August 19th, 2011
Is Coronary Calcium Better Than CRP for Predicting CV Events?
Larry Husten, PHD
A new study suggests that people with low LDL levels and high CRP levels may benefit from coronary artery calcium (CAC) scans to identify those who are most likely to benefit from statin therapy. In a paper published in the Lancet, Michael Blaha and colleagues analyzed data from 950 people enrolled in the Multi-Ethnic Study of Atherosclerosis (MESA) who met the entry criteria for the JUPITER study.
After 5.8 years of of follow-up:
• 47% of the subjects had a calcium score of zero. CHD event rates in this group were 0.8 per 1000 person-years. They calculated that in this group 124 patients would need to be treated with rosuvastatin to reduce one cardiovascular event (NNT = 124). Overall, 6% of coronary events and 17% of cardiovascular events occurred in this group.
• 28% of the subjects had a CAC between 1 and 100. In this group the cardiovascular NNT was 54.
• 25% of the subjects had CAC scores over 100. This group accounted for 74% of all coronary events. The CHD event rate was 20.2 per 1000 person-years, and the cardiovascular NNT was 19.
The investigators also report that unlike CAC scores, CRP levels did not predict outcome in this population of patients who already had CRP levels ≥2 mg/L.
The investigators concluded:
CAC seems to further stratify risk in patients who meet eligibility criteria for JUPITER, and might be used to target a subgroup of patients expected to derive the most and the least absolute benefit from treatment. Focusing of treatment on the subset of individuals with low LDL cholesterol with measurable atherosclerosis might represent a more appropriate allocation of resources, reduce overall health-care cost, and prevent the occurrence of a similar number of events.
In an accompanying comment, Axel Schmermund and Thomas Voigtländer write that “although definitive proof of treatment effects is scarce, CAC identifies high cardiovascular risk, and statin therapy is most effective in high-risk patients. In our practice, we therefore focus on CAC… for expanded risk stratification in asymptomatic patients.”
To join the panel discussion about this study, click here.
The take home here is that CAC scores might predict who would benefit LESS from statins, and thus not take them since everyone would likely benefit to some degree.
The subject of “competition” between biomarkers is popular, and certainly the favorite from the “xxx trumps yyy” category has been CAC vs CRP. One is a soluble biomarker of inflammation with predictive capacity that is nonspecific, while the other is an imaging assessment of existing arterial calcification, that cannot detect early noncaclfied vulnerable plaque–but picks up “age-related calcium”–and, as Dr Nissen remarked, lacks evidence for effectiveness using outcome improvement data in randomized trials. Each may have theoretical or practical advantages and disadvantages in particular patients, but neither has compelling indications as do, for instance, as with antihypertensive or antianginal agents.
The members of SHAPE tend to recommend CAC in the general population as per their site, http://www.shapesociety.org/, to the point of using CAC to decide whether statin therapy should be initiated, and repeated (I suppose) every 4-5 years when the “warranty” of the exam expires. On the other hand the USPSTF, authors of some papers in the Archives Int Med and others are wary, particularly concerning the radiation exposure (although it is lessening greatly as technology improves).
Although both CRP and CAC are reasonable according to guidelines, if I had a CRP which was high and my CAC were 0 (for which I be most thankful), I’d go further with LDL-P, but if that were indecisive, I would still certainly consider statins to bring down the CRP, unless lifestyle changes could do so, especially if I had more risk factors.
I could anticipate a reduction in the CRP, but no significant change in my CAC. No guarantees, however, either way.
Nonetheless, a high CRP in the face of a 0 CAC score still needs explaining, and carries a poor overall progosis. A 0 CAC does not eliminate the existence of significant lesions, as noted in other studies, and evidenced by the 6% of coronary events and 17% of cardiovascular events reported in the instant trial… unless one is ready to “write” these patients off.
Both CRP and CAC are useful. Although CAC looks directly at the disease itself rather than a surrogate, those lesions are stable. There are no statin trials based on CAC scores. Assertions that doing more CACs in a given practice virtually eliminates events among those patients needs trial vertification. This is unlikely simply because the follow-up to a positive CAC is intensive statin therapy, which still leaves a considerable number (70-80%)of events to unfold. This residual risk is the elephant in the room, and the second elephant is poor adherence. Their quantitative importance may exceed any difference between add-on methods to stratify those with FRS intermediate risk.
Until hard outcome CAC RCTs are available, arguments will continue, depending upon the agenda and personal preference of the physician.
The questions posed in Dr. Kones’s post have mostly been answered. Coronary calcium measurement adds greatly and incrementally to the risk determined by conventional risk factors by a factor of 10. (MESA) The same study showed that baseline CRP was higher in the subset of patients who did not have an event during the study than it was in the subset of patients who did have an event during the study.
Coronary calcium adds great predictive value to CRP however CRP adds no predictive value to coronary calcium.
Although one small study may find 6% of events occurring in subjects with a 0 calcium score, most studies demonstrate a much lower rate of events associated with a 0 score. However,compare the 6% event rate associated with the worst study on coronary calcium with a 66% event rate in women considered to be low framingham risk and coronary calcium begins to look superb indeed.
How does one tell if the calcium seen in coronary vessels is stable or not, easy, repeat the scan in a year or two. If the scan is stable, the plaque is stable and safer, if the plaque is increasing, the plaque is unstable and of greater risk. This has been shown in observational as well as prospective studies.
This is not about agendas or personal preferences, and to state this as such is insulting to the investigators who have worked hard to acquire the necessary information. This is about exhaustive review of the literature. The answers are there, we simply need to read.
The radiation dose for coronary calcium imaging is and has always been exceedingly low. EBT calcium, the initial technology, exposes the patient to 0.7msv of radiation, or about 1/20th the radiation of a nuclear stress test. The radiation argument is a smoke screen for those trying to cloud the issues.
Competing interests pertaining specifically to this post, comment, or both:
I work at and have an ownership interest in a facility that can do EBT calcium imaging and carotid ultrasound as a basis to identify individuals at risk for heart attack and stroke.
Clarification to above: True 6% of all events occurred when CAC=0. However the CAC=0 group is large (50% of most asymptomatic cohorts). The corresponding event rate is <1 per 1000 person-years, corresponding roughly to a 1% 10-year event rate. Very incorrect to state that "event rate" is 6%.
Competing interests pertaining specifically to this post, comment, or both:
Thank you Dr. Blaha for the statistical correction.
Let me modify my comments based upon Dr. Blaha’s appropriate criticism of our distortion of statically science.
Comparing a study where 6% of heart attacks occur in subjects with a 0 calcium score, resulting in an annualized risk of0.1% with other studies showing that 66% of heart attacks occur in women considered to be “low risk” by conventional risk factors. Coronary calcium looks rather good.
CRP<1 comes no where close to a 0 calcium score in predicting non-events while CRP<3 does not hold a candle to a calcium score400 for predicting events.
The last paragraph of the last post got lost in the cloud. It should read:
CRP3 does not hold a candle to a calcium score >100 for predicting events. A calcium score > 400 is associated with dramatically increased risk compared to all conventional risk factors plus crp, a calcium score >1,000 is predictive of a 20% annual risk for MI.