Call For More Calcium Screening Gets Pushback

–But a zero calcium score can be useful to avoid statins, some argue

Coronary artery calcium (CAC) scans should be widely used in routine clinical practice to improve the detection of coronary disease in people without known disease, according to the authors of a new review in JACC: Cardiovascular Imaging.

But many experts urged caution, and the paper also offered plentiful ammunition for the argument against routine use of CAC scanning. The authors acknowledge that persuasive evidence in the form of randomized clinical trials is unlikely to appear any time soon.

Coronary scanning has already been shown to improve prognosis of CV disease, the authors, led by Alan Rozanski (Mount Sinai St. Luke’s Hospital, New York City) wrote. But adoption has been slow because CAC scans have not been proven to improve outcomes, they wrote.

The argument in favor of improved CV screening— what the authors deem “the imperative to screen for CVD”— is based on the fact that cardiovascular disease continues to be the leading cause of death in the western world, despite enormous reductions and the use of current screening methods based on established CV risk factors.

“By using imaging for screening, we can detect problems early on, which gives the patient an opportunity to make lifestyle changes to help avoid developing heart disease—such as by improving nutrition, starting to exercise or quitting smoking,” Rozanski said in a press release. “We believe this will not only help improve and save lives but that it can ultimately contribute to lower health costs since the earlier adoption of positive health habits can reduce patients clinical risk and potentially eliminate the need for more costly interventions later on.”

But the authors also acknowledged that clinical benefits have not been proven in clinical trials. They explained in detail why it is difficult to prove the benefits of routine CAC scans in outcomes trials. Because the event rate in the primary prevention population is low, trials need to enroll large numbers of patients who must be followed for a long time. Over time, the burden of proof becomes even tougher, due to the long term overall decline in CV disease.

Given these difficulties, the authors argued that it may be necessary to use trials with “intermediate endpoints”: “Because of the very low event rate in current imaging populations, the long duration of follow-up that would be required, and the growing ubiquity of aggressive medical management in general, developing prospective imaging trials for prevention that would be based on hard clinical events may be increasingly impractical. However, a variety of intermediate endpoints may be suitable for study in smaller trials.”

But the authors don’t address the obvious alternative perspective: If they are so difficult to prove, the benefits are unlikely to outweigh the costs and the harms.

Skepticism About Screening

I asked Khurram Nasir (Baptist Health South Florida, Miami), a self-described “screening skeptic,” to comment on the paper.

The argument in favor of screening is based on “the general assumption” that the existing tools to assess risk “are inadequate to identify early disease,” said Nasir. The result is often either “initiation or intensification of establishment treatment,” with the underlying assumption “that this will translate into lower downstream risk, usually in [the] form of preventable events.”

But Nasir argued that this is a “myopic view,” especially for CAC screening. First, he pointed out, under the current guidelines two-thirds of adults are already statin candidates and CAC testing is “unlikely to impact management decision for upgrading risk for treatment purposes.” In addition, among those adults who are not eligible for treatment under the current guidelines, “we need to scan more than 25 individuals to find one with early disease, and even that does not change the projected risk in a meaningful manner in most cases.”

Nasir warned against the “seductive concept” that surrounds the “CAC screening paradigm.” Summarizing, he said “it is very unlikely to work because of a) low disease burden in population, b) better outcomes than expected with improvements in treatment and lifestyle interventions, c) little marginal utility with more aggressive treatment as seen in recent studies, and more importantly, d) very few are no longer candidates for treatment (what’s left to screen?).”

A Better Role For CAC Screening

Nasir has previously proposed a different role for calcium scans. Instead of being used as a broad screening tool, he believes calcium scans may be useful in select cases as a decision tool in patients for whom the benefits of statin treatment are unclear. In particular, a zero calcium score may help many people at intermediate risk avoid statin therapy.

Nasir’s position received support from Saurabh Jha (University of Pennsylvania, Philadelphia), a radiologist who has thought deeply about the complexities of screening. Here is his lengthy response to the Rozanski paper:

“Being a radiologist, I rarely speak to patients, but I was asked to counsel Mrs. Patel (not her real name, so calm down HIPAA totalitarians), who was worried about the risks of radiation from cardiac calcium CT scan. Because of her risk factors for atherosclerosis, her cardiologist wanted her to take statins for primary prevention, but she was reluctant to start statins. They eventually reached a truce. If she had even a speck of calcium in her coronary arteries she would take statins. If her calcium score was zero she wouldn’t. This type of shared decision making is the most frequent reason why cardiologists order calcium scans at my institution.

“A calcium scan is a nifty test, not because it improves outcomes – that’s a population-based consideration – but because it changes management, specifically when there is zero calcium. It does this by so lowering the patient’s risk profile that they no longer meet the risk threshold deemed by the AHA [American Heart Association], and endorsed by the USPSTF [U.S. Preventive Services Task Force], for starting statins for primary prevention.

“You can quibble about the threshold for recommending statins, but there is no quibbling that [a] zero-calcium scan often reclassifies risk bringing the person to a lower risk than previously thought. Zero calcium portends a happier future. In one study, of those who were eligible for statins because their estimated risk of cardiovascular events over 10 years was > 7.5%, nearly half had zero calcium, which put them at a lower risk profile.

“Calcium scans unmask a tremendous amount of risk heterogeneity, even amongst those at high risk for cardiovascular events. A zero calcium even in someone with the highest Framingham Risk Score (FRS) shifts their risk profile to within the safer territories of the lower FRS. The shift is quite dramatic, and affects the intensity of primary prevention or the need to indulge in it. I’m reluctant to use the term ‘precision medicine’ because Vinay Prasad will chew me alive on Twitter, but careful use of calcium scans is an example of ‘preciser medicine.’

“The rhetoric with calcium scans can be misleading. For example, some call it a ‘mammogram of the heart’ – aside from the insensitive gender nihilism, this also insinuates, to borrow a slogan from the screening world, that ‘calcium scans save lives.’ This is not strictly correct. First, a zero-calcium scan does not exclude the deadlier, non-calcified plaque, which is often the culprit lesion in young people with fatal myocardial infarctions. But such rhetoric would also need a sufficiently powered randomized controlled trial. The sample size is likely to be immense, perhaps greater than the population of Mumbai, because in an undifferentiated population, many of whom may already be eligible for statins for primary prevention, and thus be optimally medicated, the incremental mortality benefits, at population level, from change in therapy, conditional on knowing that there’s calcium is not likely to be tremendous. Calcium scan is an example of subtractive medicine – it subtracts statins from people. We need more subtractive medicine.

“Calcium scans can induce downstream testing, but radiologists are getting better at limiting the field of view to the heart, to avoid seeing the thyroid and adrenal glands – organs which often beg to be over tested. See no thyroid nodule, hear no thyroid nodule. There’s no getting away from the pesky lungs, unfortunately, which means sometimes we will see and follow-up lung nodules, and sometimes we will catch an incidental lung cancer and pretend to have saved the patient’s life. There is also the ascending aorta, a tortuous structure jutting out of the heart, and wandering aimlessly for some distance, which, but for the grace of a radiologist’s caliper-happy fingers, is minding its own business (I tell radiology residents not to measure the ascending aorta unless it’s blitheringly obvious that it’s large).

“Carefully used, with the intent of changing management, a calcium scan, specifically when there is zero calcium, is a useful test. It’s like a GPS which tells you to avoid the NJ [New Jersey] Turnpike.”



  1. James H. Stein says:

    Regarding primary prevention, screening for subclinical arterial disease and refining prediction used to be really important when statins were expensive and were perceived to have serious side-effects like hepatotoxicity. Starting a statin used to be a big decision. Now, with over 25 years of safety data, generic statins, and a more thorough understanding of their effectiveness and side effects, the gain from any additional testing is small and limited to small niches of patients, including some discussed above. When one considers lifetime rather than 10-year risk and that statins are a modest lever among the many interventions that affect CVD outcomes, imaging ahead of time seems like long run for a short slide that just gives the illusion of much more precision. More important is having a discussion about patient preferences, NNT, NNH, drug costs, drug interactions, and other personal considerations. IMHO, the greatest advance in the 2013 AHA/ACC guidelines was adding in shared decision making.

  2. James H Stein says:

    I’d add that the excuses of “it would take too many people to show …” or “it would take too long to show …” ring hollow to this imager and preventive cardiologist, as it suggests that the intervention is not effective. Indeed, I always teach my students that if something has been around 20 years and people are still debating its value, it can’t be very much. Nowadays, a CAC screening strategy isn’t really the alternative to traditional risk assessment. The strategy of “treat all” is the real competition.

Leave a Reply

%d bloggers like this: