some very smart folk who swear by the CCTA. as of now, is it fair to say that the verdict is still pending? several years ago, i heard Jerry Hoffman rant against the CCTA, in his brooklyn-ish ridiculing laughing way; i do not know where he stands on the issue now. i recognize that this discussion is about study methods, and perhaps faulty, biased conclusions. but as far as CCTA, what’s an ER doc (or hospitalist), who has the capability, to do? is a good quality CCTA that shows “completely normal” or a study that shows… Read more »
I wish I’d remember the interobserver reproducibility of CTCA results, and the concordance between CTCA and conventional coronary angiography. Do you ?