Low-risk chest pain patients should be sent home without imaging, concludes a new study in JAMA Internal Medicine. But several experts are disputing the study's main finding that treadmill stress tests, as well as CT and other imaging tests, have no value in preventing heart attacks in low-risk individuals.
The massive study analyzed the insurance claims of more than 400,000 chest pain patients who presented to U.S. emergency departments (EDs) in 2011. It found that low-risk patients who did not undergo imaging or a treadmill stress test when they presented to the ED had no more heart attacks than patients who did -- either at seven days of follow-up (0.11% myocardial infarction [MI] rate) or 190 days of follow-up (0.33% MI rate), according to authors from Penn State University's Hershey Medical Center (JAMA IM, January 26, 2015).
"I think the most important finding is that early cardiac stress testing compared to no stress testing was not associated with lower odds of having a heart attack in the future," said lead author Dr. Andrew Foy, an assistant professor of medicine and health science at Penn State College of Medicine, in an interview with AuntMinnie.com.
Other noninvasive imaging tests such as echocardiography, SPECT myocardial scintigraphy, and CT coronary angiography did no better, he said. Early cardiac testing wasn't associated with better outcomes in the study, but scans and treadmill tests were associated with higher odds of downstream tests and procedures such as revascularization.
"Given that these early procedures were not associated with reductions in concomitant myocardial infarction, it's likely that these were unnecessary," he said.
Chest pain problems in the U.S.
An estimated $10 billion to $12 billion is spent each year to evaluate chest pain patients, but the risk of heart attack is low in patients without objective evidence of myocardial ischemia, noted Foy and colleagues including Guodong Liu, PhD, and Dr. William Davidson Jr. The cause of these symptoms is noncardiac in many patients, and the optimal management strategy is unknown.
Moreover, even though the American Heart Association has endorsed both the safety and usefulness of stress tests to detect ischemia or coronary artery disease in chest pain patients, "there is no evidence that testing reduces the risk of future cardiac events compared with a more conservative approach," the authors wrote.
The study of 421,774 patients aimed to compare the association between exercise electrocardiography, stress echocardiography, myocardial perfusion scintigraphy (MPS), coronary CT angiography (CCTA), or no imaging and events such as downstream cardiac catheterizations, revascularization procedures, future noninvasive imaging tests, and hospitalizations for myocardial infarction.
The retrospective analysis included health insurance claims data for all of 2011 in the cohort of privately insured patients, and it was limited to patients with a primary or secondary diagnosis of chest pain in the emergency department. The researchers measured the number of patients in each group who received cardiac catheterization, coronary revascularization, or a noninvasive future test, as well as those who were hospitalized for myocardial infarction at seven and 190 days of follow-up.
Overall, the results showed that the percentage of patients hospitalized with myocardial infarction was very low: only 0.11% at seven days of follow-up and 0.33% at 190 days of follow-up, representing a total of 464 patients, the authors wrote.
Compared with the no-testing cohort, there were no significant differences in the rates of hospitalization, catheterization, revascularization, or myocardial infarction in any of the noninvasive imaging groups.
| Adjusted odds ratios (AOR) after initial noninvasive test | |||||
| Test | Revasc.: AOR at 7 days | Revasc.: AOR at 190 days | MI: AOR at 7 days | MI: AOR at 190 days | Additional noninvasive test: AOR at 190 days |
| No test (reference) | 1 | 1 | 1 | 1 | 1 |
| Stress echo | 1.54 | 1.08 | 0.99 | 0.83 | 1.47 |
| Exercise ECG | 2.41 | 1.55 | 1.33 | 0.94 | 1.63 |
| MPS | 2.40 | 1.71 | 1.25 | 1.03 | 2.31 |
| CCTA | 3.56 | 1.95 | 1.70 | 1.20 | 1.58 |
The results showed that patients with CCTA had the highest rate of immediate revascularization after their scans. Could these revascularizations have contributed to the paucity of problems in longer-term follow-up?
"No, our interpretation would be that they were finding more disease, but it was disease that didn't necessarily need to be found," Foy told AuntMinnie.com. "A problem with CCTA is that it doesn't provide functional information, and a lot of people as they get older develop atherosclerotic coronary artery disease. We see more disease and there is more treatment, but that treatment is probably not helpful and could potentially be harmful."
Based on the results, compared with no testing, exercise electrocardiography, myocardial perfusion scintigraphy, and CCTA were all associated with significantly higher odds of cardiac catheterization and revascularization procedures, yet they did not confer an improvement in the odds of experiencing a myocardial infarction, according to the authors.
They estimated that for every 27 patients who undergo myocardial perfusion scintigraphy instead of a no-testing approach, one will undergo an unnecessary catheterization. In addition, for every 71 patients who undergo CCTA instead of no testing, one will undergo unnecessary catheterization.
The across-the-board low prevalence of MI did not seem to be affected by the initial testing strategy, and given this result, deferring noninvasive testing until it proves to be needed appears to be "a reasonable approach," the authors wrote.
They concluded that a randomized controlled study is the only way to truly clarify the best testing strategy for low-risk patients being evaluated for chest pain in the ED. But cost is paramount.
"Given today's concerns regarding healthcare cost growth, especially the portion attributable to noninvasive cardiac imaging, and patient safety issues related to radiation exposure as well as overdiagnosis, performing such a study should be a priority," they wrote.
Redberg agrees
In an accompanying editorial, cardiologist and professor of medicine Dr. Rita Redberg, from the University of California, San Francisco, said the study goes a long way toward explaining the intricacies of what to do with low-risk patients presenting to the emergency department with chest pain. After six months of follow-up, only 0.33% were hospitalized with myocardial infarction. So what's the best solution for many of them? Send them home.
The findings suggest that the current practice of performing a stress test on low-risk patients in the ED is "unnecessary and prolongs the length of stay in EDs as well as increases unnecessary medical imaging, with a significant associated radiation risk for tests that include nuclear imaging," Redberg wrote.
"It is time to change our guidelines and practice for treatment of chest pain in low-risk patients," she continued. "Such patients should be given a close follow-up appointment with a primary care physician who can determine, based on the patient's condition, whether further evaluation is necessary."
But sick patients were excluded
On the other hand, several experts contacted by AuntMinnie.com who reviewed the study disputed its findings and disagreed with its methodology. The study's retrospective design "creates a selectively truncated cohort for study," Dr. Geoffrey Rubin, professor of cardiovascular research, radiology, and bioengineering at Duke University, wrote in an email. Moreover, the patients who underwent imaging versus those who did not were substantially older and sicker.
"Averaging nine years older and with more than twice the prevalence of diabetes, hypertension, hyperlipidemia, and ischemic coronary disease, those receiving noninvasive testing were at higher risk than those who were not tested," Rubin said.
Rubin also disagreed with the researchers' decision to exclude certain patients from the study.
"Physicians in the emergency department triage patients based upon level of risk determined from information at hand. By excluding a large group of subjects who underwent noninvasive testing, who were admitted, and subsequently proven to have myocardial infarction, the authors amputate a key part of the acute chest pain population," Rubin continued. "If the ED docs knew in advance who was having an MI, then they wouldn't have needed to do the test."
When discussing study limitations, Foy and colleagues wrote in their paper that the "no-testing cohort was significantly younger and had fewer comorbid conditions than the cohort undergoing an initial noninvasive test." But the effect on the study results was not likely to be significant given the high prevalence of false positives in imaging tests, they added.
"To limit confounding as much as possible, strict exclusion criteria were applied to the cohorts to make them representative of low-risk patients with chest pain," they wrote.
However, Dr. U. Joseph Schoepf, professor of medicine, radiology, and pediatrics at the Medical University of South Carolina, countered that the result of this decision was anything but even-handed.
"The authors state that their strategy aimed at carefully avoiding bias; however, their exclusion maneuvers introduce the most fundamental bias conceivable, setting up the study design against imaging to predictably produce the desired results," he said.
The reported mean age of the noninvasive imaging cohort was nearly 50 years versus 59 for the imaging cohort, and patients who underwent initial noninvasive testing had more comorbidities and a greater likelihood of being hospitalized on their index chest pain encounter. Initial imaging patients were also more likely to have undergone noninvasive testing within six months before their index chest pain visit to the ED.
The JAMA piece "is just another egregious example of how an investigation based on a shallow, flawed (some might call it 'quick and dirty') study design is abused in an attempt to influence healthcare policy," Schoepf wrote in an email to AuntMinnie.com.
The researchers also chose to exclude anyone with a diagnosis of noncardiac reasons for chest pain. This severely narrowed the study focus to a question of whether chest pain is related to myocardial infarction or not, which is unrealistic in clinical reality.
"Narrowing the outcomes selectively to patients with a major adverse cardiac event also blatantly disregards the opportunities arising from the early detection of disease with subsequent risk stratification," according to Schoepf.
Foy responds to criticism
Foy said there were too many individual points in Rubin and Schoepf's commentaries to address each of them, but "in general, they have valid criticisms of the study, most of which we acknowledged in the discussion section of our paper."
He said via email that he and his colleagues accounted for differences in age and risk factors as well as additional covariates in their statistical design.
"In our opinion, the fact that we excluded patients with an MI on the index encounter actually favors the stress testing population, because if patients had a cardiac stress test and were then diagnosed with an MI, it would imply the MI was due to the intervention -- otherwise, it would have been discovered in the labs prior to the stress test and a stress test would not have been performed. Stress testing is contraindicated in patients with an MI or objective evidence of ischemia on the ECG," Foy said.
The stress test is not performed to diagnose MI, he added. It is done to delineate patients who may be at higher risk in the near future and may benefit from more aggressive therapy.
"Ultimately, we concluded our manuscript by saying that the only way to definitively address this issue would be to perform a randomized trial," Foy said. CCTA trials such as Rule Out Myocardial Infarction Using Computer-Assisted Tomography (ROMICAT) do not suffice because they were designed to compare CCTA to usual care, which generally included stress testing.
A randomized trial has never been undertaken in the population examined in this study, Foy said.
Still, Schoepf believes that the study's flaws were serious enough to question why it was published at all.
"A transparent, unbiased peer-review process should have detected these study flaws and resulted in manuscript rejection," Schoepf concluded. "By the selective publication of such questionable contributions, otherwise highly regarded journals allow themselves to be abused as a vehicle for the shortsighted opinion of few. This does a disservice to the vast field of researchers who are genuinely interested in developing novel, contemporary concepts in our current diagnostic and therapeutic strategies, which are admittedly flawed. More importantly, they are doing a disservice to our patients who are potentially put in harm's way, based on unsustainable conclusions derived from an unrealistic, artificial study design."
