Mixed Clinical, Biomarker Score Might Boost CAD Risk Prediction

A novel clinical and biomarker scoring system has the potential to identify obstructive coronary artery disease in a wide range of patients, according to proof-of-concept research.

Today, most patients suspected of having significant coronary artery stenosis—typically defined as at least 70%—undergo a range of often costly tests, such as CT angiography and stress testing, that are not entirely without risk. Biomarker scoring has shown promise in predicting the likelihood of adverse outcomes in A-fib, but it has had mixed success in predicting CAD.

Researchers led by James Januzzi, Jr., MD (Massachusetts General Hospital, Boston), created a scoring system using clinical variables (sex and PCI history) and biomarkers (midkine, adiponectin, apolipoprotein C-I, and kidney injury molecule-1 [KIM-1]). They then tested and validated it in 649 and 278 patients, respectively, from the CASABLANCA study.

Results published in the March 7, 2017, issue of the Journal of the American College of Cardiology showed that elevated scores were predictive of ≥ 70% stenosis in all patients (OR 9.74), men (OR 7.88), women (OR 24.8), and those with no previous CAD (OR 8.67). Additionally, in the validation cohort the score could effectively predict which patients would or would not have CAD in 42% of subjects, with a positive predictive value of 93% and a negative predictive value of 91%. Even in those with “nonsignificant” stenosis, scores were correlated with the degree of narrowing, Januzzi said.

Over a mean follow-up of 3.6 years, the scoring system independently predicted incident acute MI in age- and score-adjusted models (HR 2.23; 95% CI 1.53-3.25).

“Having a widely available and easily interpretable blood test would be expected to be an advantage over much of the conventional ways that we evaluate patients for coronary disease,” Januzzi told TCTMD. “This is the first time that these four proteins that we identified have been put together in this way together with clinical variables to predict coronary artery disease presence. So it is very, very novel.”

Januzzi said that while he was optimistic in general about the usefulness of this scoring system, he was surprised at how the biomarker information “overwhelmed” the effect of clinical variables. Specifically, his team entered “100 or more” clinical variables initially, and only male sex and previous PCI remained as influential. Also noteworthy is how the testing strategy “can both exclude and identify the presence of significant disease in a substantial percentage of patients,” he said. “Biomarkers in general are good for either excluding or identifying disease. To find this balance of sensitivity and specificity is unique and really remarkable I think.”

A Potential ‘Gatekeeper’

In an accompanying editorial, Robert Vogel, MD (University of Colorado Denver), writes that this scoring system stands out for its specific goal of finding “significant” disease as opposed to disease of any severity, but describes this as both a pro and a con.

“Assigning patients to the ‘nonsignificant’ category of coronary disease, as opposed to ‘no angiographic disease,’ is a disservice to patients because they have almost the same risk for major coronary events], and secondary preventive measures are usually not suggested with the same vigor as for those with significant disease,” he says, adding that the results question whether “high grade” coronary disease is associated with different biomarkers than atherosclerosis in general.

Vogel observes that the score’s “combination of clinical and biomarker parameters mimic clinical judgment,” but he too expressed surprise at how only two clinical variables “proved useful.” He also noted that midkine and KIM-1 have never before been linked to coronary atherosclerosis, but remarked that “it is interesting that none of the four biomarkers identified is a marker of inflammation as would be expected if the goal had been determining unstable disease.”

Ultimately calling these results “preliminary” due to the division of patients into separate training and validation cohorts, Vogel said this scoring system should be further studied “in an independent population selected with more varied inclusion criterion.”

Januzzi agreed. “It is a step toward a personalized precision approach for using biological information with validation to predict the presence of a very common and very important clinical diagnosis,” he said. “With validation and further data regarding this score strategy, I think we could envision seeing it used clinically. More data are needed, and we're in the process of collecting it.”

As for whether this system could ultimately change routine practice, Januzzi said it could “quite conceivably [be used as] a gatekeeper toward more expensive testing like stress testing or CT angiography, eliminating the need for some cases and increasing the need in others depending on the results of the score.”

Specifically, he plans to investigate the score’s use in both stable, office-based patients with “relatively subtle symptoms” and in unstable patients presenting the emergency department with ambiguity as to whether or not they have an ACS. “There are a wide range of potential applications here that still remain to be examined, but there's definitely a lot of promise,” Januzzi concluded.