Testing for CRP: red flag or red herring?
From the March ACP-ASIM Observer, copyright © 2003 by the American College of Physicians-American Society of Internal Medicine.
By Jason van Steenburgh
Over the past several years, a growing number of internists and cardiologists have begun debating the usefulness of a blood test to measure levels of c-reactive protein (CRP). At issue is whether physicians should use the test as a screening tool to assess patients' risk of cardiovascular disease.
Proponents of the inexpensive, high-sensitivity test claim that CRP, which indicates inflammation, can help identify some at-risk patients that other measurements—like cholesterol and blood pressure—can miss. Critics counter that elevated CRP levels indicate only the presence of underlying disease that treating other risk factors will address.
Now, two recent developments have intensified the CRP debate. One highly publicized study published late last year in the New England Journal of Medicine suggested that CRP levels predict cardiovascular disease better than cholesterol levels alone.
Then in late January, a panel convened by the American Heart Association and the CDC (AHA/CDC) issued the first—albeit cautious—expert guidance on using CRP to gauge patients' cardiovascular risk. While the panel stopped short of recommending CRP screening for everyone, it did find evidence to support using CRP "as an adjunct" to other measurements when assessing coronary vascular disease risk.
Many physicians warn that while testing for CRP can be useful, it should not replace existing heart disease screening methods.
The guidelines disappointed many advocates of CRP testing, who maintain that detecting high levels of inflammation can ultimately save lives. But the guidance reflects the thinking of many physicians who warn that while the test can be useful, it should not replace existing methods to screen patients for heart disease.
For many physicians, however, the new guidance and studies may raise more questions than answers. The biggest question: How should physicians translate "bad" CRP test results into a course of clinical action?
Questions about causation
Since the 1930s, physicians have used CRP levels to measure inflammation in the body. Researchers have long agreed that CRP is elevated in patients with diseases like rheumatoid arthritis and lupus. They also recognize that temporary events such as infection and injury can cause CRP levels to rapidly rise by a factor of up to 100 or more.
Inflammation's role in atherosclerosis and cardiovascular disease, however, is less clear. Researchers know that CRP can accelerate the oxidation of LDL cholesterol, a key element of plaque formation. Data have also shown that elevated CRP levels may help activate macrophage inflammation and adhesion in vessel walls, another dangerous step in forming the kind of plaque that can rupture.
Paul M. Ridker, MD, director of the center for cardiovascular disease prevention at Brigham and Women's Hospital in Boston, is lead author of last year's study and a patent-holder on CRP tests. He said that high levels of CRP may magnify the damage caused by even low levels of cholesterol.
Dr. Ridker's study, which appeared in the Nov. 14, 2002, New England Journal of Medicine, found that women with high levels of CRP were twice as likely to have a heart attack or stroke as women with high levels of LDL cholesterol. He concluded that testing for both CRP and LDL cholesterol levels provided much better risk assessment information than either test alone.
But the exact relationship of inflammation to heart disease is unknown. It is unclear, for instance, whether elevated CRP is a risk factor in its own right, or if it is a marker of risk factors that are already present.
No definitive evidence shows that elevated levels of inflammation cause atherosclerosis, or that higher CRP levels indicate advanced stages of the disease. Even more importantly, there is no clear link between reduced CRP levels and a reduced risk of dying from heart disease.
"There is not good evidence that altering CRP alters risk," said David S. Siscovick, FACP, co-director of the cardiovascular health research unit at Seattle's University of Washington. "I focus my clinical efforts on traditional risk factors—smoking, high blood pressure and cholesterol, physical inactivity, obesity and diabetes—for which there is consensus that we can do more good than harm."
Treating those known risk factors also changes CRP levels, he pointed out. But it is unclear whether patients benefit because they have less inflammation, or if the drop in CRP is a side effect that in itself does not reduce cardiovascular risk.
Because of this uncertainty, many physicians question the point of testing for inflammation to assess heart attack and stroke risks. If the only known ways to lower CRP are to treat already established risk factors, many physicians ask, why bother to measure it at all?
A wider view of risk
But physicians and researchers who support CRP testing say that questions about cause and effect ignore the value of CRP screening in assessing patient risk and deciding treatments.
Russell P. Tracy, PhD, professor of pathology and biochemistry at the University of Vermont in Burlington, said it's not important whether CRP itself actually causes heart disease. Dr. Tracy, who helped redesign the traditional CRP test to make it sensitive enough to detect chronic low-level inflammation, urged physicians to use elevated levels of inflammation—even if they result from conditions like obesity and insulin resistance—much like the Framingham Heart Study global risk score.
Physicians, he said, can tally the total impact of many different factors on patients' health. While you might not be able to pinpoint the reason CRP levels are rising, he explained, you should consider elevated levels a warning that cardiovascular trouble may be just around the corner.
"It's a red flag and a general indicator that a patient's risk factors are not being adequately controlled," said Eric J. Topol, FACP, chairman of the department of cardiovascular medicine at the Cleveland Clinic Heart Center in Clevelandº. Dr. Topol, who has been measuring patients' CRP levels for more than three years, is a strong backer of testing and recommends screening all patients who have what he called "a confluence of risk factors."
"To physicians, high CRP means they have to tighten up their whole approach," he said. "It tells patients that they are basically walking around with increased risk."
Some proponents say the test can be most helpful in primary prevention. Boston's Dr. Ridker, for instance, said he supports CRP screening for a large patient population.
"Half of all heart attacks and strokes occur among people with normal cholesterol levels," he said. "Many of those events occur in people with no real clustering of risk factors at all."
Dr. Ridker outlined one approach that would exclude all patients with very high or low Framingham scores from CRP testing. Patients with high scores, he explained, should already be getting aggressive treatment. Patients with low scores have such a low risk profile that even if their CRP levels jumped, they would still not be treatment candidates.
That leaves a very large intermediate group that consists of healthy, middle-aged patients who have an intermediate risk of heart disease as determined by their Framingham score. Patients in this group found to have elevated CRP levels, Dr. Ridker claimed, should be treated like other high-risk patients: encouraged to modify their lifestyles and considered for medications like statins.
While the new AHA/CDC guidelines are considerably more narrow, they take a middle-of-the-road approach to CRP testing. The guidelines say that CRP testing is an option physicians can use with healthy patients who, based on other assessments, face an "intermediate risk" of heart disease. (The guidelines, which appeared in the Jan. 28, 2003, issue of Circulation, are online.)
The guidelines suggest testing patients who have a 10% to 20% risk of coronary heart disease over 10 years. They also recommend performing the test when physicians need additional information to guide evaluation—such as whether to order diagnostic imaging or exercise testing, for example—or to guide therapy, such as drugs.
The AHA/CDC panel also identified the following CRP risk scale: under 1 mg/L is low risk, 1 to 3 mg/L is average risk, and more than 3 mg/L is high risk.
Interpreting the results
Even physicians who rely on CRP testing say they use screening in conjunction with—not as a replacement for—cholesterol testing. But once you decide to go ahead and test a patient for CRP, how do you interpret the results?
Because jumps in inflammation are common in patients with multiple medical problems, experts say that CRP testing should be limited to healthy outpatients. If you do find huge CRP spikes—a patient jumps from a 2 to a 13, for example—you can probably dismiss it as the result of inflammation from an injury or some other condition. (Certain parasitic diseases such as malaria will also boost CRP levels.)
Some experts say you should be concerned when repeated CRP tests show substantial variation at low levels. Testing proponents say that if a patient who normally has a CRP level of 1.8 jumps to a 3, for example, that patient has significantly elevated risk for heart disease.
Other researchers claim, however, that such a change can be accounted for by normal variation in CRP levels, which some say can vary much more than cholesterol readings. That point brings up another CRP testing controversy: Physicians disagree about exactly what small changes in CRP levels mean.
Some research has found that a person's CRP varies less than 30% over time, while other studies estimate that CRP levels can change as much as 84%. Experts caution that some of these numbers are inaccurate because researchers have included large spikes from temporary conditions that clinicians would ignore. However, even a 30% variability would mean that a shift from a CRP level of 3 to 3.9 would tell you nothing about a patient's risk.
How often do test variations interfere with physicians' ability to interpret results? Again, there is no consensus.
Cleveland Clinic's Dr. Topol, for instance, tests most patients once a year, although he screens patients every six months who have several risk factors or have shown high levels of CRP. He said he hasn't seen as much unwarranted variation in his practice as has been reported in the literature.
Tony G. Farah, FACP, medical director of the cardiac catheterization lab at Allegheny General Hospital in Pittsburgh, said he sees occasional false elevations in CRP readings, but that CRP levels don't vary unreasonably. When he is concerned about variability in patients, however, he measures a patient's CRP every two to three months.
"Serial tests neutralize some of the variability," Dr. Farah explained. "If something is causing the level to be falsely elevated, you want to give it some time to come down."
But other physicians say that test variability is a problem, one that adds greatly to patient stress. They also argue that the costs of serial testing to smooth out variability may not be justified by the test's benefit.
In an editorial accompanying Dr. Ridker's study, Lori Mosca, MD, director of preventive cardiology at New York-Presbyterian Hospital in New York City, suggested that widespread CRP testing is not yet justified because it is unknown if the test is cost effective. She also noted that physicians still don't know if altering treatment strategies based on CRP levels will help patients.
Because CRP readings can vary so much, she said that patients who get tested frequently are on an emotional roller coaster ride. "I call them the 'CRP cripples,' " Dr. Mosca said. "They are going nuts over what CRP means."
For example, the wife of one of her patients with heart disease calls Dr. Mosca every time her husband's level is elevated, wondering if a heart attack is imminent.
Because patients find readings so stressful, Dr. Mosca is very selective about which patients she tests. Many patients come to her wanting a second opinion after they have convinced their doctor to order the test.
What about treatment?
Dr. Mosca does use CRP testing in cases that are "just outside the realm of evidence-based medicine." She said she uses the test as a "tiebreaker."
Patients with LDL cholesterol between 130 and 160 mg/L and two cardiovascular risk factors, for example, don't meet the standard criteria for drug therapy. But if they have an elevated CRP reading as well, Dr. Mosca said she would consider prescribing statins. She would also push those patients harder to make necessary lifestyle modifications.
Dr. Topol of the Cleveland Clinic agreed that high CRP levels give new urgency to his discussions with patients about the importance of lifestyle changes. "It is so easy to fix LDL in most patients with statins that physicians have a false sense of security that they can just ignore the other things," he explained. "For patients with high CRP, physicians need to emphasize lifestyle changes considerably more."
Dr. Topol gives patients with elevated CRP three to six months to start making necessary diet, exercise or smoking adjustments. "If their CRP is still significantly elevated," he said, "it might be time to add a drug."
To reduce CRP, he typically turns to drugs like aspirin, statins, clopidogrel, and rosiglitizone. Dr. Topol also uses drugs that raise HDL like niacin and fibrates.
Dr. Farah of Allegheny General said he encourages patients with multiple risk factors to lose weight, stop smoking and exercise more, regardless of their CRP levels. He also said he treats high CRP levels very aggressively, first with antiplatelet agents such as aspirin and clopidogrel then statins and ACE inhibitors because of their anti-inflammatory action.
Besides using CRP testing to monitor established cases, Dr. Farah said he tests patients who have several risk factors but are otherwise healthy. Not surprisingly, given the controversy about CRP, other physicians disagree with that approach.
Lewis H. Kuller, MD, DrPH, head of the graduate school of public health's center for healthy aging at the University of Pittsburgh, for instance, said he finds CRP readings valuable for deciding treatments only in patients already diagnosed with coronary disease. "In people with atherosclerosis, if you measure their CRP over time, an increase could indicate that a plaque is getting hot or a thrombus is getting very active," he said.
A patient's elevated CRP reading would lead him to use more aggressive anti-inflammatory or lipid-lowering therapy. "However, if you don't already know they have atherosclerosis," Dr. Kuller claimed, "CRP is worthless."
Brigham and Women's Dr. Ridker, on the other hand, argued that a prime benefit of CRP screening is to identify new patient populations that need to be treated. Without the testing, he said, patients with moderate cholesterol levels who are still at risk of a heart attack might fall through the cracks.
While the debate continues, some physicians say that patients have already begun demanding CRP tests, spurred on by media coverage of the new studies and guidelines.
"The public is listening in on the communication among scientists in journals and presentations," said University of Washington's Dr. Siscovick. "Sometimes they are ahead of the physicians, and sometimes they're ahead of the science."
Dr. Siscovick said he looks forward to results of a study Dr. Ridker is now conducting on 15,000 patients with moderate lipid levels and high levels of CRP. Those patients are being randomly chosen to receive either statin treatment or placebo to see if the statins are an effective preventive strategy.
Dr. Mosca, however, said she would rather see a study designed to show that lowering CRP lowers risk. She described the ideal study as one that randomized patients with high CRP vs. low CRP to receive statin therapy or placebo. That study would then see if there is benefit to reducing CRP independent of reducing LDL.
"If there were a drug that lowered CRP without altering other risk factors," Dr. Mosca said, "then we could really test the inflammation hypothesis. But we don't have that drug yet."
The information included herein should never be used as a substitute for clinical judgment and does not represent an official position of ACP-ASIM.
Internist Archives Quick Links
Internal Medicine Meeting Early Registration Discount
Register early for Internal Medicine Meeting 2015 in Boston, MA to lock in the lowest possible rate. Learn more or register now!
Are You Using ACP Smart Medicine®?
This online clinical decision support tool is a FREE benefit of ACP membership delivering point-of-care access to evidence-based recommendations. Includes more than 500 modules, images and reference tables. Start now or watch the video tour.