One year later, cholesterol guidelines sinking in
One option for a more complete profile
The first National Cholesterol Education Program guidelines
for cholesterol testing and treatment, released in 1988, created a
minor sensation, with the public immediately embracing the link they
established between lowering cholesterol levels and lowering the risk
of heart disease.
Following the 1993 release of the second version of the NCEP guidelines,
people were almost as quick to catch on to the notion of "good"
cholesterol versus "bad" cholesterol.
Because heart disease continues to be the leading cause of death
in the United States, and because other cholesterol-related diseases
are reaching epidemic status, the latest revision of the guidelines
known as the Adult Treatment Panel III, or ATP III,1
could prove at least as significant. Released a year ago, it substantially
changed laboratory testing and treatment recommendations, bringing
millions of new patients into the cholesterol-testing fold and expanding
the cardiac risk-factor menu.
For laboratories, it means conducting many more lipid tests, changing
reporting formats, and paying even closer attention to accurate,
standardized lipoprotein measurements, says G. Russell Warnick,
MS, director of the Pacific Biometrics Research Foundation, Issaquah,
Warnick says the 2001 guidelines nearly tripled the number of
patients expected to qualify for drug treatment, increasing it from
13 million to 36 million, or 18 percent of all adults.2
The number of persons qualifying for lifestyle intervention will
rise to 65 million, fully one-third of the adult population. "The
treatment net falls on a much higher proportion of the population,"
Warnick is the former director of the lipoprotein laboratory at
the University of Washington’s Northwest Lipid Research Clinic—one
of 12 research centers in the United States and Canada that participated
in the Coronary Primary Prevention Trial, reported in 1984. This
was the first major study to demonstrate that lowering cholesterol
could lower the risk of heart disease.
"That was really the watershed study that focused attention on
treating cholesterol and led to the formation of the NCEP, which
has been the leader in promoting improvements in lowering cholesterol
and reducing risk," Warnick says.
Studies dating back to the 1950s show cholesterol’s link with
heart disease, says Gerald R. Cooper, MD, PhD, medical research
director of the Clinical Chemistry Branch, Centers for Disease Control
and Prevention. "The thing that really made a difference was the
Lipid Research Clinics Trial. They knew that cholesterol changed
whenever you eat a lot of fat or didn’t watch your diet, but they
did not know whether increased serum cholesterol levels caused an
increase in death. But the Lipid Research Clinics Trial between
1972 and 1982 definitely proved that lowering cholesterol actually
Warnick, who co-edits a series called The Handbook of Lipoprotein
Testing, describes the shifts in NCEP recommendations over the
three reports since 1988. "The major change from 1988 to 1993 was
to elevate the status of HDL [high-density lipoprotein] cholesterol
as a protective or inverse risk factor for coronary artery disease,
and NCEP began recommending measuring HDL whenever total cholesterol
The 2001 guidelines took a new tack by substantially expanding
the range of disease that cholesterol testing can address as well
as the indications for treatment. ATP III recommended categorizing
people with diabetes and multiple risk factors but who did not have
heart disease to the risk level of "CHD risk equivalent." It called
for identifying patients with multiple metabolic risk factors (metabolic
syndrome) as candidates for intensified therapeutic lifestyle changes.
These changes increase the number of patients who qualify for treatment
and mean that much more laboratory testing is likely to be required.
ATP III also calls for a complete lipoprotein profile, as the
preferred initial test, rather than simply screening for total cholesterol
and HDL alone. "Basically what they’re saying is that anybody 20
or over should have a profile consisting of total cholesterol, HDL,
LDL [low-density lipoprotein], and triglycerides once every five
years," says Gary Myers, PhD, chief of the Clinical Chemistry Branch
at the CDC. "And then if total cholesterol is greater than 200,
or HDL is less than 40, a followup lipoprotein profile would be
The guidelines have also increased the number of patients to be
selected for treatment by modifying the classification of lipids
and lipoprotein levels. "There were three big changes," Warnick
says. "The first is that the cutpoint for LDL cholesterol has been
lowered—particularly for LDL levels that would trigger drug
treatment." Identifying LDL cholesterol of less than 100 mg/dL as
optimal in part reflects the success of the statins. These not only
can lower cholesterol, he adds, but can reduce the risks of coronary
disease and may alleviate the risk of type 2 diabetes and Alzheimer’s
disease and mitigate bone loss contributing to osteoporosis.
"So far, the benefits of these drugs have far outweighed the side
effects, so the Adult Treatment Panel was comfortable about pushing
drug treatment into the lower categories," Warnick says.
The second change was raising the risk cutpoint for HDL from less
than 35 mg/dL to less than 40 mg/dL, which the panel called a better
measure of a depressed HDL. The third change—lowering the
risk cutpoints for triglycerides to include moderate elevations—"has
probably not gotten as much attention but is very dramatic," he
says. The triglycerides cutpoint at which followup is triggered
went from 400 in the original guidelines, to 200 in 1993, to 150
"The underlying rationale is that there’s more and more evidence
that high triglycerides are also associated with increased risk
of coronary disease, and triglyceride levels should be treated as
well as high cholesterol levels," Warnick says.
In some cases, both analytes may be elevated, "but there are instances
or conditions that elevate cholesterol without elevating triglycerides
and vice versa," he adds. When both are elevated, "it can indicate
a metabolic syndrome, a new buzzword emphasized more strongly in
the latest guidelines. It’s a somewhat complicated constellation
of symptoms that can include high blood pressure, insulin resistance,
high triglycerides, low HDL, and central obesity, or the so-called
’apple’ shape. So those conditions, as well as diabetes itself,
were given a lot more attention in 2001."
The wider net called for by ATP III is well-founded on the science,
he says. "Studies have shown that treating a person with a moderate
cholesterol elevation is as effective as treating a person with
a high cholesterol elevation. Regardless of initial LDL level, if
you lower LDL there’s a proportionate reduction of risk."
In 1993, treatment was recommended for people who had already
had a heart attack or heart disease symptoms, but "now the cost-effectiveness
of treatment has become better established, so the guidelines justify
treatment over a higher proportion of the population," Warnick says.
Science aside, the time frame for widespread adoption
of the ATP III recommendations is hard to predict. "When the first
guidelines came out," Warnick says, "there was a lot of skepticism
both in the popular press and among many medical professionals who
did not accept the lipid hypothesis or did not accept that high
cholesterol was a major contributor to heart diseases. So there
was some resistance. But over the subsequent 14 years, I think the
skepticism has largely faded."
In 1996, "a group of us did a survey after the release of the
ATP II guidelines,"3 he says.
"We found that about three years after release of the guidelines,
very few clinical laboratories had actually adopted the recommended
cutpoints for all lipids and lipoprotein analytes. But I would think,
as time goes by, laboratories will be somewhat more aggressive in
getting up to speed and implementing the new guidelines."
Warnick notes that professional organizations generally have supported
the NCEP guidelines. Among those supporters have been the American
Heart Association, American Association for Clinical Chemistry,
and American College of Cardiology. "When they put out followup
reports, it has been more in the sense of fine-tuning or updating
the recommendations of NCEP," he says.
Laboratories primarily are hindered by a lack of communication
and inertia, he adds. "Cholesterol, LDL, and HDL are only three
of hundreds of analytes that clinical laboratories have to deal
with." With labs having to keep abreast of so many analytes, "simply
getting up to speed on the information is probably one factor,"
Warnick says. "The second issue would be figuring out how to incorporate
the guidelines into a standard reporting format."
Triglycerides, which now have four cutpoints and five categories
of risk, exemplify the difficulty. "When I go out and talk to laboratory
people," Warnick says, "I ask them, can you actually make these
fit into your laboratory information system? And it sometimes requires
some juggling, streamlining, or simplifying to get complicated cutpoints
to fit into the system."
An NCEP working group informally suggested that laboratories streamline
reports by categorizing patients based on the three treatment targets
of 100, 130, and 160 mg/dL for low-density lipoprotein cholesterol,
rather than the five categories of medical decision value—optimal,
near optimal, borderline high, high, and very high. For patients
at highest risk, treatment is designed to bring LDL-C values below
100 mg/dL, while the target for patients with two or more risk factors
is less than 130, and for all others is less than 160.
Sometimes, Warnick says with a chuckle, the guidelines have become
so complicated that no one can figure out how to follow them to
the letter. One example is the ATP III guidelines recommendation
to use Framingham projections of 10-year absolute CHD risk—that
is, the percent probability of having a CHD event in 10 years—through
a formula based on the patient’s gender, age, total cholesterol,
high-density lipoprotein cholesterol, blood pressure, and cigarette
smoking. "So you can go to the National Institutes of Health and
order a risk calculation and load it into a Palm Pilot or a computer,
or go online and calculate it,4
but the process is somewhat cumbersome to go through, and I’ve heard
some professionals question whether doing the calculations is really
worth the trouble," Warnick says.
The Framingham Heart Study, a well-known Massachusetts research
project that has followed a large sample of patients for more than
50 years, "has provided a huge and long-term base of data," he adds.
"They have not only sampled the original cohort but the offspring
of that original cohort as well, and the statisticians and cardiovascular
epidemiologists associated with the study have been quite astute
in continually mining the data for information."
Nevertheless, while cholesterol has proved to be a powerful marker,
one sign of its limitations is that half the Framingham cohort that
developed heart disease had not been identified with traditional
lipid panels. Total cholesterol is obviously missing a lot of people,
the CDC’s Dr. Myers says. "There are a percentage of individuals
that have what are considered to be normal lipids, or slightly elevated
lipids, but still suffer from some cardiovascular event like a myocardial
infarction. We estimate that 50 percent of people with MI are individuals
who may have normal or slightly elevated cholesterol."
ATP III did not make detailed recommendations about the measurement
of emerging biochemical risk markers. "You can consider the NCEP
Adult Treatment Panel recommendations more or less the tip of the
iceberg," Warnick says. "It is only advocating measurements and
conclusions that have been very well proven over years of research."
ATP III, however, recognized that cardiac risk is influenced by
other factors, such as Lp(a), homocysteine, high fasting glucose,
and prothrombic and proinflammatory markers.
Expert laboratory panels of the NCEP had, earlier, recommended
developing alternative methods to the routine LDL calculation approach
using the Friedewald formula. This method, Warnick says, has been
shown to be reasonably reliable for triglycerides of less than 200
mg/dL and marginal but acceptable up to 400 mg/dL. The diagnostics
industry responded to this by developing fully automated tests for
LDL, the so-called homogeneous assays.
Opinion differs on whether measuring LDL directly is better than
calculating it. A recent study compared four of five new methods
against each other and against the reference method at the CDC,
says Dr. Myers. It found that because there are some specificity
problems, these methods "didn’t really do much better than the Friedewald
estimation that’s been used for years.5
There is pretty much a general feeling that the estimation triad
is quite appropriate and good for triglyceride levels below 300
milligrams," he says.
Warnick points out that a draft policy from the Centers for Medicare
and Medicaid Services could induce more clinicians to choose the
direct LDL-C assay over calculation. CMS is proposing that the number
of reimbursed lipid panels would drop from four to one while the
number of ancillary individual tests would rise to six. The agency,
therefore, would reimburse for any combination of total cholesterol,
HDL-C, LDL-C, or triglycerides up to six tests.
A research review by Warnick and colleagues,6
published this year, found evidence to support using the homogeneous
assays for LDL-C to supplement the Friedewald calculation in cases
where the calculation is known to be unreliable, such as when triglycerides
exceed 400 mg/L. But in routine practice, the authors call for more
evaluation before recommending that the assays replace the calculation.
Atherotech, Liposcience, and Berkeley Heart Laboratories, says
Warnick, are "fairly aggressively promoting" the measurement of
lipoprotein subclasses, especially of small/dense LDL, as superior
measures of risk (see "One option for a more complete profile,"
page 80). Noting that ATP III acknowledges the lipoprotein subclasses
but does not recommend measuring them, Warnick says, "The idea is
that by measuring not just the major lipoprotein classes but also
the subclasses, one can get a better prediction. However, these
tests are relatively new, and the jury is still out on their routine
But other emerging markers that might improve the cardiac risk
detection rate have been a hot topic throughout the standard-setting
world, especially homocysteine and the proinflammatory acute-phase
reactant high-sensitivity C-reactive protein. Recent prospective
epidemiological studies have shown that CRP is a strong and independent
predictor of future coronary events in apparently healthy men and
"It’s quite a controversial question whether the lipoprotein subclasses
are really helpful in improving prediction," says Warnick, maintaining
that the evidence is not yet strong. "If one had to make a choice,
the inflammatory marker is probably more useful than lipoprotein
subclasses. The evidence has been accumulating in recent years that
the CRP may be as important as HDL and LDL in identifying patients
Herbert Naito, PhD, asserts that both markers are helpful in assessing
patient risk. "We need more studies on both to observe the outcome,"
says Dr. Naito, chief of ancillary testing and satellite facilities,
Veterans Affairs Medical Center, Cleveland. "The next stage is whether
there is evidence that if you lower the high-sensitivity C-reactive
protein the outcome is more positive, and there’s some evidence
for that, although we need more."
The CDC and the American Heart Association, he adds, recently
held a conference to develop a position paper on emerging markers,
particularly inflammatory markers, and the paper is expected to
be issued soon.
Much is at stake as the medical community awaits further
consensus on emerging markers. For example, a coverage policy bulletin
on Aetna’s Web site says the "medical literature does not support
the effectiveness" of C-reactive protein, homocysteine testing,
intermediate and small density lipoproteins, or lipoprotein(a) enzyme
Even taking into account the ATP III recommendations, the insurer
maintains that there is insufficient evidence to warrant covering
these tests, either because they are only theoretically linked to
CHD, there are conflicting study results, or the test methods "are
not applicable to clinical practice."
Dr. Myers warns against putting too much credibility into the
favorite "new risk factor of the month."
"I think where we are is we have new risk factors that have the
potential of adding more information for assessing individuals at
risk for coronary heart disease," he says. "These will all be supplemental
to the traditional lipids and lipoprotein profile that’s recommended
by the NCEP ATP. I don’t see any of these new risk factors taking
the place of the traditional ones, but I do see them as possible
adjuncts." The new risk factors would not be used in screening but
would be used to supplement the lipid and lipoprotein profile. "In
other words, if someone presents a certain risk classification,
then we may look at them," he adds.
Dr. Myers emphasizes the need for more clinical trials, however.
"Before these emerging markers can make the transition from research
into clinical practice, we have to have clinical data to support
their associations," he says.
Dr. Naito concurs. "However," he says, "about 18 percent of the
patients undergoing bypass surgery have normal lipid and lipoprotein
profiles. Obviously there are other etiologic factors that are contributing
to the coronary artery disease process." These individuals and other
high-risk persons, such as persons with type 2 diabetes or a family
history of premature deaths due to CAD, "may benefit from assessing
the presence of the emerging risk factors," he says.
Lipoprotein(a) standardization will be the topic of a workshop
sponsored by the National Heart, Lung and Blood Institute later
this year, and the CDC has finished the first phase of a project
to standardize commercial methods for CRP.
While a few cardiologists now order tests like these, they are
not widely performed in routine clinical practice and are still
being transitioned from the medical research environment. Because
of this, and because ATP III lowers and narrows the gap between
cutpoints, the need for precise cholesterol testing has increased,
highlighting the central importance of standardization, according
to many experts. Fortunately, that’s an area in which cholesterol
has been the leader, says Warnick, who has been active in standardization
issues for years. "We don’t try to make everyone use the same method,
but we put programs in place so that hopefully all the methods will
give equivalent results."
It was the Framingham project leaders who approached the CDC nearly
45 years ago about standardizing cholesterol test results, says
the CDC’s Dr. Cooper. "They said the cardiovascular epidemiologists
had been swapping specimens and didn’t agree on the results, and
everybody thought their own answers were right. So they wanted a
neutral laboratory to set up a standardization methodology research
laboratory and a standardization office."
Today, CDC’s cholesterol reference methodology laboratory network
works with manufacturers to calibrate their instruments and ensure
they give accurate results with the right coefficient of variation.
"Right now, over 90 percent of the manufacturers are participating
in that program," Dr. Cooper says.
Having a common cutpoint is necessary "so the whole country is
on the same page," says Dr. Naito. "I’m pleased that CDC is stepping
forward to standardize the country. It’s extremely important from
analytical and chemical standpoints. They’ll also set up guidelines
on preanalytical and postanalytical issues."
But Warnick worries that the emphasis on standardization could
suppress innovation. "CDC has provided an invaluable service in
allowing the entire world to tie into a common gold standard, but
on the other hand, an overly rigid demand that every new test demonstrate
agreement with the reference method could stifle innovation and
basically lock practice into a 30-year-old technology," he contends.
The reference methods have never been proved to fractionate the
lipoproteins in a way that gives the best prediction of risk, he
says. "This has to do with the subclasses and, depending on how
a particular method cuts those subclasses, it can affect the association
with cardiovascular disease," Warnick says. An example is the "convenient
and precise fully automated methods for HDL and LDL, the so-called
homogeneous assays. The assays for HDL have been quite well accepted,
but those for LDL have encountered resistance—and part of
that is they don’t exactly separate the lipoproteins the way the
reference method does," he says.
To address the problem, Warnick has outlined ways manufacturers
could demonstrate that their methods are as good as or better than
the reference methods in predicting risk. "I have long been an advocate
of standardization, but when I talk to people about the concept,
many of them seem to agree that we have been so fixed on standardization
and making everyone agree with reference methods, there’s been a
tendency to throw the baby out with the bathwater, to discourage
access to some of the newer and more innovative methods, or even
to create a disincentive for manufacturers to innovate."
Says Dr. Myers: "So much rests on the measurement of total cholesterol
and LDL and HDL that it’s still very important that we have good
reliable measurement. We can’t take it for granted that we’ve solved
cholesterol measurement problems and everything’s okay. It’s still
something that needs to be watched and maintained. We’re still basing
intervention decisions on a very specific number, and because of
that, reliability, accuracy, and precision in those measurements
are still very important."
As to the impact of ATP III, Dr. Cooper believes laboratories
are well-equipped to handle the new parameters of cholesterol testing.
Managing analytical performance guidelines and finding ways to report
easily interpretable results will be the chief concerns. "I think
with the help of the manufacturers in developing really excellent
equipment, and with the programs like CDC’s and CAP’s and those
of other quality assurance groups, laboratories as well as cardiologists
have been sensitized to expect accurate results," he says. So adjusting
to ATP III "will be a question of replanning what they’re doing.
All it takes is for them to be a little flexible and recognize what’s
going to happen-and get in there."
1. National Cholesterol Education Program Expert
Panel. Third report of the National Cholesterol Education Program
(NCEP) Expert Panel on Detection, Evaluation, and Treatment of High
Blood Cholesterol in Adults (ATP III). NIH Publication. Bethesda,
Md.: National Heart, Lung, and Blood Institute; 2001.
2. Warnick GR, Myers GL, Cooper, GR, et al. Impact
of the third cholesterol report from the Adult Treatment Panel of
the National Cholesterol Education Program on the clinical laboratory.
Clin Chem. 2002;48:11-17.
3. Kroll MH, Cole TG, Rifai N. Standardization
of lipoprotein reporting. Am J Clin Pathol. 2000;114:696-702.
4. See www.nhlbi.nih.gov/guidelines/cholesterol/
5. Miller WG, Waymack PP, Anderson FP, et al.
Performance of four homogeneous direct methods for LDL-cholesterol.
Clin Chem. 2002;48:489-498.
6. Nauck M, Warnick GR, Rifai N. Methods for measurement of
LDL-cholesterol: a critical assessment of direct measurement by homogeneous
assays versus calculation. Clin Chem. 2002;48(2):236-254.
Anne Paxton is a writer in Seattle.