COMMENTARY

Donald Trump Has Coronary Calcium: Is That Helpful to Know?

John Mandrola, MD

Disclosures

February 16, 2018

The debate over Donald Trump's heart health  (or unhealth) raises a core question in cardiology: Will knowing he has coronary calcium, which is present in about 85% of white men his age, lead to better cardiac health?

The idea behind coronary calcium scans is that finding disease is good. Good because we have treatments—statins and aspirin— that reduce the rate of cardiac events. Good because we can personalize care. Good, of course, because doctors control outcomes.

But in medicine, good ideas don't always work out. When you add up the pros and cons of measuring coronary artery calcium (CAC), you find little evidence the scans improve health.

The Case for CAC

Proponents make five main arguments for doing the test:

1. Current scores are imprecise: Current risk scores, such as the atherosclerotic cardiovascular disease (ASCVD) risk estimator, based on the pooled cohort equations are imprecise. Although both the 2013 ACC/AHA cholesterol treatment guidelines[1] and the 2016 US Preventive Services Task Force guidelines[2] use this calculator to determine statin eligibility, numerous studies have shown that it overestimates risk.[3,4] Overestimating risk on this scale raises the possibility of more than a billion humans worldwide taking statins.[5]

2. CAC beats the equation: Many studies that follow people over time show that CAC outperforms the ASCVD risk score or other risk markers for predicting future cardiac events.[6,7]  The novel MESA-CAC score adds CAC to traditional risk factors and improves the 10-year coronary heart disease risk prediction from a c-statistic of 0.75 to 0.80 (0.5 is a roll of the dice).[8]  Also, CAC scores form a stepwise gradient of risk from low to high CAC.[9]

3. Zero CAC warranty: Knowing you have a zero calcium score can be useful. In a cohort of nearly 4800 participants in the MESA study, half were eligible for statins, largely on the basis of a 10-year ASCVD risk of  7.5% or greater. Yet nearly half of this group (44%) recommended or considered for statins had zero calcium at baseline and an observed 10-year ASCVD event rate of 4.2 per 1000 person-years.[10] Proponents say CAC could be used to help patients take fewer meds. 

4. Motivator for lifestyle changes: People who know they have calcium in their coronary arteries may be motivated to make healthy lifestyle changes. In support of this theory, a meta-analysis of six studies with more than 11,000 patients found a two- to threefold increase in the initiation of aspirin, lipid-lowering drugs, blood pressure–lowering drugs, and lifestyle changes in those with calcium vs those with a zero score.[11] 

5. Improved shared decisions: More accurate risk prediction may improve shared decision-making. Statin drugs reduce the risk for a cardiac event by, say, 25%. Shared decisions get more precise with CAC because a 25% reduction in someone with a 15% 10-year risk for an event (lots of calcium) is much bigger in absolute terms of risk reduction than a 25% reduction in someone with only a 3% 10-year risk (no or very low calcium).

The Case Against CAC

Those are the upsides of CAC testing. I am more convinced by the cons.

1. Lack of RCT evidence: The strongest argument against CAC screening is the lack of evidence from randomized controlled trials (RCTs). In the single-center St. Francis heart study, about 1000 people with a calcium score above the 80th percentile for their age and sex were randomly assigned to atorvastatin, 20 mg daily, or placebo. Yes, the rate of cardiac events was lower in the statin group, but this did not reach statistical significance (6.9% vs 9.9%; P=0.08).[12] There are no other adequately powered outcomes trials on CAC screening.

Low cardiac event rates in intermediate-risk people remain the major barrier to studying the effects of CAC screening on outcomes. Experts have estimated that it would require a trial of about 30,000 people, which would be costly.[13] This is my main argument against CAC: If it takes a trial this big to show differences, the benefits—if any—have to be tiny.

2. Oversell of reclassification: A good test should accurately up- and down-classify people, as measured by the test's net reclassification improvement (NRI). Proponents tout CAC's ability to change one's category of risk, but as Dr Andrew Foy (Penn State University Medical Center, Hershey, PA) explained to me, a look at the raw numbers tells a different story.

Let's explore a recent paper in which researchers used the MESA cohort to measure the incremental gains of CAC over and above the ASCVD risk estimator. They started with about 5200 participants (mean age, 61 years) and observed an event rate of 6.2% over 10 years.[7]

They first studied the sensitivity side of the equation, looking at patients who had events. CAC scans correctly up-classified 18% of participants to a higher-risk group.

When it came to specificity, however, CAC scans incorrectly up-classified many more participants to high-risk categories than it correctly down-classified to low-risk categories. This equated to a net reclassification of –6% (negative).

Adding the two yields an overall NRI of 12%. CAC improved classification in 12% of participants. That sounds good, right? Don't be misled. The percentages hide the raw numbers: Far fewer participants have events (n=320) than don't have events (n=4865). That means the overall NRI is a negative number:  (18% × 320) – (6% × 4865) = –234.

In absolute terms, if you undergo CAC screening and are reclassified to a higher-risk group than you would according to the ASCVD calculator, there's about a 1 in 6 chance that this upgrade in risk is correct and a 5 in 6 chance that it is incorrect.

3. Using radiation to scare people into health: In 2014, Dr Steven Nissen from the Cleveland Clinic in Ohio nixed the idea of using CAC scanning as a nudge: "Exposing patients to radiation in order to motivate them…take a deep breath and think about that," he said.

I'd add to Nissen's common sense that the evidence in support of a nudge effect is weak. The EISNER trial, the largest RCT (n=2137) studying CAC screening effects on medical management, reported "statistically" positive results.[14] But this trial is a classic case of statistical significance not equating to clinical significance. Participants who underwent scanning had an incremental 2–mm Hg drop in systolic blood pressure (–7 mm Hg vs –5 mm Hg)  and an incremental 6-mg/dL drop in LDL cholesterol (–17 vs –11 mg/dL) compared with those who didn't. And these were the "positive" results; body weight, waist circumference, and smoking cessation did not change between the two groups.

The remaining evidence that knowing your calcium score can change behavior stems from observational studies, which—at best—find a higher odds that people with CAC will initiate and maintain preventive medications and lifestyle changes.[11] There's a massive gap between the promise of taking more meds and making behavioral changes and having fewer cardiac events.

4. Incidental findings: "The investigation of incidental findings [from cardiac CT] is not without cost or risk," wrote Canadian authors in a paper published in the Journal of the American College of Cardiology.[15] In this series of nearly 1000 patients who underwent cardiac CT, they found incidental noncardiac findings in 41% of patients. Most were not clinically significant, but one patient experienced a major complication from the workup. A group in California reported a similar incidence of noncardiac abnormalities during cardiac scans.  In this series, 1 in 4 of the incidental findings was deemed notable enough to necessitate further medical testing.[16]

One way to reduce incidental findings is to restrict the field of view to the heart. But radiologist Dr Saurabh Jha (University of Pennsylvania, Philadelphia) told me that the search for extracardiac findings is operator dependent, and many radiologists will not restrict the field of view.

In an era where the threshold to investigate any aberration, however benign, is low, the burden of increased anxiety, potential harm, and higher costs of incidental findings cannot be ignored.

5. Downstream testing: No guideline statement recommends that CAC should start a cascade of cardiac testing. The only therapeutic recommendation experts make is to use CAC to inform the decision to initiate or intensify preventive therapies with lifestyle or statins/aspirin.[17] That is it.

But anyone who practices medicine knows the truth: Finding arterial disease in the heart instills fear. And fear often means more tests. I regularly see asymptomatic "patients" with nonzero CAC scores getting stress tests, including nuclear scans, coronary angiography, and, yes, even a stent.

Does CAC Matter Given Current Evidence?

The Society of Cardiovascular Computed Tomography expert consensus is that intermediate-risk people garner the most benefit from CAC screening.[18] 

But we already know how to affect outcomes in this group. We have strong evidence from trials such as HOPE-3[19]  and JUPITER[20] that statin drugs significantly reduce cardiac endpoints in intermediate-risk primary prevention patients. 

CAC  proponents posit that there are heterogeneous treatment effects within these populations. There may be. But we don't have evidence for using CAC to fine-tune who benefits more (or less). Using calcium scores to guide decision-making on statins is based on belief, not evidence.

Summary

The notion of making a personalized diagnosis and treatment decision with CAC is attractive. You can see the white specks, the disease! But…to say seeing things on a scan leads to healthier people requires a lot more evidence. 

What's more, given the (lucrative) downstream testing that often occurs when coronary calcium is found in asymptomatic people, the biggest winners from CAC screening may be the testers rather than the tested.

If clinicians want to improve the health of Donald Trump and millions of other sedentary overweight people like him, they'd be better off prescribing more exercise and less junk food rather than distracting them with calcium scans.

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as:

processing....

Recommendations