College of American Pathologists
Printable Version

  Picking new winners for
  cardiac risk


cap today

The link with peripheral arterial and cerebrovascular disease

November 2001
Cover Story

William Check, PhD Assessing coronary heart disease is becoming an even riskier business these days.

While physicians harbor virtually no doubts about the central role of cholesterol in initiating and promoting CHD, abnormal low-density and high-density lipoprotein cholesterol levels fail to identify all persons at risk. To detect more persons with an elevated risk, researchers have proposed turning to a handful of newer risk factors, ranging from C-reactive protein and lipoprotein(a) to homocysteine and albuminuria.

Many novel analytes appear promising. Indeed, the third and most recent report from the Adult Treatment Panel of the National Cholesterol Education Program, published earlier this year, took note of several "emerging" risk factors, as the panel put it. "This is the first time the guidelines have mentioned pro-thrombotic or pro-inflammatory factors, homocysteine, and Lp(a)," says Herbert K. Naito, PhD, chief of ancillary testing and satellite facilities at the Louis Stokes Cleveland Veterans Affairs Medical Center.

But their place in clinical practice remains tenuous, as pathologists and their clinical colleagues explore the standards of evidence the risk factors will need to meet to become part of everyday patient management.

Today, when cholesterol levels are so well accepted, it is easy to forget the rigorous series of clinical trials that had to be done before skeptics of the cholesterol hypothesis were converted. As recently as 1989, a best-selling book titled Heart Failure contended that trying to reduce blood cholesterol is a huge waste of time and that drugs that lower cholesterol do not always improve or extend lives. In 1990, when the POSCH trial showed that partial ileal bypass surgery reduced cholesterol and overall mortality, it was considered a groundbreaking result. Claude Lenfant, MD, then director of the National Heart, Lung and Blood Institute, said, "What the study does is knock down the critics of the cholesterol hypothesis."

However, POSCH was a symbolic victory in a sense, since intestinal bypass surgery is not feasible to use on a population-wide basis. Only in 1995 was it convincingly demonstrated that a drug that lowered cholesterol could also lower overall mortality. That demonstration came from the 4S trial, which tested simvastatin, an HMG-CoA reductase inhibitor, or statin. In the next three years, three further trials using this new and potent class of drugs repeated the finding of 4S.


Will the newer markers be held to the same demanding standard of proof as cholesterol profiling? Should they be?

Though inflammatory markers and other novel risk factors are "very interesting," says Jerome D. Cohen, MD, professor of medicine in cardiology and director of preventive cardiology programs at St. Louis University, "they need to pass the high hurdles that we have set up for making standards." As for a general recommendation regarding screening for cardiovascular risk factors, Dr. Cohen says, "None of these new or emerging risk factors have made it to a routine list so far. They are not ready for prime time."

No prospective clinical trials have shown that any of the newer risk factors can be altered in a way that has clinical meaning or that the result of a laboratory test can be used to make treatment decisions. In addition, he says, "No laboratories are standardized for these tests, which poses problems."

A researcher who studies one of the novel markers, Lp(a), shares this view. "I don't see any quick use of non-traditional factors in clinical practice for patient risk assessment," says Santica M. Marcovina, PhD, ScD, research professor of medicine and director of the Northwest Lipid Research Laboratory at the University of Washington. "For years to come, classic atherosclerosis risk factors—lipid profile, clinical and family history, obesity—will be used to determine CHD risk," she predicts.

As for acceptance, the newer markers will not be adopted any more easily than cholesterol measurement was. "If anything, I think it will go the other way," Dr. Naito says. "It takes so long to get acceptance of anything new in the medical community." In his view, slow acceptance is a hazard: "For some patients, I don't think we can afford to wait 10 to 20 years to have a simple measurement process."

Dr. Naito's current approach is to use newer markers "cautiously and selectively" in specific populations if an effective treatment is available. "Obviously I don't recommend generalized screening," he says. "But we see patients with normal lipid profiles having heart attacks or coming in for repeat angioplasty. In these patients we need something more to define risk. So we measure homocysteine."

Dr. Naito's laboratory is examining whether to also use Lp(a) in such patients now that it can be quantitated conveniently with a newly available electrophoretic system that he evaluated, which simultaneously measures Lp(a), LDL-C, HDL-C, and very-low-density lipoprotein cholesterol directly on a drop of blood. "The main point," Dr. Naito emphasizes, "is that we can treat abnormal levels of Lp(a) and homocysteine. So if we are searching for additional risk factors, under selected conditions these analytes can be recommended."

While definitive evidence is being collected, other specialists are also selectively measuring some of the newer putative risk markers. Dr. Cohen describes a candidate similar to that targeted by Dr. Naito: a patient who has a family history of heart disease or premature cardiovascular disease and in whom traditional risk factors, such as lipids, smoking, hypertension, and diabetes, are not present. "In this patient, we recommend measuring non-traditional risk factors," Dr. Cohen says. "If you have a genetic defect causing high homocysteine, wouldn't it be nice to know that?" In the general population, novel risk factors would be uncommon. "But in a patient where we have an unexplained family history, for example, we would more often hit a home run," Dr. Cohen says. "I tell medical students and housestaff it's like kicking over the other rocks."

Richard Cannon, MD, head of the clinical cardiology section of the Cardiovascular Branch of the National Heart, Lung and Blood Institute, endorses this approach. "Certainly when I encounter a patient with a strong family history [of CHD] or myocardial infarction at a young age, I start looking for other things," he says. "I might measure homocysteine and maybe Lp(a), although I have not yet started using CRP."

In clinical practice, the laboratory component of coronary heart disease risk assessment has been based for several years on the fasting lipid profile—measured total cholesterol, HDL-C, and triglycerides, and calculated LDL-C. Showing greater confidence in these parameters, the Adult Treatment Panel report, or ATP III, set more aggressive values for treatment initiation and goals. In high-risk patients, the LDL-C goal is lowered from 130 to 100 mg/dL; the threshold for abnormal HDL is raised from <35 to <40 mg/dL. In addition, Dr. Naito says, "If you look at the guideline quite carefully, it holds some interesting surprises." For instance, ATPIII mentions use of very-low-density lipoprotein cholesterol. "A lot of people seem to overlook that," Dr. Naito says. ATP III recommends evaluating VLDL-C specifically in persons with the metabolic syndrome—high insulin, high triglycerides, and obesity. "What they are saying is that plasma triglycerides in these individuals should be <150 instead of <200. So that is a new wrinkle," he notes.


Among the novel or non-traditional risk markers, C-reactive protein is getting the most attention. It's an acute phase protein that is part of the response to endothelial injury due to ongoing inflammation. "Elevated CRP tells you that the subject's atherosclerotic disease is in a very active state, which increases risk for an MI," Dr. Naito says.

Says Dr. Cannon: "It is not universally agreed which marker or collection of markers might best identify inflammation, which is an important component of atherosclerosis. CRP has been most studied, both in patients with CHD and apparently healthy persons." Paul Ridker, MD, of Brigham and Women's Hospital and Harvard Medical School, has led much of this work. (Dr. Ridker is listed as coinventor on patent applications filed for the use of inflammatory markers in coronary artery disease.)

Studies of healthy populations have shown a graded association between CRP levels and risk of heart attack and stroke. Elevated CRP levels are also associated with anatomic evidence of early atherosclerosis in the carotid artery or aorta. "Levels of CRP that appear to identify this risk would generally be considered to fall within normal ranges," Dr. Cannon says, "even below the sensitivity of many older assays." (Recent clinical studies use high-sensitivity CRP immunoassays that are sensitive to 0.01 mg/dL.)

However, Dr. Cannon notes, these are all correlative kinds of arguments. More direct evidence that CRP could have clinical relevance came from a re-analysis of data from the AFCAPS/TexCAPS study, a primary prevention trial that enrolled patients with normal or only mildly elevated LDL-C (<160 mg/dL), patients who would not generally warrant statin treatment under National Cholesterol Education Program guidelines. In AFCAPS/TexCAPS, among all patients randomly allocated to lovastatin, coronary heart disease events dropped by almost 40 percent. However, says Nader Rifai, PhD, associate professor of pathology at Harvard Medical School and director of clinical chemistry at Children's Hospital, who works with Dr. Ridker, "We don't want to put every American adult on statins."

So Drs. Ridker and Rifai and colleagues retrospectively divided the AFCAPS/TexCAPS population based on whether LDL-C was above or below the median for the entire cohort (149 mg/dL) and further subdivided these groups based on whether CRP was above or below the median. In patients with above-median LDL-C, the event rate dropped by 50 percent, irrespective of C-reactive protein level. Lovastatin did not decrease events significantly among subjects with low LDL-C and low CRP. However, in subjects with low lipids but above-median CRP, treatment with lovastatin cut the event rate by 50 percent, as much as in those with high LDL-C (N Engl J Med. 2001;344: 1959-1965). Dr. Rifai points out that the number needed to treat to prevent a single coronary event in those with low LDL-C but high CRP was almost identical to the number in those with high LDL-C—about 50 patients.

"What we learned," Dr. Rifai says, "is that CRP provides a very important tool to help identify those individuals who are not candidates for statin therapy using the current guidelines but who will benefit from statin therapy in terms of lowering their future coronary risk."

These data suggest, Dr. Cannon says, that treating cholesterol is of benefit even in people who have normal population LDL-cholesterol but CRP above the median. "This article is saying that a simple laboratory test can make that distinction for you. And if CRP is low, don't treat." What's needed, he says, is a "proper clinical trial testing the hypothesis that in relatively low-risk people, at least as determined by cholesterol, that CRP can identify a higher-risk group and justify treatment." Such a trial is "in the planning phase," Dr. Rifai confirms.

Referring to a separate, prospective trial (PRINCE) by Dr. Ridker's group (JAMA. 2001;286:64-70), in which statin therapy reduced CRP levels independently of LDL-C, Dr. Naito says, "I think their study using pravastatin to lower CRP is potentially important as a prophylactic endpoint for treatment. Maybe in the future we will use CRP to determine how low we should reduce lipids." It is not unusual to find individuals who have heart disease and a total cholesterol of 190, he notes. "Are we not lowering cholesterol enough?" he asks. "Or are there other markers we are not measuring and treating?" Perhaps when CRP drops to a pre-specified level, which remains to be defined, that will indicate that the inflammatory process is in a quiescent state. "I think CRP is going to turn out to be a very useful marker to determine the status of the inflammatory process going on with atherogenesis in the blood vessel wall," Dr. Naito predicts. "This would be a totally different approach to treatment."

However, the CRP assay is not ready now for general introduction, in Dr. Naito's view. "While there have been several major cross-sectional and prospective studies, additional large-scale and prospective studies are warranted and should include women as well as men," he says. A standardization program to obtain accurate measurements is also needed, says Dr. Naito, as is further information on how to interpret CRP levels, how to define cutoffs, and how to integrate CRP levels with standard lipid
risk factors.

Dr. Cohen agrees that further work is needed before CRP is introduced into general practice. "It is still not part of the treatment algorithm," he says. "If CRP is high and cholesterol low, I will not give a statin. But I may give aspirin." (In fact, he suggests, it may be that aspirin works because of its anti-inflammatory mechanism, rather than its anti-platelet effect.) "Will lowering CRP lower the CHD rate?" Dr. Cohen asks. "We don't know that intervention will be of value." CRP contributes to the basic physiological process in atherosclerosis, and work on it should continue, he notes. "But at present I don't think it has risen to the level of routine clinical use," he says. He, too, believes cutoffs need to be defined: "Before intervention studies are conducted, a definition of 'normal CRP' is needed."

Creating a population of "worried well" is one potential problem of using CRP in low-risk groups, Dr. Cannon says. Persons who believe themselves to be at low risk might get this additional test and the result might indicate increased risk, causing patient and physician concern that might not be justified. "I think we should not use CRP on a widespread basis until we have convincing evidence from a prospective treatment trial," Dr. Cannon says. Unfortunately, he suspects some such testing is already being done, because assays are available. "The test is not the problem," Dr. Cannon says. "The problem is what to do with the information. That is why many people, and I include myself, don't get CRP levels in clinic populations but only in research studies."

Dr. Rifai says knowing how to report CRP is important: "It is somewhat different from other analytes. We don't recommend values reported to the clinician in mass units, such as mg/dL, but in quartile or quintile of risk." But, says Dr. Naito, "Using quartiles or quintiles is not a standard format we use in the laboratory to report or interpret data. It is a totally new approach. So if Dr. Rifai is saying that is the way CRP should be handled, then we need to further educate clinicians in this new way. If we can make it simpler, it would be better for medicine."


Turning to homocysteine, Dr. Naito suggests testing may be useful in select populations. "For now, I would target several groups documented potentially to have high homocysteine levels," he says. Such groups include:

  • elderly persons, who may have inadequate intake of folate and vitamins B6 and B12.
  • diabetic patients.
  • uremic patients on long-term hemodialysis.
  • perhaps hypothyroid patients.

In persons who present with thromboembolic events, homocysteine is 40 percent to 50 percent higher, Dr. Naito says, and it is seven to 10 percent higher in postmenopausal women. A recent publication documented that patients who have had a heart attack and are not treated for elevated homocysteine levels are almost three times more likely to have a recurrent heart attack (Circulation 2000;102:605).

Homocysteine is easier to measure now that commercial HPLC kits are available, Dr. Naito notes. However, he adds, "No one has standardized homocysteine assays yet. This must be done."

Detecting elevated levels and controlling them is important. "Evidence is increasing to suggest that we need to treat elevated homocysteine, especially since treatment is so easy and relatively cheap," he says. Folate supplementation costs $3 to $4 per month; vitamin B6 or B12 supplementation is also inexpensive.

Dr. Cohen describes a 40-year-old physician who presented with chest pain but didn't have classic risk factors. "We couldn't understand why he had disease," Dr. Cohen says. "We found his homocysteine was high, so now we have him on folic acid. But that is a rare occurrence." Even so, he finds that many physicians are taking folic acid prophylactically, just as physicians were in the vanguard of prophylactic low-dose aspirin use.

A large trial of homocysteine lowering (VA HOST) is being conducted by the VA medical centers, in which patients with renal disease are treated with folate plus vitamins B6 and B12 to see if cardiovascular events are reduced.


Lp(a) varies greatly within populations, ranging from close to zero to several hundred nanomoles/L. While Lp(a) has been associated with coronary heart disease risk in population studies, Dr. Marcovina says, there is no defined threshold now that consitutes risk high enough to warrant lowering, a situation she says applies to all novel markers. "So this creates already a clinical problem," Dr. Marcovina says. "There is no clear idea on how to use them in laboratories.

"Data from the Quebec Heart Study clearly showed that higher Lp(a) levels potentiated the risk of high LDL and counterbalanced high HDL levels," Dr. Marcovina says. But it is not yet known whether persons with normal LDL-C and elevated Lp(a) have increased coronary heart disease risk.

To complicate matters, Lp(a) is difficult to measure; validated assays suitable for the clinical laboratory have been slow in coming. Dr. Marcovina has an NHLBI contract to standardize the Lp(a) assay. "Maybe the recommendations we make will improve the state of the art of Lp(a) measurement," she says. "Only then can we do clinical measurements to set up meaningful ranges."

Lp(a) can play a role in select situations, such as in a patient with family risk for coronary disease and normal or only moderately elevated LDL-C. "If Lp(a) is very low, perhaps a physician would be more likely to prescribe diet and exercise and wait to do drug therapy," she says. She acts as a consultant to physicians in these cases, but says, "This is not the role of a general laboratory."

Therapy for Lp(a) is problematic. Even the most effective Lp(a)-lowering drug, niacin, reduces this lipoprotein only by 30 to 40 percent. And no drug lowers Lp(a) alone. Niacin also raises HDL-C and lowers LDL-C, which makes it difficult to know whether the decrease in Lp(a) contributes to regression of coronary artery disease. A study was conducted at the University of Washington several years ago in which patients whose LDL-C and Lp(a) were both very high were treated with statins. "Patients whose LDL-C was lowered substantially had beautiful regression [of CHD]," she says, "even though their Lp(a) remained very high." One possible way to resolve this question would be to randomly assign patients to treatment with a statin or a statin plus niacin. "This may be done, but it would require a very large number of subjects," Dr. Marcovina says.

The ability to use the markers of oxidative stress in a clinical setting is also some time away, says Sridevi Devaraj, PhD, assistant professor of pathology and associate director of chemistry at the University of Texas Southwestern Medical Center, who does research on this topic. "Measurement of LDL oxidizability remains in research settings," Dr. Devaraj says. "I don't think there are assays that are well standardized." Several possible indicators of oxidative stress are being explored.

Compelling data exist for a direct measure called F2-isoprostanes. "Isoprostanes can be measured in plasma but are relatively unstable," Dr. Devaraj says. "But they are very stable in urine, and validated assays for measuring urinary isoprostanes are commercially available."

Some laboratories have made ELISA assays to measure autoantibodies to oxidized LDL, which is the most atherogenic form of this lipoprotein. It is taken up by monocytes, which are transformed into foam cells and accumulate in plaque. In coronary heart disease patients, autoantibodies to oxidized LDL are increased; in several trials they have correlated with the degree of atherosclerosis. Dr. Naito says there is considerable solid evidence in the literature that oxidized LDL plays a crucial role in CHD, but determining how and what to measure is difficult. "CRP data interpretation to assess CHD risk poses an even greater challenge," he adds. "What would seem prudent and practical at this stage is for an individual to eat a healthy diet rich in vitamins A, E, and C."

Another measure of oxidation is total antioxidant capacity or oxygen radical absorbance capacity. Commercial assays are available for both analytes in plasma. There is considerable evidence for the prognostic value of these indicators, Dr. Devaraj says.

When it comes to remedying elevated oxidation states in patients, at least six large clinical trials have been conducted to test the antioxidant vitamin E, of which four have provided evidence of efficacy, Dr. Devaraj says. It has worked only in trials in which doses of 400 or 800 mg have been used. In CHAOS (Cambridge Heart Antioxidant Study), a 77 percent reduction in myocardial infarction was seen; similar benefits occurred in SPACE, which tested secondary prevention in patients with end-stage renal disease and heart disease. "So evidence is emerging to show that vitamin E at high doses will probably be effective, especially because it has been shown to have not only antioxidant effects, but also anti-inflammatory effects at high doses," Dr. Devaraj says.

Despite the promise of these novel markers in guiding therapy, Dr. Cannon suggests, "For now our major effort should be to encourage the use of medications in patients in whom evidence shows they are of benefit."

We should also aim to treat the current risk factors adequately, which one expert describes as still a "struggle." "Even with proof of the efficacy of intervening on lipids, we struggle to perform enough lipid profiles," says Alan T. Hirsch, MD, associate professor of medicine at the University of Minnesota Medical School. "And it takes considerable health care resources. I wonder if we are ready to expand that to new risk factors where therapeutic interventions are not yet proved."

Dr. Hirsch would like to see "our community do more intervention in patients with proven risk profiles and in whom we know what to do."

"I find it intriguing but hardly mainstream," he says, "to consider inflammatory markers, when evidence for long-term treatment is still in development."

William Check is a medical writer in Wilmette, Ill.