Predicting CV risk among the healthy: Is there a 'best' method?
The phrase “a ticking time bomb for a heart attack” is commonly used to describe the middle-aged obese smoker who never exercises and has a family history of cardiac disease. Nonetheless, it is not uncommon for healthy, normal-weight individuals to suffer a myocardial infarction (MI) or other cardiovascular (CV) event without warning.
Indeed, it is now accepted that some individuals who appear to be healthy actually carry a higher than average CV disease risk, thus raising the question: What is the best method for differentiating the “at-risk” healthy from the low-risk healthy.
MedWire reporter Lauretta Ihonor spoke to Michael Blaha, lead investigator on a recent MESA (Multi-Ethnic Study of Atherosclerosis) analysis, and to Paul Ridker, chief investigator of JUPITER (Justification for the Use of Statins in Primary Prevention: An Intervention Trial Evaluating Rosuvastatin), about potential ways of identifying CV risk in healthy individuals.
Michael Blaha is a cardiology fellow at Johns Hopkins Hospital in Maryland, Baltimore, USA.
Paul Ridker is the Director of the Center for Cardiovascular Disease Prevention at the Brigham and Women's Hospital in Boston, Massachusetts, USA. He was the principal investigator of the Justification for the Use of Statins in Primary Prevention: An Intervention Trial Evaluating Rosuvastatin (JUPITER).
At the heart of the matter lies a paradigm shift
Michael Blaha (Johns Hopkins Hospital, Baltimore, Maryland, USA) and team presented an analysis of the Multi-Ethnic Study of Atherosclerosis (MESA) at the American Heart Association (AHA) Scientific sessions in late 2010.
The findings indicate that measuring coronary artery calcium (CAC) levels, as a marker of subclinical atherosclerosis, may offer great predictive accuracy for CV risk in healthy individuals.
Computed tomography-based CAC scanning
- measures calcification in atherosclerotic plaques found in the coronary arteries
- detectable early in the atherosclerotic process
Measuring CAC may therefore offer a means of determining which healthy individuals “could actually expect to accrue a benefit from statin treatment,” Blaha claims.
But he says that such an approach will require a “paradigm shift” toward medicating only patients who have measurable disease.
His team analyzed CV event rates among 1451 CV disease-free MESA participants with low levels of low-density lipoprotein (LDL) cholesterol (<130 mg/dl, 3.36 mmol/l) but elevated levels of the inflammatory marker high-sensitivity C-reactive protein (hsCRP, ≥2 mg/l).
The researchers intentionally recruited a population similar to that of the 2008 JUPITER study, which assessed statin therapy in asymptomatic patients with high levels of hsCRP.1
Blaha explains: “These patients were selected because they represent the patient subpopulation in which there is currently the most doubt about whether to treat aggressively with medications. Current clinical guidelines do not necessarily support treatment of most of these individuals.”
In this MESA subgroup, 53 CV events occurred over a median follow-up period of 6 years, and the frequency of CV events correlated with CAC score, with 60.4%, 22.6%, and 17.0% of events occurring in participants with CAC of >100, 1–100, and 0, respectively.2
Blaha reported that only 47% of patients with elevated hsCRP levels had a CAC score of 0.
“The key here is targeting the testing at the appropriate individuals. The added time and cost will be well worth it in patients in whom the testing will directly change clinical decision making.” Michael Blaha
An elevated hsCRP level was not linked with a higher CVD event risk than a normal hsCRP level, but CAC scores of 1–100 and greater than 100 were associated with respective 1.7- and 9.0-fold increased risks for CV events, relative to a score of 0, irrespective of hsCRP status.
Blaha stresses that “CAC scans are easy to do. They take just a few minutes, do not require contrast injection, and do not require medications to slow the heart rate.”
He says: “The key here is targeting the testing at the appropriate individuals. The added time and cost will be well worth it in patients in whom the testing will directly change clinical decision making.”
Blaha notes, however, that before screening via CAC scanning can be implemented clinically, the question of who to scan still needs to be addressed.
“I do not think everybody should be scanned,” he remarks, adding that patients with a known risk for CV disease are a good example of individuals in the general population who do not need CAC scanning.
This, he says, is “because in general those patients [who have a known high CV risk] should be treated with medications. CAC will not change decision making when the pre-test probability of disease is very high.”
Blaha explains that “the patients that benefit from CAC are those in which treatment decisions (for example with statins or aspirin) are unclear. This is in general patients at low to intermediate risk [for CV events].”
The shortcomings of CAC scanning
“The only question that matters is this: ‘is there evidence that individuals identified by the biomarker of interest benefit from a therapy they otherwise would not have received?’” Paul Ridker
Conversely, JUPITER lead investigator Paul Ridker (Brigham and Women's Hospital, Boston, Massachusetts, USA) expresses doubt over the clinical value of CAC-based screening.
He says: “I would certainly hope that an expensive imaging test does a better job predicting risk than some simple biomarkers like LDL, high-density lipoprotein, and hsCRP. But remember, that is not the relevant question to even ask in primary prevention. The only question that matters is this: ‘is there evidence that individuals identified by the biomarker of interest benefit from a therapy they otherwise would not have received?’
“The clear answer for LDL cholesterol and for hsCRP is ‘yes’ and the clear answer for CAC is ‘no.’ So in my opinion, there is very little reason to ever get a CAC scan in primary prevention because we have no evidence that the CAC scan will change our behavior or improve patient outcomes.”
Ridker also states: “CAC scans radiate the patient and cause considerable expense for the healthcare system. Because there is no evidence that CAC scans will impact positively on patient outcomes, we do not recommend them for our patients at the Harvard hospitals.”
Blaha, however, argues that CAC is a powerful means of risk stratifying outwardly healthy patients, “but it [will] require an important paradigm shift in this generation of physicians.”
He says: “We have reached the point where we can directly measure atherosclerosis; we no longer have to guess whether a patient has the disease. To me, it makes good sense to focus [pharmacologic] treatment on those who have a measurable quantity of the disease we propose to treat.”
By also identifying individuals without subclinical atherosclerosis, CAC scanning helps to avoid medication and repeat imaging of patients who do not need it. These patients, Blaha states, could still minimize their future risk by adopting lifestyle changes that benefit CV health.
Blaha stresses that it is nonsensical to forgo using an accurate predictive tool in favor of less precise, but easily determined, variables.
He remarks: “Risk factors [such as hsCRP] used alone as predictors [of CV risk] made more sense in an era where we couldn’t measure subclinical disease.”
Inflammatory markers: An alternative approach?
But Ridker remains unconvinced.
He says that “inflammation is a major determinant of heart disease that has until recently been largely ignored, particularly among clinicians in Europe.”
He adds: “More than 50 prospective studies consistently confirm that the magnitude of independent risk associated with elevated levels of the inflammatory biomarker hsCRP is at least as large (if not larger) than that of cholesterol and blood pressure. So continuing to ignore inflammation is a major mistake for patient care.”
Indeed, analysis of data from the JUPITER study show that elevated hsCRP is a highly accurate predictor of CV risk in asymptomatic individuals.3 Therefore Ridker surmises that an expensive test, like CAC scanning, which carries a risk for adverse effects, is not at all necessary.
JUPITER involved 17,802 apparently healthy men and women with a low CV disease risk according to Framingham risk score. They had a LDL cholesterol level of less than 130 mg/dl and no history of CV disease, but an elevated hsCRP of at least 2 mg/l.
All participants were randomly allocated to receive rosuvastatin 20 mg daily or placebo, and were followed-up for 4 years (or until 520 CV events occurred).
By 1.9 years of follow-up, however, the study was terminated due to what the researchers described as “an extreme level of statin benefit.”
Results showed that at 1 year, rosuvastatin therapy significantly reduced hsCRP levels by 37%, MI risk by 55%, and stroke risk by 48%, compared with placebo.
Ridker describes the study as “a fantastic win for patients and a crucial study to get physicians to understand that those with low LDL but high hsCRP are in fact at high risk [for CV events] and greatly benefit from treatment.”
He adds: “In a variety of follow-up studies, we have also shown that those who achieved low levels of LDL cholesterol and low levels of hsCRP got the greatest clinical benefit, that the JUPITER strategy is not just cost-effective but actually cost-saving for society, and that the magnitude of benefit is directly related to the underlying levels of hsCRP. In other words, the more inflammation a patient had, the greater the benefit.”
Despite this evidence, the use of hsCRP in a CV risk-prediction capacity is not advocated by all. Several researchers have reported study results that refute an association between elevated hsCRP and increased CV event risk.
For example, a post-hoc analysis of the ASCOT trial (ASCOT-CRP; Anglo-Scandinavian Cardiac Outcomes Trial), presented at the 2010 AHA Scientific Sessions in Chicago, Illinois, USA, indicated that hsCRP measurement does not offer much improvement in CV risk classification beyond that obtained with the use of traditional Framingham risk factors.4
ASCOT investigator Peter Sever (Imperial College London, UK) reported that a low hsCRP (<1.83 mg/l) was no more likely to be significantly associated with a lower 6-month risk for CV events than was a high hsCRP (>1.83 mg/l). Furthermore, among the study participants allocated to receive atorvastatin, a significant reduction in LDL cholesterol, but not hsCRP, corresponded to a significant drop in CV event risk at 6 months.
This observation led Sever to conclude that “neither baseline nor on-treatment CRP provide useful information about the efficacy of statin treatment to reduce CV events beyond LDL-cholesterol reduction.”
He suggested instead that LDL cholesterol remains the best guide to determining if a person will benefit from statin treatment.
The ASCOT study data are as yet unpublished; however, other published studies have echoed the deductions made by Sever and the ASCOT team.
For example, in an analysis of the Heart Protection Study (HPS), published in The Lancet in February 2011,5 the risk for CV-related death, MI, stroke, or revascularization (composite or as separate components) did not vary with baseline CRP concentration, and treatment with simvastatin 40 mg reduced overall CV event risk irrespective of baseline CRP.
Ridker maintains that the differing designs of these studies from that of JUPITER limit the extent to which they can be used as a counterargument against hsCRP use in risk stratification.
The art of compromise
“I think the JUPITER study in general reflects the sort of patients that benefit from CAC, although clearly CAC predicts risk well regardless of CRP status,” says Blaha.
He reports that MESA findings showed that “50% of JUPITER patients have no CAC, have extremely low event rates, and are unlikely to benefit from aggressive medical therapy.”
“The benefit of CAC is that it identifies who is likely to benefit from treatment and also very importantly, who will likely not benefit from treatment in the short term.” Michael Blaha
Thus, CAC may help to risk stratify and estimate treatment benefit in populations with individuals at high risk for CV events.
“The benefit of CAC is that it identifies who is likely to benefit from treatment and also very importantly, who will likely not benefit from treatment in the short term,” explains Blaha.
He adds: “Identifying people with no CAC has tremendous public health implications, because this is a group of patients who do not need costly medications and repeat imaging studies, and in whom we can focus on cheap lifestyle modifications which will lower the lifetime risk of developing advanced atherosclerosis.”
The next step
“It is a very exciting time in cardiology as we watch the core concepts of inflammation translate into real clinical benefits for our patients.” Paul Ridker
Since the publication of JUPITER, a few guideline updates, such as those of the Canadian CV society,6 have included the recommendation for primary CV disease prevention in individuals without traditional CV risk factors, namely, elevated levels of LDL cholesterol.
However, this strategy is yet to achieve widespread clinical adoption, reflecting that the task of predicting CV risk in healthy individuals is far from complete.
Ridker reports that “with JUPITER completed, we are now embarking on a series of trials designed to address whether reducing inflammation per se reduces cardiac risk... It is a very exciting time in cardiology as we watch the core concepts of inflammation translate into real clinical benefits for our patients.”
Whereas Blaha concludes: “Our MESA study should increase awareness of the importance of a finding of a CAC score of 0... Our study should push the field even harder for a randomized controlled trial showing the value of CAC in guiding treatment decisions.”