CARDIOVASCULAR deaths in Australia have seen an 82% relative reduction from their peak in 1968. Much of this decline has been due to widespread decreases in smoking, blood pressure and cholesterol levels in the population. Increased use of blood pressure-lowering and lipid-lowering therapy has also played a significant role.

Deaths and disability from cardiovascular disease (CVD) might be reduced even further by improved detection of people at high risk for treatment with blood pressure-lowering and lipid-lowering medication.

Our best current means to determine who is at high risk (and thus might benefit from treatment) is a CVD risk calculator. Single risk factors, such as blood pressure and lipid levels, are poor predictors of overall (or absolute) cardiovascular risk, and clinicians are not able to intuitively combine these factors accurately. The Framingham risk equation, based on the Framingham Heart Study and the Framingham Offspring Study cohorts, accurately predicted CVD events in a recent Australian population study. Other risk equations such as QRISK3, developed for the UK population, and PREDICT, developed for the New Zealand population, also include health equity-associated risk factors, such as self-identified ethnicity and socio-economic status.

Many other new factors have been proposed as potentially useful additions to risk assessment. One prominent recent proposal has been coronary artery calcium (CAC) scans, which are increasingly used in clinical practice in Australia and overseas. However, it has been unclear how much information the scans add beyond that from traditional risk assessment using a CVD risk calculator.

To find out, we undertook a systematic review and meta-analysis of published studies that provided relevant data on the incremental value of CAC scans, beyond that provided by traditional cardiovascular risk scores. Using the 2018 US Preventive Services Task Force review on the topic as our starting point, we screened 2772 titles and abstracts, and 53 full text articles. We included six community-based cohort studies from the US, the Netherlands, Germany and South Korea, with CVD event rates ranging from 0.9% to 9.4% over 4–10 years’ follow-up.

The studies showed modest gain in discrimination when scans were added to the risk equation (ability of the models to separate those who did and did not have a CVD event), and this appeared consistent across studies. The C statistic (area under the receiver operating characteristic [ROC] curve) for the CVD risk models without CAC scan results ranged from 0.693 to 0.80. The pooled gain in C statistic from adding a CAC score was 0.036 overall, with possible larger gain for studies using coronary heart disease events as the outcome (gain in C statistic of 0.049) than for studies using CVD events (gain in C statistic, 0.029).

We also found that the test may have incremental prognostic value for some participants. But the clinical meaning of these small gains in discrimination and prognostic value was unclear.

Among participants reclassified from low risk by the risk score to intermediate or high risk by the CAC scan result, only 3.6% to 14.5% had a CVD event during follow-up, while 85.5–96.4% did not. Among participants reclassified from intermediate or high risk by the risk score to low risk by CAC scan result, 91.4–99.2% did not have a CVD event during follow-up, while 0.8–8.6% did have a CVD event.

In summary, our study suggests that the CAC scans offer a small improvement to risk prediction based on traditional CVD risk assessment equations. These potential (but uncertain) benefits need to be weighed against potential harms from unnecessary tests, diagnoses and treatments, radiation risk, and costs. An additional consideration is that the scans have a large carbon footprint, and their indiscriminate use contributes to health care emissions.

CAC scans may have a role for refining risk assessment in selected patients. This possible role is suggested by modest improvement in discrimination (C statistic) and evidence of prognostic value with hazard ratios that ranged from 1.29 to 1.75 per standard deviation, after adjustment for other CVD risk factors.

The groups most likely to benefit are patients for whom, after standard CVD risk score assessment, reasonable likelihood exists that a CAC scan could help “in clarifying whether the risk is high enough to justify primary prevention medications”. Further refinement to assessing the incremental gains is likely to need individual patient data to define who may benefit from a CAC scan. Methods are needed that avoid using strata defined by CVD risk equations to assess the incremental gain, because such stratified analysis is seriously flawed.

At present, there is no direct evidence that adding the CAC scores to traditional risk scores provides clinical benefit.

Katy Bell is a clinical epidemiologist and health services researcher in the School of Public Health at the University of Sydney. She has expertise in the evaluation of the clinical effectiveness of health care, and her research aims to identify sustainable models of health care that benefit health and not cause harm.

Anna Mae Scott is an Assistant Professor at the Institute of Evidence-Based Healthcare at Bond University. She is an epidemiologist, with interest in conducting and developing new methodology for evidence syntheses.

Dr Lin Zhu is a statistician and postdoctoral researcher in the School of Public Health at the University of Sydney.

Paul Glasziou is Professor of Evidence-Based Medicine at Bond University was a part-time General Practitioner for many years. His key interests include identifying and removing the barriers to using high quality research in everyday clinical practice, and improving the process of research through automated reviews and better reporting.



The statements or opinions expressed in this article reflect the views of the authors and do not necessarily represent the official policy of the AMA, the MJA or InSight+ unless so stated.

Subscribe to the free InSight+ weekly newsletter here. It is available to all readers, not just registered medical practitioners.

If you would like to submit an article for consideration, send a Word version to


I am interested in training as a voluntary assisted dying physician
  • No (61%, 90 Votes)
  • Yes (24%, 36 Votes)
  • I'm thinking about it (15%, 22 Votes)

Total Voters: 148

Loading ... Loading ...

8 thoughts on “Do coronary artery calcium scans add benefit?

  1. Brett Forge says:

    As MESA showed 50% of people judged suitable for statin therapy had scores of zero. In this group there is a huge saving in money and unnecessary medicalisation. A problem with this kind of data is that people at higher risk will get statins so that their risk will drop substantially. It will appear as if the scan has not helped but in fact it shows that those at high risk will drop their risk to those with low risk. To assess benefit a randomised trial would have to withhold treatment from those who have high CTCS. Such a trial is unlikely.

    The huge advantage is that CTCS is a measure of disease. Completely different from measuring risk.

  2. A/Prof Paul Langton says:

    Too much emphasis is placed on arbitrary risk bands and categories, the value of which,at best, is to ease communication to the patient .
    We know that aside from gender, all CV risks are Continuous, not Categorical. I personally find that the risk calculators are the best starting point, to which I add individual consideration of family history, diabetes (which automatically implies higher risk), CKD, family history and lifestyle factors are added. Biomarkers such as hsCRP and homocysteine can be useful in some individuals, as can CACS.
    More often than not the patient has already had a CACS, and that has prompted Cardiology referral. I personally use CACS very selectively, but often use CTCA is prognostic assessment of symptomatic patients (including those with atypical features).
    Primary prevention remains focused on treatable lifestyle and medical risk factors. The major benefit of interventional cardiology practice are limited to (1) STEMI & (2) reduction in angina in symptomatic patients – BUT patients with high risk anatomy (tight Left Main or proximal LAD stenosis) have not been included in trials of Optimal Medical Therapy +/- revascularisation, so there is still some uncertainty in these groups. These are the patients I’m looking for with CTCA.

  3. George Crisp says:

    Dear JulianP – anecdotes are NOT evidence.

    Whilst it clearly is possible, we can not really know whether you would have benefited from this “diagnosis” or not. The absolute risk reduction of taking a statin is alos low.

    And you could have quite easily instituted the dietary and lifestyle measures without risk stratification.

    By the way have you looked at reducing your exposure to ambient and indoor air pollution? ( which is also a major risk factor for both contributing to atherosclerosis and acute cardiac and vascular events).

    So avoiding car commuting, trips to polluted cities, staying indoors on days smokey days, avoiding wood heaters and not using gas appliances in your home etc.

  4. Andrew Wettenhall says:

    CV risk calculators miss predicting ~45% of ASCVD events, CAC misses only 5%. Cost aside its better. Major limitation is current reporting often gives Agatston score only but not plaque counts/density. Score of zero can prevent unnecessary treatment and allow reasonable reassurance so long as test not done too early in life or too soon after onset of prediabetes/diabetes (i.e. before calcification). Non zero scores esp those >100 can motivate statin compliance. Its not perfect and shouldn’t be used to ignore modifiable risk factors, but its better than what we had prior.

  5. Sue Ieraci says:

    As in most areas of medical practice, we need to consider the pathophysiology. This is the basis of scientific medicine. When many of us were training, we saw a lot of so-called “widow-maker” infarcts in men in their 50’s and 60’s who were generally smokers. They had proximal clots in major coronary vessels due to rupture of non-calcified (unstable) plaque. This was the basis of the “vessel-opening movement” that started with thrombolysis with STK in the 1980s, and then spawned the huge interventional cardiology industry. Cardiologists were once the gate-keeper to the cardiac surgeon – now they refer their own patients to themselves for highly lucrative procedures, including non-evidence-based stenting in many cases (the evidence is only solid for STEMI).

    The patient population has changed. With massive reduction in smoking and greater longevity, the single-vessel acute “widow-maker” clot is not the most common pattern. Many older women have diffuse coronary disease, often in anatomically small vessels – not single ulcerated plaques. Calcified plaques are, by definition, more stable and less prone to rupture and clot. Generalised disease in smaller vessels with some plaque further narrowing it is not best-served by the current style of interventional cardiology, which is being fed by cardiac imaging.

    Calcification of all arteries – not just coronary – is a degenerative condition that increases with age (and is accelerated by the causes we all know).

    Having said all that, a zero calcium score provides some reassurance that there is no need to drive down the lipid levels in someone with otherwise a low risk score. A more than zero score appears to be less useful.

  6. Anonymous says:

    This article arrived in my inbox the afternoon of the morning in which I had a CAC and CT coronary angiogram. I am a recently retired, non-smoking, normal weight and cholesterol, slightly elevated BP, exercise-loving male with no family or personal family history of ischaemic heart disease – but I do come from the West of Scotland and I have late-onset T1DM! I recently had 4 short-lived episodes of severe epigastric pain with sweating and my medical friends declared the symptoms to be cardiac. The verbal results suggest of this morning’s tests that they are correct. So now I have to proceed to formal investigations – a real nuisance but clearly inevitable! Risk estimates are by nature imprecise and this has been recognised for many years. By the way, the inclusion of the carbon footprint of CT scans and contribution to “Health Care emissions” with no data whatsoever on the proportion of scans that are indiscriminate or clinical outcomes were unscientific at best.

  7. CN Aroney, MBBS, MD, FRACP, FCSANZ says:

    CVD risk tools are actually very poor predictors of individual risk…and are really only predictive of population risk. Large studies using CVD risk tools shows that 75% of persons having a myocardial infarction were considered low risk, the day before their infarct. A CACS score of zero is incredibly good information, and these persons are vitrually immune from heart attack for the next 15 years. The above study does not admit that the CACS is superior to all the risk factors combined in predicting risk…and the C statistic is improved for CHD events in incremental risk prediction. Any positive score provides incremental risk, and asprin and statin therapy should be considered, especially if the score is greater than 100.

  8. JulianP says:

    As a very fit doctor with an excellent BP, an enviable lipid profile and a normal BMI, I was initially reluctant to undergo a CAC. My only risk factor was a family history of coronary artery disease. My clever GP also ordered a CT angiogram for good measure. The CAC score placed me at low risk, but the CT angiogram was an entirely different story. With plaques in some crucial places, she decided to start me on a statin. I altered my diet a little, as I was already in a mostly Mediterranean diet. What do I think of CAC scores? If your plaques have not calcified, it will provide false reassurance that all’s well. Until it isn’t. Just to cover all bases, I’m taking early retirement and will hopefully enjoy a stress-free life.

Leave a Reply

Your email address will not be published. Required fields are marked *