November 9th, 2011

Obstructive Lesions Found in Patients with Calcium Scores of Zero

A small but significant number of symptomatic patients with calcium scores of zero have obstructive coronary artery disease (CAD), according to a new study published in the Journal of the American College of Cardiology.

Todd Villines and colleagues studied 10,037 symptomatic patients without known CAD enrolled in the CONFIRM (Coronary CT Angiography Evaluation for Clinical Outcomes: An International Multicenter) registry who underwent coronary CT angiography (CCTA). Half of the patients (51%) had a coronary artery calcium (CAC) score of 0 and although most (84%) had no CAD, 13% had a nonobstructive stenosis, 3.5% had a stenosis of 50% or greater, and 1.4% had a stenosis of 70% or greater. Some 3.9% of patients with a CAC score of 0 and obstructive CAD had an adverse event during follow-up as compared to 0.8% in the group with a CAC score of 0 and no obstructive CAD.

The authors write that their results reaffirm “the importance of properly assessing patient pretest probability for obstructive CAD” when calcium scores are used in symptomatic patients. They conclude that “the absence of CAC reduces but does not fully eliminate the occurrence of obstructive CAD” and that calcium scoring “does not appear to offer significant incremental prognostic information when combined with clinical risk factors and CAD severity on CCTA.”

7 Responses to “Obstructive Lesions Found in Patients with Calcium Scores of Zero”

  1. Stephen Fleet, MD says:

    This is why calcium scoring is not used for symptomatic people, for example, in the ER.

    Better for more general risk stratification in asymptomatic people, where the Framingham score may not be sensitive enough.

  2. Amit Khera, MD, MSc says:

    In my opinion, there are two ways to view this study: 1) From a purely imaging perspective, if a symptomatic patient undergoes a CTA, the presence or absence of CAC in those with obstructive CAD does not add prognostic information to the extent of CAD on the CTA 2) A small proportion of symptomatic individuals who undergo a CTA will have obstructive CAD. While this observation is not surprising and the actual percent will vary depending upon the sampled population, perhaps a more important observation is prognostically how well patients with CAC score of 0 do in the short term, even when symptomatic. My back of the envelope math provides a figure of 0.9% event rate in the next two years for the entire group with CAC=0, with no difference in death, and with a significant proportion of these events being revascularizations. Although the hazard for events is higher in those with CAC score of 0 and obstructive CAD, the actual MI rate was only 1% over 2 years (2 out of 177).

    There are a few important additional points—again these are short-term outcomes. Also, those who were older, diabetic, smokers, and with a family history had a higher probability of obstructive CAD (and maybe worse outcomes?). Thus, one could conclude that in symptomatic outpatients with a low-to-intermediate pretest probability, a CAC score of 0 portends a good short term prognosis. However, the value of CAC score of 0 for long term outcomes in symptomatic outpatients is unclear, and a zero score may not be adequately reassuring in symptomatic patients with a higher pretest probability.

  3. JOSEPH BARRY, BS, MD says:

    Once again a test that I thought resolved the issue of the presence of coronary plaque disease, like all othe tests, is flawed. Maybe flawed is the wrong adverb. The write up by Dr.Khera is confusing. Simplistic thoughts like mine are; The CAC score is no guarentee that my patient with no a score of zero has no plaques or plaque mimic, that will resulti an MI

  4. Michael J. Blaha, MD MPH says:

    It is critical to note that this was a study of symptomatic patients. This important point was missed by most press releases for this study. Symptomatic patients are much different that asymptomatic patients, with a much higher pre-test probability of finding obstructive coronary disease.

    From an epidemiologic perspective, if a researcher were to try to find patients with zero coronary calcium but with obstructive disease, where would they look? They would look for a population of patients similar to this study — symptomatic, CAC=0, and referred by their doctors for further testing due to increased risk of finding obstructive CAD.

    Despite this, coronary artery calcium performed extremely well. A total of 51% of these symptomatic patients had CAC=0. The negative predictive value for finding any stenosis >50% was 96% for CAC=0 in this study. This is far better than a negative stress echo or a negative myocardial perfusion study. Did all of these patients benefit from the extra radiation, contrast exposure, and cost of CTA? During a median follow-up of 2.1 years, there was no difference in mortality among patients with a CAC score of 0 irrespective of obstructive CAD. The rate of MI for the entire CAC=0 group was very low, ~1%. The composite endpoint was increased when the CTA showed obstructive disease, and no doubt revascularizations are increased when physicians are “aware” of a stenosis.

    The authors demonstrate that CAC scoring did not add incremental prognostic information to CCTA using ROC analysis. However, this does not seem like the most pertinent question. My question is: did CTA add to the area under the curve for CAC when CAC=0? (ref:

    The phenotype of CAC=0 with obstructive soft plaque exists but it is rare. Physicians need to be aware of the phenotype, and in higher risk symptomatic patients a CTA makes great sense. For low risk symptomatic patients, I use a “gatekeeper” approach, performing a CTA only when an initial calcium score is greater than zero.

    For symptomatic patients, the authors remind us that CTA is the preferred test for patients with medium to high risk features. In this population, a calcium score is not needed, just the CTA.

    It remains clear that asymptomatic patients with CAC=0 don’t need a CTA for risk stratification (ref: The phenotype of CAC=0, obstructive soft plaque in this population is exceedingly rare.

  5. Stephen Fleet, MD says:

    Here’s a simple takeaway.

    If someone is asymptomatic and has a calcium score of zero, the prognosis is very good. (Nothing is perfect, though.)

    If someone is symptomatic in a worrisome way, you have to do a physiological test (stress test)or anatomic test (coronary CTA or cath)to sort things out. Don’t do a calcium score for symptoms because an obstructive plaque doesn’t have to be calcified.

  6. Dan Hackam, MD PhD says:

    I agree with Stephen’s take. In addition, a carotid ultrasound will not help sort this situation out either – I have seen a number of patients with zero total plaque area score who have presented with MI or ACS. The disease quite clearly is localized to the coronary circulation and may be non-calcified. On the other hand, both CUS and CAC are good tests for asymptomatic patients for general risk stratification and modification of intensity of treatment.

  7. Finally had time to digest the article. Very interesting and enlightening piece of research!

    A couple of observations.

    1. The divergence in the composite endpoints in subjects with CAC of 0 did not occur until after year 2. I suspect that had the CAC been repeated at year 2, it would have identified a significant number of the subjects who were destined to have an event that it missed in the initial score. AS this study did not do a follow up score, we cannot assume this to be fact however a study to look at this could prevent the need for CCTA is symptomatic subjects with 0 CAC.

    2 In subjects with “atypical pain”, a CAC of 0 was a very strong predictor of non-events and this study might reinforce the value of a 0 CAC in the evaluation of “atypical” chest pain.

    3. Among symptomatic patients with a CAC of 0, the 10 year risk for MI was 1%. This is at the bottom of what Framingham defines as “low risk” and is dramatically stronger predictor of non events than a “normal” stress test. (I am hesitant to look at revascularization as a significant endpoint as it is so subjective and often misapplied).

    4. My read of the DATA would suggest that if we did CCTA on individuals with a CAC of 0 but exhibiting typical symptoms plus any combination of smoking, family history of CAD,or diabetes, we would likely find everyone to be found. Conversely, doing CCTA on subjects with a CAC of 0 but lacking these 3 factors is likely of no incremental value.

    5. I would urge caution in interpreting the fact that individuals with CAC of >0 and normal CCTA had so remarkably few events. As CCTA can find plaque that does not narrow the lumen while conventional angiography cannot, this finding of very low events cannot be extrapolated to subjects with CAC >0 and normal catheter angiography.

    Competing interests pertaining specifically to this post, comment, or both:
    I work at a facility that does EBT calcium imaging and carotid ultrasound. We use atherosclerosis imaging to determine who needs increased primary prevention. We use serial imaging to determine who needs more aggressive preventive strategies.