Author + information
- Roger S. Blumenthal, MD⁎ ( and )
- Rani K. Hasan, MD
- ↵⁎Reprint requests and correspondence:
Dr. Roger S. Blumenthal, The Johns Hopkins Ciccarone Center for the Prevention of Heart Disease, Blalock 524C, Ciccarone Center, 600 North Wolfe Street, Baltimore, Maryland 21287
In a classic scene from the iconic movie Ghostbusters (1984, directed by Ivan Reitman), Dr. Peter Venkman (played by Bill Murray) confesses to Dana Barrett (played by Sigourney Weaver), “I make it a rule never to get involved with possessed people.” However, after Dana starts to seduce him, Dr. Venkman replies, “Actually, it is more of a guideline than a rule!”
Although 90% to 95% of the global population-attributable risk for myocardial infarction has been ascribed to 9 modifiable risk factors (1), clinical risk prediction models for hard events (myocardial infarction, cardiac death) remain suboptimal. In the United States, 40% to 60% of myocardial infarctions and sudden death occur as unheralded first manifestations of atherosclerotic cardiovascular disease (ASCVD) (2).
Current U.S. guidelines for identifying and treating people at increased risk for ASCVD events with proven therapies, such as aggressive lipid lowering and aspirin, are based on the Framingham Risk Score (FRS), which is derived from several generations of long-suffering Caucasian Red Sox fans in Massachusetts—clearly, a unique group. While useful and widely accepted as an office-based risk assessment tool, the Adult Treatment Panel (ATP) III version of the FRS has demonstrated limitations in predicting the risk of a major atherosclerotic disease event, particularly among patients with a family history of premature ASCVD and metabolic syndrome phenotype. Misclassification of risk results in both under- and over-treatment of many persons on the basis of the current ATP III guidelines (3–5).
A number of noninvasive imaging modalities for assessing the degree of subclinical atherosclerosis have been proposed to improve cardiac risk prediction. Professional organizations have sought to evaluate the literature and distill the available data into a summary form that is succinct and conveniently accessible to the busy clinician: hence, the guideline.
In this issue of the Journal, Ferket et al. (6) present a systematic review of the guidelines on imaging of asymptomatic coronary artery disease (CAD) published by major professional organizations between 2003 and early 2010. The authors based their search on the Institute of Medicine definition for clinical practice guidelines, and they limited their selections to guidelines that included recommendations for imaging for primary prevention of CAD in presumably healthy nondiabetic populations. They also performed an assessment of the guideline generation process for each of the included guidelines using the 7-item Rigor of Development domain of the Appraisal of Guidelines Research and Evaluation (AGREE) instrument.
Fourteen guidelines published by the U.S. Preventive Services Task Force (USPSTF), the New Zealand Guidelines Group (NZGG), the American Heart Association (AHA), the American College of Cardiology Foundation (ACCF), the National Cholesterol Education Program (NCEP), the Canadian Cardiovascular Society (CCS), and the Canadian Society of Radiologists (CSR) were included in the review, with AGREE rigor scores ranging from 93% (most rigorous) to 21% (least rigorous). Imaging modalities considered among these guidelines included resting and exercise electrocardiography, stress echocardiography, myocardial perfusion imaging (single-positron emission computed tomography and positron emission tomography), computed tomography (CT) and magnetic resonance angiography, and CT coronary artery calcium scoring (CAC).
The authors found wide variability with regard to consideration of these testing modalities, with most guidelines recommending against or noting insufficient evidence for the majority of noninvasive imaging modalities, with the only positive recommendations made in reference to intermediate-risk or selected higher-risk populations. Logistic regression analysis suggested no relationships between the likelihood of a guideline recommending for or against testing and the AGREE rigor score or the proportion of guideline panelists with reported industry relationships. Hence, industry relationships did not appear to have any bearing on the directionality of guideline content in this exploratory analysis.
The only imaging modality that was addressed by a majority of the included guidelines was CAC, with 10 of the 14 guidelines making specific recommendations about this modality as an adjunct to current risk prediction. Among the intermediate-risk population, 1 guideline (CSR) made a favorable recommendation for use of CAC, 4 guidelines (ACCF, AHA, NCEP, CCS) recommended consideration of CAC, and 3 guidelines (USPSTF, NZGG, ACCF) found insufficient evidence to make a recommendation. The 10 guidelines were unanimous in recommending against CAC among very low- and high-risk subjects.
The authors note that while there were widespread inconsistencies and several low AGREE rigor scores among the included guidelines, there was general support for consideration of CAC among intermediate-risk subjects. After a cogent discussion of the limitations of the present review, Ferket et al. (6) explore the paucity of randomized controlled trial (RCT) data for early detection of CAD, and touch on the challenges in generating such data.
While we agree with the authors that RCT data on this topic is lacking, the challenges and potential pitfalls in pursuing such studies are substantial. These challenges include a large sample size and expense as well as a complicated study design. Most would advocate randomizing the study population to receive the screening test or not and then utilizing various intensities of lipid-lowering therapy based on how the screening test might influence perceived risk.
Should the basis for trial inclusion be the presence of multiple ASCVD risk factors or by FRS criteria? Should lipid-lowering therapy be mandated by the study protocol, and if so, what lipid-lowering algorithm should be used? Should such a study be placebo controlled as opposed to comparing various intensities of lipid-lowering therapy? Given that statin therapy is now generally used aggressively by many clinicians for intermediate-risk patients even without imaging tests, the power to show an incremental gain when adding an imaging test may prove to be limited.
It is likely that most men over the age of 55 years and women over the age of 65 years will have some degree of coronary calcification. Are we then obligated to treat them with at least a low-dose statin and aspirin if they have below average CAC for their age, and use a high dose of a potent statin only for those with Agatston scores >100 or for those with scores >75th percentile for their age?
How will costs and adverse effects of additional testing on quality of life be measured? How does one track the cost and emotional anguish surrounding incidental findings such as noncalcified lung nodules that must be followed up by 2 or more chest CT scans over the next 18 to 24 months to document stability or make a diagnosis of possible malignancy? Conversely, how does one measure the potential increased motivation for lifestyle change and increased compliance with medication when patients can directly visualize advanced atherosclerosis in their coronary arteries? Appropriate design, interventions, and duration will prove critical and costly for such a trial, but necessary. Efforts by the National Heart, Lung, and Blood Institute to address this need are ongoing.
One note of caution is that a suboptimally designed or executed RCT could be very confusing to clinicians and to the public. Recently, the DIAD (Diagnostic Imaging in Asymptomatic Diabetics) trial assessed whether screening of diabetic patients with single-positron emission computed tomography myocardial perfusion imaging could enhance detection of those with high ASCVD risk and whether the detection of this risk was associated with an improvement in clinical outcomes (7).
Although 22% of subjects in this study had an abnormal myocardial perfusion imaging, only 6% of the defects were moderate or large, and there was no ability to detect persons with advanced subclinical atherosclerosis without overt ischemia. As a result, there was only a slightly higher rate of coronary angiography in the group that was screened, but there was no difference in the intensification of secondary prevention measures in the screening group. Ultimately, there was no difference in clinical events between the screened and unscreened arms. One wonders if CAC testing had been employed and identified persons with advanced subclinical atherosclerosis for their age, it would likely have resulted in more appropriate use of aggressive secondary prevention measures, as recently demonstrated by Nasir et al. (8) in a multiethnic population with elevated CAC scores.
In addition to its ability to identify persons with advanced subclinical atherosclerosis for their age, the absence of CAC has been associated with a very low risk of cardiac events over the next 5 years (9). That provides a rationale to emphasize lifestyle changes and scale back on expensive high-potency statins and focus on generic statin therapy if the low-density lipoprotein cholesterol is >130 mg/dl despite improved dietary and exercise habits. Clinicians may also decide to refrain from ordering stress imaging tests in the setting of atypical chest discomfort.
Some have suggested that one should restrict aggressive pharmacotherapy to patients with at least moderate subclinical atherosclerosis to lower the number-needed-to-treat for expensive pharmacotherapy. For example, in the ASCOT (Anglo-Scandinavian Cardiac Outcomes Trial), 93 adults would have needed to be treated for a mean of 3.3 years to prevent a single cardiac event (10). Could the use of atherosclerosis imaging have targeted those patients likely to benefit from aggressive pharmacotherapy and reduced the number-needed-to-treat?
No randomized, controlled, interventional outcome trial evaluating aggressive lipid and blood pressure lowering as well as aspirin therapy guided by the FRS exists. The FRS has become widely accepted as a reasonable approach to CAD risk prediction and stratification in guiding primary prevention on the basis of prospective, observational cohort data. This is certainly not the first instance in which a medical intervention or strategy is widely used or accepted in the absence of RCT evidence of mortality benefit: consider the examples of nitrates for myocardial ischemia, furosemide for decompensated heart failure, oxygen for hypoxemia, or parachutes for free fall (11).
Conservative guidelines often withhold support for a therapeutic or diagnostic intervention in the absence of robust RCT evidence. However, the absence of such evidence is not equivalent to evidence against selective use of a diagnostic test. In the case of imaging asymptomatic CAD, ample prospective cohort and observational data support the consideration of the use of CAC in improving risk prediction in appropriate groups of patients (12,13), as currently reflected by several of the guidelines reviewed by Ferket et al. (6).
In a recently published study of 5,878 participants in the MESA (Multi-Ethnic Study of Atherosclerosis) with 209 CAD events over 6 years of follow-up, CAC correctly reclassified one-half of the subjects deemed at intermediate risk into both higher- and lower-risk categories (14). The statistical methods in this study demonstrated that CAC had an additive effect on risk prediction, with accurate reclassification and improved estimation of risk, with the caveats that this study used the FRS modified to account for ethnicity for initial risk prediction and truncated 10-year risk categories for the 5-year follow-up.
Hence, selective use of CAC in intermediate-risk populations may prove beneficial in avoiding over-treatment of some persons while identifying others who previously would not have qualified for intensive pharmacologic and lifestyle prevention efforts based on FRS alone. There is generally a low risk with and low cost for aspirin and generic statin therapy, and there is a very low radiation risk of a single CT scan for CAC screening (∼1 to 1.5 mSv). Withholding the potential benefits of selective screening of persons at indeterminate risk, such as those with family history of premature cardiovascular disease and at least a 6% risk of a myocardial infarction over the next decade in the absence of RCT evidence of improved mortality, seems overly dogmatic.
Guidelines are intended to help clinicians navigate issues and challenges in patient care by application of the best available evidence. However, they are not dicta to which all care decisions should perfunctorily adhere. In the case of screening for asymptomatic CAD, a double standard exists for what is the currently accepted standard—the FRS—and for novel strategies for risk prediction, as there are no clamors for an outcome-based RCT of FRS-guided interventions.
Given the difficulties inherent in generating robust RCT evidence, judicious application of the available evidence from well-executed prospective, observational cohort studies is needed to continue to improve risk prediction and primary prevention of CAD events. We must avoid the “cognitive dissonance” that often impedes forward progress and confines guidelines to the necessity of the RCT to optimize the care of our patients in light of the available evidence.
In summary, when there's something wrong in the neighborhood of cardiovascular risk prediction, who should we collectively call (upon)? The data in the preceding text argue for thoughtful interpretation of the available evidence as summarized in current guidelines rather than narrow adherence to rules. We think that Dr. Venkman would agree.
The authors have reported that they have no relationships to disclose.
↵⁎ Editorials published in the Journal of the American College of Cardiology reflect the views of the authors and do not necessarily represent the views of JACC or the American College of Cardiology.
- American College of Cardiology Foundation
- Gibbons R.J.,
- Jones D.W.,
- Gardner T.J.,
- et al.
- Berger J.S.,
- Jordan C.O.,
- Lloyd-Jones D.,
- Blumenthal R.S.
- Ferket B.S.,
- Genders T.S.S.,
- Colkesen E.B.,
- et al.
- Nasir K.,
- McClelland R.L.,
- Blumenthal R.S.,
- et al.
- Blaha M.,
- Budoff M.J.,
- Shaw L.J.,
- et al.
- Sever P.S.,
- Dahlof B.,
- Poulter N.E.,
- et al.
- Smith G.C.,
- Pell J.P.
- Scheuner M.T.,
- Setodji C.M.,
- Pankow J.S.,
- et al.