This article is more than 5 years old. Information may no longer be current.
Genomic blood test predicts revascularization events in PROMISE cohort
An age- and sex-specific gene expression score calculated from the results of a blood test was a significant predictor of revascularization events in symptomatic patients without diabetes from the PROMISE trial.
“These hypothesis-generating results indicate this score may have a role in the noninvasive evaluation of people presenting with stable chest pain,” Pamela S. Douglas, MD, professor of medicine and the Ursula Geller Professor for Research in Cardiovascular Disease at Duke University School of Medicine, told Cardiology Today. “We have not tested it head-to-head vs. more conventional approaches such as doing a stress test or a CTA, but it seems to provide good information that will help with care.”
The researchers retrospectively measured the gene expression score derived from samples of a blood test (Corus CAD, CardioDx), previously validated to detect obstructive CAD, in 2,370 participants from the PROMISE trial without diabetes (median age, 60 years; 48% men; 1,137 with CTA data).
The outcomes of interest were obstructive CAD, defined as at least 70% stenosis in at least one vessel or at least 50% left main stenosis on CTA; and a composite of death, MI, revascularization or unstable angina.
In unadjusted analyses, a gene expression score > 15 was associated with obstructive CAD (OR = 2.5; 95% CI, 1.6-3.8) and the composite endpoint (HR = 2.6; 95% CI, 1.8-3.9), according to the researchers.
After adjustment for Framingham risk score, the association between gene expression score > 15 and the composite endpoint remained (HR = 1.7; 95% CI, 1.1-2.64), driven by revascularization (HR = 2.69; 95% CI, 1.52-4.79); the other components of the composite endpoint were not statistically significant, Douglas and colleagues wrote.
“If people are looking for prediction of events other than revascularization, the score may not be that effective, but if they need to predict future revascularization, the score did seem to predict that, though not in a prospectively adjudicated sense,” Douglas told Cardiology Today. “Clinicians made the decision to revascularize in an observational way; they did not have access to this information at the time of the decision. We weren’t predicting the need [for revascularization], but whether in routine care the patient had received it.”
Compared with noninvasive testing, the gene expression score improved prediction for the composite endpoint (improvement in C-statistic, 0.036; continuous net reclassification index, 43.2%), according to the researchers.
Patients with a gene expression score 15 had a 3.2% rate of the composite endpoint, similar to the 2.6% rate for those with negative results of noninvasive testing (P = .29), Douglas and colleagues wrote.
“We don’t yet know the impact of integrating the clinical score into decision-making processes,” Douglas said. “But importantly, the association [between the score and revascularization] provides support for doing future research that would prospectively assess the value in clinical decision-making. You could easily randomly assign patients to the usual care, which would be stress testing or CT, or to the blood test, and see what the clinical outcomes were in terms of continued chest pain, the need for catheterization, catheterization without obstructive disease, and hard clinical endpoints such as death and MI.”
In practice, it would be “convenient to draw blood at the first health care encounter for symptoms of chest pain rather than having to schedule another visit to perform a noninvasive test,” Douglas told Cardiology Today. “There are potential efficiencies for the care pathway.” – by Erik Swain
For more information:
Pamela S. Douglas, MD , can be reached at 7022 North Pavilion DUMC, Durham, NC 27715; email: pamela.douglas@duke.edu.
Disclosure: The PROMISE trial was funded by the NHLBI. CardioDx provided blood tests and laboratory services for free for the present study. Douglas reports serving on a data and safety monitoring board for GE Healthcare and receiving grant support from HeartFlow. Please see the full study for a list of the other researchers’ relevant financial disclosures.
Perspective
Back to Top
This is an important study for a couple of reasons. No. 1, it is a large study that reinforces what we saw in previous smaller studies such as COMPASS and PREDICT. It’s common for cardiologists to wait for a large study to confirm smaller studies before adapting practice. No. 2, clinical and academic cardiologists like to see independent corroboration of a concept. The PROMISE study was independently funded by the NHLBI and conducted by Duke Clinical Research Institute, and not by a company. While the studies run by the companies who make the products are helpful, independent studies provide confidence. The beauty of PROMISE is that it’s very large and completely independent of CardioDx. It’s better evidence than we had up until now that the Corus blood test is very valuable in identifying CAD as well as predicting who is at high risk for future events, which is very powerful information.
The Corus blood test is a tool that clinical cardiologists can use to help us differentiate which patients have coronary disease and which don’t. I have used this test judiciously but with great success for 2 or 3 years. Once these results are disseminated, I think more cardiologists and eventually more primary care physicians are going to start using this test. Coronary disease is the No. 1 killer in the U.S., and diagnosing and treating it is the No. 1 job of a cardiologist. This test gives us an opportunity to diagnose it relatively easily at little or no risk. A lot of cardiologists don’t know about it, and these results may get their attention.
The primary use of the test is in a patient that is believed to be at low risk for major coronary disease. For instance, chest discomfort is one of the top complaints that cardiologists and primary care doctors have to address. We know that 80% to 90% of chest discomfort turns out to be noncardiac. One of my main jobs is playing detective for any potential cardiac symptom. Often there is a relatively low clinical suspicion that the patient has cardiac chest discomfort, but even so, testing that proves that suspicion correct is warranted. With a quick, relatively inexpensive blood test, you get an answer in about 3 days with a 96% negative predictive value. If it comes back negative, that reassures the clinician and the patient with at least 96% accuracy that the source is noncardiac, and no further tests are required. In cases like these, I use it for proof that my gut feeling is right. If the test comes back positive, then we proceed to a stress test.
The second type of patient for which the test has shown to be valuable is where I believe a patient might have angina, especially if he or she has a family history of coronary disease, but the stress test was normal. It’s well known that stress tests can have false-negative results. Most patients don’t have further testing after a negative stress test, and there is a chance we could miss something. If one is concerned that there is a cardiac problem despite a negative stress test, the blood test can be used as a tiebreaker. If the blood test is also negative, that’s a good indication that the problem is noncardiac. If the blood test is positive, then a decision whether to proceed to cardiac catheterization must be made.
One of the first times I used the blood test was in a patient in his 50s with a family history of early coronary disease and described angina. His stress test was normal. My gut feeling was that he had coronary disease, but I couldn’t justify performing a cardiac catheterization at that time because the symptoms were not bad and the stress test was normal. I ordered the blood test based on its negative predictive value, but the results came back with a high score. With the added information, the patient was agreeable to a cardiac catheterization, which revealed a tight proximal left anterior descending artery lesion, despite the nuclear scan being normal.
Doctors are being held more accountable for use of health care dollars and have to be smart with the money we use. Cardiac catheterizations are very important, but they are very expensive and present a risk for injury. The value of the blood test is that it should decrease the number of cardiac catheterizations that reveal a normal result, putting patients at less risk for complications and saving money in the long run.