Circulation. 1999;100:1481-1492
(Circulation. 1999;100:1481-1492.)
© 1999 American Heart Association, Inc.
AHA/ACC Scientific Statement |
Assessment of Cardiovascular Risk by Use of Multiple-Risk-Factor Assessment Equations
A Statement for Healthcare Professionals From the American Heart Association and the American College of Cardiology
Scott M. Grundy, MD, PhD;
Richard Pasternak, MD;
Philip Greenland, MD;
Sidney Smith, Jr, MD;
Valentin Fuster, MD, PhD
Key Words: AHA/ACC Scientific Statement • coronary disease • risk factors • risk assessment
 |
Introduction
|
|---|
The past decade has witnessed major strides in the
prevention
of coronary heart disease (CHD) through modification
of its
causes. The most dramatic advance has been the demonstration
that
aggressive medical therapy will substantially reduce the
likelihood
of recurrent major coronary syndromes in patients
with established
CHD (secondary prevention). The American Heart
Association (AHA)
and the American College of
Cardiology (ACC) have published
joint recommendations
for medical intervention in patients with
CHD and other forms of
atherosclerotic disease.
1 A similar
potential exists for
risk reduction in patients without established
CHD (primary
prevention). However, the risk status of persons
without CHD varies
greatly, and this variability mandates a
range in the intensity of
interventions. Effective primary prevention
thus requires an assessment
of risk to categorize patients for
selection of appropriate
interventions. The present statement
is being published jointly by
the AHA and ACC to outline current
issues and approaches to global risk
assessment for primary
prevention. The approaches described in this
statement can be
used for guidance at several levels of primary
prevention; however,
the statement does not attempt to specifically
link risk assessment
to treatment guidelines for particular risk
factors. Nonetheless,
it provides critical background information that
can be used
in the development of new treatment guidelines.
The major and independent risk factors for CHD are cigarette smoking of
any amount, elevated blood pressure, elevated serum total
cholesterol and low-density lipoprotein
cholesterol (LDL-C), low serum high-density lipoprotein
cholesterol (HDL-C), diabetes mellitus, and advancing age
(Table 1
). The quantitative
relationship between these risk factors and CHD risk has been
elucidated by the Framingham Heart Study2 and other
studies. These studies2 show that the major risk factors
are additive in predictive power. Accordingly, the total risk of a
person can be estimated by a summing of the risk imparted by each of
the major risk factors. Other factors are associated with increased
risk for CHD (Table 2). These are of 2
types: conditional risk factors and predisposing risk factors. The
conditional risk factors are associated with increased risk for CHD,
although their causative, independent, and quantitative contributions
to CHD have not been well documented. The predisposing risk factors are
those that worsen the independent risk factors. Two of them—obesity
and physical inactivity—are designated major risk factors by the
AHA.3 4 The adverse effects of obesity are worsened when
it is expressed as abdominal obesity,5 an indicator of
insulin resistance.
|
Clinical Importance of Global Estimates for CHD Risk
|
|---|
Preventive efforts should target each major risk factor. Any
major
risk factor, if left untreated for many years, has the
potential to
produce cardiovascular disease (CVD). Nonetheless,
an
assessment of total (global) risk based on the summation
of all major
risk factors can be clinically useful for 3 purposes:
(1)
identification of high-risk patients who deserve immediate
attention
and intervention, (2) motivation of patients to adhere
to
risk-reduction therapies, and (3) modification of intensity
of
risk-reduction efforts based on the total risk estimate.
For the latter
purpose, patients at high risk because of multiple
risk factors may
require intensive modification of
1 risk factors
to maximize risk
reduction. Guidelines for the management of
individual risk factors are
provided by the second Adult Treatment
Panel report (ATP II) of the
National Cholesterol Education
Program
(NCEP),
6 the sixth report of the Joint National
Committee
(JNC VI) of the National High Blood Pressure
Education Program,
7 and the American Diabetes Association
(ADA).
8 All of these
guidelines are currently endorsed or
supported by the AHA and
the ACC. These reports
6 7 8
advocate adjusting the intensity
of risk factor management to the
global risk of the patient.
In ATP II and JNC VI,
6 7
overall risk is estimated by adding
the categorical risk factors. They
do not use a total risk estimate
based on summation of risk factors
that have been graded according
to severity; this latter approach has
been advocated recently
by Framingham investigators.
2 The
use of categorical risk factors
has the advantage of simplicity but may
be lacking in some of
the accuracy provided by graded risk factors.
Some researchers and clinicians believe that the summation of graded
risk factors provides advantages over the addition of categorical risk
factors. For instance, the use of graded risk factors has been
recommended in risk-management guidelines developed by joint European
societies in cardiovascular and related
fields.9 Advocates of this approach contend that the
increased accuracy provided by the grading of risk factors outweighs
the increased complexity of the scoring procedures. If the Framingham
system is to be used, however, its limitations as well as its strengths
must be understood. The AHA's Task Force on Risk Reduction recently
issued a scientific statement10 that reviewed and assessed
the utility of Framingham scoring as a guide to primary prevention. The
present report expands on this assessment and considers factors
that must be taken into account when the Framingham algorithm is
used.2
|
Primary Versus Secondary Prevention
|
|---|
The present report focuses mainly on risk assessment for
coronary
disease and not on risk for other
cardiovascular outcomes. Framingham
scores estimate
risk for persons without clinical manifestations
of CHD.
2
Therefore, the scores apply only to primary prevention,
ie, to
prevention in persons without established CHD. Once coronary
atherosclerotic
disease becomes clinically manifest, the risk for
future coronary
events is much higher than that for patients
without CHD,
6 regardless of other risk factors, and in
this case, Framingham
scoring no longer applies. The AHA and ACC have
issued joint
guidelines for the management of risk factors for patients
with
established CHD and other forms of atherosclerotic
disease.
1
|
Definition of CHD
|
|---|
Interpretation of risk estimates for CHD requires a precise
definition
of CHD. Framingham estimates traditionally predict total
CHD,
which includes angina pectoris, recognized and unrecognized
myocardial
infarction, coronary insufficiency (unstable
angina), and CHD
deaths. In contrast, many clinical
trials
11 12 13 14 that have
evaluated specific risk-reducing
therapies have specified major
coronary events (recognized
acute myocardial infarction and
CHD deaths) as the primary
coronary end points. In accord, the
recent Framingham
report
2 also provided estimates for "hard"
CHD,
excluding angina pectoris. The inclusion of coronary
insufficiency
(unstable angina) and unrecognized myocardial infarction
(defined
by electrocardiography) probably gives
estimates of hard CHD
that are somewhat higher than combined end points
reported in
several clinical trials.
11 12 13 14 A recent
clinical trial,
the Air Force/Texas Coronary Artery Prevention
Study (AFCAPS/TexCAPS),
15 specified acute coronary
events, including unstable angina,
acute myocardial infarction, and
coronary death, as the primary
end point. This combined end
point probably corresponds closely
to the Framingham study's
definition of hard CHD. Definitions
of coronary end points
assume critical importance when risk
cutpoints are defined to select
patients for specific therapies.
|
Absolute Risk Estimates
|
|---|
Absolute risk is defined as the probability of developing CHD
over
a given time period. The recent Framingham report
2
specifies
absolute risk for CHD over the next 10 years. Although
absolute
risk scores can be used to evaluate preventive strategies, 4
caveats
must be kept in mind. First, Framingham scores derive from
measurements
made some years ago; the possibility exists that absolute
risk
for any given level of risk factors in the general population
may
have changed since that time. Second, absolute risk in the
Framingham
population for any given set of risk factors may
not be the same as
that for all other populations, for example,
those of differing ethnic
characteristics. Third, Framingham
risk scores represent
average values; however, considerable
individual variability in risk
exists within the Framingham
population. For example, several other
factors not included
in the Framingham scores potentially modify
absolute risk for
individuals (see Table 2
). Finally, Framingham
scores are not
necessarily elastic; the magnitude of risk reduction
achieved
by modifying each risk factor may not equal (in reverse) the
increment
in risk accompanying the factors.
|
Definition of Low Risk
|
|---|
The Framingham report
2 defined low risk as the risk
for CHD
at any age that is conferred by a combination of all the
following
parameters: blood pressure <120/<80
mm Hg, total cholesterol
160 to 199 mg/dL (or LDL-C 100 to
129 mg/dL), and HDL-C
45 mg/dL
for men or
55 mg/dL for women in a
nonsmoking person with no
diabetes (Table 3
). This definition of low risk seems
appropriate
and should be widely applicable; for example, in the
follow-up
of 350 000 screenees of the Multiple Risk Factor
Intervention
Trial,
16 most of the excess mortality from
CHD could be explained
by the presence of the major risk factors above
these levels.
The NCEP
6 designated a total
cholesterol level of <200 mg/dL
(or LDL-C of <130 mg/dL)
as a desirable level. Framingham
investigators
2 included
total cholesterol levels in the range
of 160 to 199 mg/dL
(and LDL-C of 100 to 129 mg/dL) in their
definition of the low-risk
state. In addition, NCEP
6 recognized
an LDL-C level of
100 mg/dL as optimal and as the goal of therapy
for secondary
prevention. This level corresponds to a total
cholesterol
level of
<160 mg/dL. An elevated LDL-C level
appears to be the
primary CHD risk factor, because some elevation
of LDL seems to be
necessary for the development of coronary
atherosclerosis.
17 A very-low-risk state
can be defined as an LDL-C level of <100
mg/dL in the presence of
other low-risk parameters (Table 3
).
Therapeutic
efforts to reestablish a very-low-risk state appear
to be justified for
secondary prevention
1 6 ; in primary prevention,
however, a
very low LDL-C level is not currently deemed
necessary.
6
|
Relative Risk Versus Absolute Risk: Estimations From
Framingham Scores
|
|---|
The relative risk is the ratio of the absolute risk of a given
patient
(or group) to that of a low-risk group. Literally, the term
relative
risk represents the ratio of the incidence in the
exposed population
divided by the incidence in unexposed persons. The
denominator
of the ratio can be either the average risk of the entire
population
or the risk of a group devoid of risk factors. The
Framingham
definition of the low-risk state provides a useful
denominator
to determine the effect of risk factors on a patient's
risk.
Both the absolute and relative risk can be derived from the
recently
published risk score sheets.
2
The first step in estimating risk is to calculate the number of
Framingham points for each risk factor (Table 4). For initial assessment, measurements
of serum levels of total cholesterol (or LDL-C) and HDL-C
are required.2 The points for total
cholesterol instead of LDL-C are listed in Table 4
because some of the Framingham database did not include LDL-C. Hence,
total cholesterol gives more robust estimates. Evaluation
for cholesterol disorders requires measurement of LDL-C,
which is also the primary target of cholesterol-lowering
therapy.6 The blood pressure value used in scoring is that
obtained at the time of assessment, regardless of whether the patient
is taking antihypertensive drugs. The average of several blood pressure
measurements is needed for an accurate determination of the baseline
level. Finally, in the present report, Framingham risk scores for
borderline elevations have been modified to assign stepwise incremental
risk in accord with current NCEP6 and JNC VI7
guidelines. Failure of Framingham scores to identify stepwise
increments in risk in borderline zones probably reflects the relatively
small size of the Framingham cohort. Diabetes is defined as a fasting
plasma glucose level >126 mg/dL, to conform with recent ADA
guidelines18 ; in the Framingham study, diabetes was
defined as a fasting glucose level >140 mg/dL. The designation of
"smoker" indicates any smoking in the past month. The total risk
score sums the points for each risk factor.
Risk ratios, relative to the low-risk state (Table 3), are shown
for men in Figure 1 and for women in
Figure 2; for each age, the number shown
gives the relative risk. In addition, 10-year absolute risk values are
shown for both total and hard CHD. The definition of hard CHD is that
used by Framingham investigators; values shown for hard CHD are
approximately two thirds those for total CHD, which are in accord with
the recent Framingham report.2 Gradations of increasing
relative risk are given in color. At the midpoint of this gradation is
the average risk for the Framingham cohort for each age range. Ratios
above average are divided into moderately high relative risk and high
relative risk. A 3-fold increase in relative risk above the lowest risk
level is designated moderately high risk; a 4-fold or greater increase
is called high risk. Absolute risk levels rise progressively with age,
even in the absence of risk factors.
View larger version (92K):
[in this window]
[in a new window]
|
Figure 1. Relative and absolute risk estimates for CHD in
men as determined for Framingham scoring.2 The number of
Framingham points is derived as shown in Table 4. Relative risk
estimates for each age range are compared with baseline risk conferred
by age alone (in the absence of other major risk factors). Relative
risk is graded and color coded to include below average, average,
moderately above average, and high-risk categories. Distinctions in
relative risk are arbitrary. Average risk refers to that observed in
the Framingham population. Absolute risk estimates are given in the 2
right-hand columns. Absolute risk is expressed as the percentage
likelihood of developing CHD per decade. Total CHD risk equates to all
forms of clinical CHD, whereas hard CHD includes clinical evidence of
myocardial infarction and coronary death. Hard CHD estimates
are approximated from the published Framingham
data.2
|
|
View larger version (71K):
[in this window]
[in a new window]
|
Figure 2. Relative and absolute risk estimates for CHD in
women as determined for Framingham scoring.2 The number of
Framingham points is derived as shown in Table 4. Relative risk
estimates for each age range are compared with baseline risk conferred
by age alone (in the absence of other major risk factors). Relative
risk is graded and color coded to include below average, average,
moderately above average, and high-risk categories. Distinctions in
relative risk are arbitrary. Average risk refers to that observed in
the Framingham population. Absolute risk estimates are given in the 2
right-hand columns. Absolute risk is expressed as the percentage
likelihood of developing CHD per decade. Total CHD risk equates to all
forms of clinical CHD, whereas hard CHD includes clinical evidence of
myocardial infarction and coronary death. Hard CHD estimates
are approximated from the published Framingham
data.2
|
|
Relative risk is useful for providing the physician with an immediate
perspective of a patient's overall risk status relative to a low-risk
state. This perspective can be helpful as a frame of reference for both
physician and patient. Moreover, relative risk probably can be used to
compare risk among individuals in populations in which baseline
absolute risk has not been established. Absolute baseline risk
(low-risk level) almost certainly varies among different populations,
but the relative contributions of individual risk factors to total risk
appear to be similar among all populations. Although the comparability
of relative risk has not been proven rigorously, examination of
available data from different epidemiological
studies19 20 21 22 23 24 25 26 27 28 suggests this to be the case.
It is apparent from Figures 1 and 2 that the relative
risk associated with a given set of risk factor levels (expressed as a
single Framingham number) declines with advancing age. At the same
time, 10-year absolute risk rises with aging. Both changes have
implications for prevention. Higher relative risk estimates in young
adults are an indication of the high long-term risk accompanying the
risk factors; they point to the need to institute a long-term
risk-reduction strategy. On the other hand, the increasing absolute
risk that accompanies advancing age reveals the opportunity for
reducing absolute short-term risk by an immediate aggressive reduction
of risk factors in older people. However, the best candidates for
aggressive risk reduction among older patients may be those with
moderately high or high relative risk. Recent guidelines have
emphasized absolute risk estimates for use in treatment guidelines.
Even so, the utility of relative risk estimates for areas of primary
prevention that are most contentious, specifically, in young adults and
elderly patients, should not be overlooked in the development of future
guidelines.
|
Absolute Short-Term Risk
|
|---|
Estimates of short-term risk (absolute risk in the next 10 years)
are
potentially useful for the identification of patients who need
aggressive
risk reduction in the clinical setting. Patients at high
short-term
risk may need pharmacological agents to control risk
factors.
The precise level of absolute risk that defines a patient at
high
short-term risk has been an issue of some uncertainty and involves
a
value judgment. Theoretically, this level of risk justifies
aggressive
risk-reduction intervention and is set through an
appropriate
balancing of efficacy, costs, and safety of therapy. Over
time
and depending on economic considerations, the thinking about
this
critical cutpoint of risk may change. Furthermore, little
dialogue has
occurred in the United States regarding the process
of choosing a
single absolute risk cutpoint for high short-term
risk. The NCEP has
taken the lead in adjusting the aggressiveness
of
cholesterol-lowering therapy to the absolute risk of
patients.
The NCEP identified patients having established CHD and other
atherosclerotic
disease as being at very high risk and deserving of
aggressive
therapy. For primary prevention, LDL-C goals were
established
by counting risk factors, but they did not define absolute
risk
in precise, quantitative terms. Future guidelines for risk
reduction
in the United States likely will put greater emphasis
on quantitative
global risk assessment.
Recently, guidelines of the joint European Societies9 have
identified high short-term risk as an absolute risk that imparts a
>20% probability of developing CHD in the next 10 years. Once a
patient reaches this threshold of risk, guidelines similar to those for
secondary prevention are triggered. This threshold may be reasonable,
but several comments must be made about how the European guidelines
were derived. The authors9 made use of older Framingham
risk equations,29 but their own risk estimates were based
only on age, cigarette smoking, blood pressure, and total
cholesterol. HDL-C levels were not included. Framingham
risk equations2 29 consistently include HDL-C,
which is a powerful independent risk factor. The absence of HDL-C as a
risk factor in European guidelines must be considered a limitation. As
previously mentioned, European guidelines9 used
Framingham's total CHD as the coronary end point,
which is a liberal coronary outcome and lowers the barrier to
initiation of secondary-prevention guidelines. Irrespective of these
details, there appears to be considerable consensus in the European
cardiovascular community that a 10-year risk for
clinical coronary end points of >20% justifies the category
of high short-term risk. One concern about European guidelines is that
although they creatively bridge the gap between primary and secondary
prevention, they seemingly deemphasize the need for long-term primary
prevention in the clinical setting.
|
Absolute Long-Term Risk
|
|---|
Framingham scoring does not directly project long-term risk
(>10
years), although such risk can be approximated by the summing
of
risk scores over successive age categories and the subtraction
of those
persons removed by having CHD events. Thus, 20-year
risk should be at
least twice the 10-year risk. An important
aim of primary prevention is
to reduce CHD over the long term
and not just over the short term. For
a patient in the age range
of 50 to 54 years, a 20-year projection
of absolute risk may
be of more interest to both the physician and the
patient than
a 10-year projection. Such a patient whose 10-year
risk for
CHD is 15% may not qualify as being at high short-term risk,
but
this same patient has a >30% probability of developing CHD
before
age 75. This latter projection needs to considered when
primary
prevention strategies are planned.
Another critical point to make about long-term risk is that any single
coronary risk factor, eg, cigarette smoking, hypertension, high
serum cholesterol, or diabetes, can lead to premature CHD
(or stroke) if left untreated over a period of many years. Therefore,
each of the major risk factors deserves intervention in the clinical
setting, regardless of the short-term absolute risk. The centerpiece of
long-term risk reduction is modification of lifestyle habits, eg,
smoking cessation, change in diet composition, weight control, and
physical activity.30 Nonetheless, in patients in whom
long-term risk is high, the use of drugs for treatment of hypertension
or serum cholesterol disorders may be warranted, as
described in JNC VI7 and ATP II,6
respectively.
|
Severity of Major Risk Factors
|
|---|
Framingham scoring takes into account gradations in risk factors
when
estimating absolute risk. The scoring does not adequately account
for
severe abnormalities of risk factors, eg, severe hypertension,
severe
hypercholesterolemia, or heavy cigarette
smoking. In such cases,
Framingham scores can underestimate absolute
risk. This underestimation
is particularly evident when only 1 severe
risk factor is present.
Thus, heavy smoking
31 or
severe hypercholesterolemia
32 can
lead
to premature CHD even when the summed score for absolute risk
is
not high. Likewise, the many dangers of prolonged, uncontrolled
hypertension
are well known. These dangers underscore the need to
control
severe risk factors regardless of absolute short-term risk
estimates.
|
Diabetes Mellitus as a Special Case in Risk Assessment
|
|---|
That diabetes mellitus is a major risk factor for CVD is well
established.
2 Both type 1 diabetes
33 and type
2 diabetes
34 confer a heightened
risk for CVD. Type 2
diabetes is of particular concern because
it is so common and usually
occurs in persons of advancing age,
when multiple other risk factors
coexist. There is a growing
consensus that most patients with diabetes
mellitus, especially
those with type 2 diabetes, belong in a category
of high short-term
risk. When the risk factors of diabetic patients are
summed,
their risk often approaches that of patients with established
CHD.
35 The absolute risk of patients with type 2 diabetes
usually
exceeds the Framingham score for hyperglycemia because other
risk
factors almost always coexist. Another reason to elevate the
patient
with diabetes to a higher risk category than suggested by
Framingham
scoring is the poor prognosis of these patients once they
develop
CHD.
36 These factors point to the need to
intensify the management
of coexisting risk factors in patients with
diabetes.
7 37 These
considerations about the very high
risk of patients with diabetes
apply to ethnic groups that have a
relatively high population
risk for CHD. The inclusion of patients with
type 2 diabetes
in the very-high-risk category may not be appropriate
when they
belong to ethnic groups with a low population risk.
|
Absolute Risk Assessment in Elderly Patients
|
|---|
One of the more prominent features of the Framingham risk scoring
is
the progressive increase in absolute risk with advancing age
(Figures
1
and 2
). This increase undoubtedly reflects the
cumulative
nature of atherogenesis. With advancing age, people
typically
accumulate increasing amounts of coronary
atherosclerosis. This
increased plaque burden itself
becomes a risk factor for future
coronary
events.
38 39 40 Framingham scoring for age reflects
this
impact of plaque burden on risk. Still, average scores
mask the extent
of variability in plaque burden in the general
population. To apply
average risk scores for age to individual
patients may lead to
miscalculation of true risk, particularly
because Framingham applies so
much weight to age as a risk factor.
Miscalculation of risk could lead
to inappropriate selection
of patients for aggressive risk-reduction
therapies. This fact
points to the need for flexibility in adapting
treatment guidelines
to older persons. The tempering of treatment
recommendations
with clinical judgment becomes increasingly important
with advancing
age, particularly after the age of 65. In the future,
measures
of subclinical atherosclerosis may improve the
accuracy of global
risk assessment in older patients. When risk scoring
is used
to adjust the intensity of risk factor management in elderly
patients,
relative risk estimates may be more useful than absolute risk
estimates.
Relative risk estimates essentially eliminate the age factor
and
are based entirely on the major risk factors. These estimates
allow
the physician to stratify and compare patients of the
same age,
and patients at highest relative risk could be selected
for the most
aggressive risk management.
|
Certain Limitations of Framingham Database
|
|---|
Certain features of the Framingham scores reflect limitations
of
the data set. For example, LDL-C and HDL-C levels are known
to be
continuous in their correlation with CHD risk. Presumably
because of an
insufficient number of subjects in all categories,
these continuous
relationships are not consistently observed
between each
incremental category.
2 Moreover, the assigned
scores for
each category are not entirely consistent with the
notations
for graded risk proposed by the NCEP
6 and the
JNC.
7 Framingham scores probably require adjustment to
account for
the continuous relationship between risk factors and
CHD.
6 7 As stated previously, this adjustment was made in
Table 4
.
Finally, there is no indication that Framingham scoring
has
been corrected for regression dilution bias
41 ; this
bias results
from the random fluctuation of risk factors over time such
that
single measures of risk factors systematically underestimate
the
association between risk factors and CHD.
Prediction scores from Framingham illustrate the substantial difference
in CHD risk between men and women before age 70. The difference between
men and women particularly stands out for hard CHD end points. The
diagnosis of angina contributes a sizable fraction of all CHD end
points in middle-aged women and accounts for the notable difference
between total CHD and hard CHD in this age group. Nonatherosclerotic
anginal syndromes may have been mislabeled among total CHD end points
in some Framingham women. The relatively small rise in risk for total
CHD events after age 55 should not obscure the progressive increase in
risk for hard CHD in older women. Framingham findings on hard end
points are more consistent with population studies that show a
sharp rise in CHD morbidity and mortality in women after age 70. Even
so, a discrepancy in CHD risk between men and women persists throughout
all age groups.
|
Use of Conditional and Predisposing Risk Factors in Risk
Assessment
|
|---|
In addition to the major risk factors (Table 1
), a series
of
other risk correlates have been identified (Table 2
). Their
presence
may denote greater risk than revealed from summation of the
major
risk factors. Their quantitative contribution and independence
of
contribution to risk, however, are not well defined. Usually,
therefore,
they are not included in global risk assessment. This does
not
mean that they do not make an independent contribution to risk
when
they are present. A sizable body of research supports an
independent
contribution of each. Their relation to CHD is more complex
than
is that of the major risk factors. In some cases, they are
statistically
correlated with the major risk factors; hence, their own
independent
contribution to CHD may be obscured by the major risk
factors.
In other cases, their frequency in the population may be too
low
for them to add significant independent risk for the entire
population;
in spite of this, they could be important causes of CHD in
individual
patients. Several of the other risk factors
represent direct
targets of therapy, either because they are
causes of the major
risk factors or because circumstantial evidence of
a role in
atherogenesis is relatively strong. Thus, even though these
other
risk factors are not recommended for inclusion in absolute risk
assessment,
their exclusion from this function should not be taken to
imply
that they are clinically unimportant. Their role in evaluation
and
management of patients at risk deserves some consideration.
Obesity
The AHA defines obesity as a major risk factor for
CVD.42 Risk is accentuated when obesity has a predominant
abdominal component.5 Obesity typically raises blood
pressure and cholesterol levels42 43 44 and
lowers HDL-C levels.43 44 It predisposes to type 2
diabetes.5 It also adversely affects other risk factors:
triglycerides43 44 ; small, dense LDL
particles45 ; insulin resistance46 47 ; and
prothrombotic factors.48 49 Although not shown by the
Framingham data,2 other long-term longitudinal studies
suggest that obesity predicts CHD independently of known risk factors.
The association between excess body weight and CHD seems particularly
strong in white Americans. For example, in one long-term prospective
study,50 men aged 40 to 65 years with body mass index
(BMI) 25 to 29 kg/m2 were 72% more likely to
develop fatal or nonfatal CHD than were men who were not overweight. In
another study,51 women whose BMI was 23 to 25
kg/m2 carried a 50% increase in risk for CHD
compared with women with lower BMIs. The overall relation between body
weight and CHD morbidity and mortality is less well defined for
Hispanics,52 Pima Indians,53 and black
American women54 ; even so, obesity is a risk factor for
type 2 diabetes, which itself is a risk factor for CHD. Much remains to
be learned about the biological mechanisms underlying the association
between obesity and CHD, but without question, a strong association
exists. Consequently, obesity is a strong risk factor for
CHD3 and is a direct target for
intervention.5 Prevention of obesity and weight reduction
in overweight persons are integral parts of the strategy for long-term
risk reduction. The recent report of the NHLBI Obesity Education
Initiative5 provides a comprehensive guideline for the
management of overweight and obese patients in clinical practice.
Physical Inactivity
The AHA also classifies physical inactivity as a major risk
factor.4 Many investigations,55 including the
Framingham Heart Study,56 57 58 59 demonstrate that physical
inactivity confers an increased risk for CHD. The extent to which
physical inactivity raises coronary risk independently of the
major risk factors is uncertain.60 Certainly, physical
inactivity has an adverse effect on several known risk
factors.60 Even though physical inactivity is an
independent risk factor, physical activity levels are difficult
to reliably measure in individual patients. For these reasons, physical
inactivity is not included in quantitative risk assessment. In spite of
these limitations in assessment, previous studies61 62
document that regular physical activity reduces risk for CHD. Physical
inactivity constitutes an independent target for intervention.
Physicians should encourage all of their patients to engage in an
appropriate exercise regimen, and high-risk patients should be referred
for professional guidance in exercise training. The AHA recently
published practical recommendations for exercise regimens designed to
reduce risk for CVD.63
Family History of Premature CHD
There is little doubt that a positive family history of premature
CHD imparts incremental risk at any level of risk factors. This
association has been shown by the Framingham Heart
Study.64 Nonetheless, the degree of independence from
other risk factors and the absolute magnitude of incremental risk
remain uncertain. For this reason, Framingham investigators did not
include family history among the major independent risk factors. The
NCEP6 counts a positive family history of CHD as an
independent risk factor that modifies the intensity of LDL-lowering
therapy. Regardless of whether family history is used to modify risk
management in individual patients, the taking of a family history is
undoubtedly important. A positive family history for premature CHD
calls forth the need to test a patient's relatives for both premature
CVD and the presence of risk factors.
Psychosocial Factors
There has long been an interest in the contribution of personality
and socioeconomic factors to CHD risk. Recently, specific factors
including hostility, depression, and social isolation have been shown
to have predictive value.65 66 67 These factors, however,
are not included in the Framingham data and cannot be incorporated into
the model currently. Nonetheless, they might be taken into account in
individual patients when an overall strategy for risk reduction is
being developed.
Ethnic Characteristics
The Framingham population represents the world's most
intensively studied population for cardiovascular risk
factors. This study is of great value in developing population-based
risk estimates in this population. Because Framingham residents are
largely whites of European origin, it is uncertain whether baseline
absolute risk is similar to that in other populations. Available
evidence suggests that absolute risk varies among different populations
independently of the major risk factors. For example, absolute risk
among South Asians (Indians and Pakistanis) living in Western society
appears to be about twice that of whites, even when the 2 populations
are matched for major risk factors.68 69 70 This higher
baseline risk should be considered when South Asians living in the
United States are evaluated. Available comparisons of non-Hispanic
white, non-Hispanic black, and Hispanic Americans71 72
point to a comparable absolute risk status, but large systematic
comparisons are in the early stages. It is also possible that some
populations have a lower baseline risk than the whites studied in
Framingham. For example, results of the Honolulu Heart
Study27 suggest that Hawaiians of East Asian ancestry have
only about two thirds the absolute risk of Framingham subjects. In the
Seven Countries Study,73 the population of Japan exhibited
a much lower risk for CHD for a given set of risk factors than other
populations. Differences in absolute risk among different demographic
groups suggest the need for adjustments in estimates of absolute risk
from Framingham scores depending on racial and ethnic origins. Although
absolute risk scores may not be transportable to all populations,
relative risk estimates probably are reliable across groups. To date,
comparison studies are insufficient to provide quantitative estimates
of the adjustments needed for Framingham scores when they are applied
to individuals from different demographic backgrounds. In spite of the
limitations of the Framingham data, absolute risk estimates as applied
to some populations seem applicable to the large populations of
non-Hispanic white, Hispanic, and black Americans in the United States.
For other groups, relative risk estimates still seem applicable.
Hypertriglyceridemia
Framingham scoring does not ascribe independence to
triglyceride levels in risk assessment. Framingham
investigators74 nonetheless have reported that elevated
serum triglycerides are an independent risk factor, as have
other reports.75 76 77
Hypertriglyceridemia is correlated with
other risk factors78 ; however, its degree of independent
predictive power is difficult to assess. Several clinical
trials79 80 81 found that drugs that primarily affect
triglyceride-rich lipoproteins reduce CHD risk when used
with patients with hypertriglyceridemia.
Elevated triglycerides consequently may become a target of
therapy independent of LDL lowering. The reduction of serum
triglyceride levels will also decrease the concentrations
of small LDL particles, another putative risk factor.82 83
Of course, weight reduction in overweight patients and adoption of
regular exercise by sedentary persons will lower
triglyceride levels, which is one way in which these
changes in lifestyle reduce CHD risk.
Insulin resistance is another risk correlate for CHD.84 85
The mechanisms of association between insulin resistance are complex
and likely multifactorial. Regardless, a large portion of all patients
who are candidates for global risk assessment have insulin resistance
and its accompanying metabolic risk factors (the
metabolic syndrome). The components of this syndrome
include the atherogenic lipoprotein phenotype (elevated
triglycerides, small LDL particles, and low HDL-C
levels),78 86 elevated blood pressure, a prothrombotic
state, and often, impaired fasting glucose.87 The
metabolic syndrome is a clinical diagnosis, but the risk
accompanying it can be assessed in large part by Framingham scoring.
This scoring does not count impaired fasting glucose as an independent
risk factor, although Framingham publications88 89 90 would
support doing so. Insulin resistance can be assumed to be present
in a patient with obesity (BMI >30
kg/m2)46 47 or overweight (BMI 25 to
29.9 kg/m2) plus abdominal
obesity,46 47 especially when accompanied by elevated
plasma triglycerides,78 91 low
HDL-C,92 or impaired fasting glucose.93
Insulin resistance is acquired largely through obesity and physical
inactivity, although a genetic component undoubtedly exists. The only
therapies presently available for insulin resistance for patients
without diabetes are weight reduction94 and increased
physical activity.95
Homocysteine
A high serum concentration of homocysteine is associated with
increased risk for CHD.96 97 98 The AHA recently published
an advisory on homocysteine that provides an in-depth review of the
relation between homocysteine and CVD.99 Several
mechanisms whereby elevated homocysteine predisposes to CVD have been
postulated. However, it remains to be proved in controlled clinical
trials that a reduction in serum homocysteine levels will reduce risk
for CHD. In some patients, nonetheless, high levels of homocysteine can
be lowered by recommended daily intake of folic
acid.99 100 101 If homocysteine levels are elevated, patients
should be encouraged to consume the recommended daily intake of folic
acid, as well as vitamins B6 and
B12. Routine measurement of homocysteine
levels was not recommended for purposes of risk assessment, but
measurement is optimal in high-risk patients.99
Other Risk Correlates
Other potential risk factors include elevated concentrations of
lipoprotein(a), fibrinogen, and C-reactive protein. Routine measures of
these risk factors currently are not recommended. An elevated serum
lipoprotein(a) correlates with a higher incidence of CHD in some
studies102 103 but not in others.104 105
Furthermore, specific therapeutics to reduce lipoprotein(a) levels are
not available; some investigators have suggested that an elevated
lipoprotein(a) level justifies a more aggressive lowering of LDL-C. An
elevated fibrinogen level also is correlated with a higher CHD
incidence.106 107 Again, no specific therapies are
available, except that in smokers, smoking cessation may reduce
fibrinogen concentrations.108 Finally, C-reactive protein
is promising as a risk predictor.109 110 The preferred
method for measurement appears to be a high-sensitivity
test.111 C-reactive protein appears to be related to
systemic inflammation; however, its causative role in atherogenesis is
uncertain.
|
Implications for Clinical Risk Reduction
|
|---|
Identification of risk factors lies at the heart of clinical
efforts
to reduce risk for CVD and/or CHD. Every major risk factor
predisposes
to CHD and other cardiovascular events,
particularly if left
unattended for long periods. In addition, when
multiple risk
factors occur in a single individual, risk is compounded,
which
justifies efforts to estimate global risk. The summation of
contributions
of individual risk factors can be a valuable first step
in planning
a risk-reduction strategy for individual patients. This
first
step should be divided into 2 phases. First, absolute risk should
be
estimated from the major risk factors (listed in Table 1
).
Framingham
risk scoring provides an acceptable tool for most
non-Hispanic
white, Hispanic, and black Americans. People of South
Asian
origin appear to have about twice the absolute risk for any
set
of risk factors as whites. In contrast, East Asian Americans
may have a
lower absolute risk than other ethnic groups in the
United States.
Second, when absolute risk has been estimated
from the major risk
factors, consideration can be given to modifying
the estimate in the
presence of other risk factors (Table 2
).
Clinical judgment is
required to estimate incremental risk incurred
by these latter factors.
Risk estimates are useful both for
short-term, high-risk primary
prevention and for long-term (or
lifetime) primary prevention.
Implications for global risk assessment
can be considered for each.
Short-Term Prevention
Recent clinical trials demonstrate that significant risk reduction
can be achieved by aggressive reduction of risk factors in high-risk
patients. Clinical trials have shown that excess risk can be reduced by
33% to 50% in 5 years. This is particularly the case when
risk-reduction strategies use smoking cessation, blood
pressure–lowering agents, cholesterol-lowering drugs, and
aspirin. Clinical trials strongly suggest that glucose control reduces
the incidence of various cardiovascular end points in
patients with either type 1 diabetes112 or type 2
diabetes.113 Other clinical trials114 115
strongly suggest that aggressive LDL-lowering therapy reduces risk for
CHD in patients with type 2 diabetes. For this reason, detection of
patients at high risk, with the aid of global risk assessment, should
be an important aim of routine medical evaluation of all patients.
Specific therapies for risk reduction in high-risk patients are
described in the NCEP ATP II report for cholesterol
management,6 the JNC VI report for treatment of
hypertension,7 and by the ADA's guidelines for treatment
of diabetes mellitus.8 Once appropriate therapies are
selected, global risk scores can also be used to help instruct patients
and to improve compliance with preventive interventions.
Long-Term Prevention
Global risk assessment is particularly useful in young and
middle-aged adults for assessing relative risk and absolute long-term
risk (Figures 1
Top
Introduction
Clinical Importance of Global...