3. Screening for Cardiovascular Risk Factors
Reducing lifetime risk for cardiovascular (CV) disease (CVD) is the principle that
underlies all CVD prevention strategies, including those beginning in childhood.
Especially important is the prevention of CVD events occurring relatively early
in life (e.g., before ages 5060 years). In these Guidelines, the Expert Panel
highlights two complementary prevention strategies: (1) primordial prevention,
which seeks to prevent the development of risk factors in all children and (2) primary
prevention, a high-risk strategy aimed at reducing risk in children with dyslipidemia,
hypertension, obesity, diabetes mellitus, or other identified factors associated
with accelerated development of atherosclerotic CVD. In contrast with primordial
prevention, primary prevention requires knowledge of risk factor levels through
the screening of individuals. This section focuses on the principles of screening
within the context of the need for practical clinical recommendations even in the
presence of insufficient evidence.
Screening is common practice in regular pediatric care with age-based recommendations
designed to identify conditions at appropriate times relative to both the disease
process and the stage of growth and development. For example, the American Academy
of Pediatrics recommends universal newborn screening for metabolic conditions, hemoglobinopathy,
and hearing loss, and selective screening for elevated levels of lead in infancy
and early childhood.[1]
Although many screening programs have been widely adopted, they are not always evidence
based. This section reviews the criteria for an effective screening program and
provides discussions of both scientific and practical considerations involved in
screening for CV risk factors in childhood.
Empirically, a recommendation for universal screening requires a high burden of
proof. First, by definition, screening is performed on asymptomatic individuals.
Second, all of the downstream consequences of screening, both beneficial and harmful,
are important to consider; sometimes they are not obvious. Third, widespread screening
programs are costly.
The highest quality evidence for establishing the utility of a screening program
derives from randomized controlled trials (RCTs) of screening versus no screening.
Such trials compare clinical outcomes among children randomly allocated to no screening
with outcomes among children allocated to screening, followed by interventions among
those with identified risk. For a CV risk factor screening trial, the children in
both groups would be followed for decades to determine disease incidence, and the
analysis would allow balancing of benefits, risks, and costs. For CV risk factors
such as dyslipidemia, hypertension, and obesity, it is unlikely that such a large,
long-term study will ever exist because of the time and costs involved, as well
as the great degree of difficulty in achieving high levels of adherence and followup
over decades. Furthermore, by the time substantial numbers of definitive end points
occurred decades later, knowledge and technology most likely would have made the
initial screening test obsolete. An RCT of CV risk factor screening with shorter
term followup to examine change(s) in risk factor levels or surrogate outcomes (such
as noninvasive measures of subclinical atherosclerosis) may be more feasible. However,
like most surrogate measures and as described in the preceding Section II, subclinical
measures of atherosclerosis do not perfectly predict clinical CVD outcomes.
One argument for screening is the knowledge that extreme elevations of risk factors
are associated with early and severe clinical outcomes. For example, children with
coarctation of the aorta have elevation of upper body blood pressure (BP) from infancy.
When surgical repair of coarctation is delayed, early death from heart attack, stroke,
and aortic rupture has been well-documented.[2] Similarly, in children with
extreme elevations of low-density lipoprotein cholesterol (LDLC) levels due
to the rare inherited homozygous form of familial hypercholesterolemia, clinical
CVD events begin as early as the first decade of life.[3]
The question that the Expert Panel faced is whether childhood screening to identify
less severe forms of these risk factors is a useful strategy to prevent CVD events
from occurring in middle-aged adults. Without definitive evidence from RCTs of screening
programs, the Expert Panel was left to determine the wisdom of recommending screening
in the face of suboptimal evidence but with knowledge that the atherosclerotic process
begins in childhood and that the long time period required between screening during
youth and clinical end points makes the most rigorous test of a CV risk factor screening
program infeasible. In this situation, assessing the usefulness of screening involves
evaluating alternative criteria, including attributes of the test, outcomes of interventions
among children with actionable levels of test results, and the program as a whole.
TEST CHARACTERISTICS
Reproducibility
Lipids, BP, height, and weight are measurements with intrinsic biologic and measurement
variability. For BP and total cholesterol (TC), LDLC, and high-density lipoprotein
cholesterol (HDLC) levels, two or three measurements, taken several days to
weeks apart, appear necessary to place most children in the categories of normal,
borderline, or high with reasonable confidence.[4],[5] As described
in Section X. Overweight and Obesity, height, weight, and body mass index (BMI)
measurements are reliably reproducible, but measurements over time are needed to
provide consistent information on growth trends and to determine whether there has
been an inappropriate change in the BMI percentile relative to age- and gender-specific
norms.
Validity/Accuracy
To be useful, a screening test must detect the condition of interest with sufficient
reliability, sensitivity, and specificity to determine whether intervention is warranted
or to mandate a second test to confirm the presence of the risk factor or disease.
For most screening tests, the frequency of the screened condition is low, so even
high sensitivity and specificity translate into a low positive predictive valuea
consideration in screening for all rare conditions.[6] From a long-term perspective, this means that a
majority of children with an identified CV risk factor in childhood will not develop
premature CVD events (i.e., myocardial infarction, sudden cardiac death, or stroke
by ages 5060 years). Although counted statistically as false positives, pathology
studies demonstrate that these individuals develop atherosclerosis faster than children
with normal risk factor levels and are at increased risk for morbidity from a range
of vascular complications (see Section II. State of the Science: Cardiovascular
Risk Factors and the Development of Atherosclerosis in Childhood). The destructive
effects of early heart attacks and strokes, the impact of multiple risk factors
in increasing the risk for such events, and especially the potential to reduce the
risk of sudden death as the first manifestation of early atherosclerotic CVD mean
that decisionmakers might consider a lower positive predictive value for CV risk
factor screening than for childhood screening for other disease processes. Specifically,
because sudden cardiac death often occurs in asymptomatic individuals, the threshold
for CV screening could be lower than that for diseases that always manifest with
symptoms.
Risk factors that are considered to be most strongly associated with disease can
nonetheless be suboptimal as screening tests. For example, adults with TC values
in the highest fifth of the population distribution have approximately a threefold
higher 10-year risk of fatal ischemic heart disease than adults in the bottom fifth
of the distribution. Although this appears to be a strong association, assessing
a screening test requires estimating the absolute risk for individuals rather than
the relative risk. If one assumes that 5 percent of unaffected individuals will
have TC levels in the top fifth of the distribution, a relative risk of 3 means
that among these individuals, the test will detect only 15 percent of those destined
to die from ischemic heart disease.[7] In fact, even a relative risk
of 200 would increase this detection rate (i.e., the sensitivity) to just over 50
percent. In the case of a 10-year death rate of 1 percent, for every 100 persons
identified with high cholesterol levels, only 3 would die of ischemic heart disease
and the other 97 would be false positives. This example is simplistic since it does
not take into account other coexisting risk factors for heart disease, other manifestations
of vascular disease, or the fact that lifetime CVD risk is higher than a 10-year
risk; however, it does convey the point that relying on relative risk can lead to
overestimation of screening test performance. These considerations also underscore
that, for maximal benefit, a strong population approach should accompany any high-risk
approach to CVD prevention beginning in childhood.
Role of Selective Screening
Updating recommendations for lipid screening was one of the most important tasks
for the Expert Panel. The original 1992 National Cholesterol Education Program (NCEP)
report National Cholesterol Education Program: Report of the Expert Panel
on Blood Cholesterol Levels in Children and Adolescents (NCEP Pediatric Guidelines)
recommended screening for elevated cholesterol levels only among children with either
a family history of early CVD or elevated TC levels.[8] The rationale
was that, compared with universal screening, this selective approach would identify
the majority of children with elevated LDLC levels by screening fewer individuals
with similar benefits at lower costs, because a group with higher prevalence was
being screened. As outlined in Section IX. Lipids and Lipoproteins, the evidence
review indicates that office-based, family history-directed CV risk factor screening
identifies significantly fewer children with abnormal LDLC levels than would
universal screening. When this evidence is combined with the knowledge that a complete
family history is unavailable for many children, a family history-based screening
approach for cholesterol does not appear to be effective. One caveat to this conclusion
is that few data address whether family history-directed screening provides a useful
approach for detecting extreme LDLC levels (i.e., those high enough to warrant
medication use). Also, as described in detail in Section IX. Lipids and Lipoproteins,
most studies of medication use involve positive family history as an entry criterion,
so that more research is needed about the efficacy of treatment among family history-negative
children with elevated LDLC.
The obesity epidemic makes more intensive lipid screening among overweight and obese
children worthy of consideration, beyond the importance of identifying obesity as
an independent risk factor for future CVD. As described in these Guidelines, particularly
in Section X. Overweight and Obesity, evaluation of obese children will identify
a large number with dyslipidemia, typically moderate to severe elevation of triglycerides
(TG), mild elevation of LDL‑C and reduced HDLC, as well as elevated
BP. A very small number also will have type 2 diabetes mellitus (T2DM). Primary
treatment for any of these risk factors is weight control. TG levels in particular
are very responsive to weight loss and to dietary change. HDLC levels rise
in response to regular exercise. As presented in Section IX. Lipids and Lipoproteins
and in Section X. Overweight and Obesity, dietary change, exercise, and weight loss
can contribute to normalization of lipid levels and BP and elimination of the metabolic
abnormalities in T2DM. An unanswered question for both children and adults, however,
is the extent to which treatment of the dominant lipid abnormalities associated
with obesity will result in reduced risk of early CV events.
Acceptability
Obtaining risk factor measurements from children requires that testing be feasible
and acceptable to parents and children. Measurement of length/height and weight
are routine and well-accepted in pediatric care from birth onward. Measurement of
BP, recommended for all children beginning at 3 years of age,[9] is also routine in most
practices. Lipid testing requires a blood test and in the past has required overnight
fasting. As described in detail in Section IX. Lipids and Lipoproteins, measuring
non-HDLC, which is accurate in the nonfasting state,[10] as the first step in
lipid screening should be a major improvement in feasibility for clinical practice
and acceptability to parents and children. Additional blood draws in the fasting
state are needed to confirm initially abnormal levels; their acceptability to children
and parents is an open question. Some older research suggests low compliance by
parents and children with followup screening recommendations in real-world practice,
but current data are sparse.[11]
Evaluation of Interventions among Children with Abnormal Risk Factors
As reviewed throughout these Guidelines, the major rationale for CV risk factor
screening in youth, followed by treatment of abnormal levels, derives from knowledge
that atherosclerosis begins in childhood and that its severity is greater in children
with a higher burden of atherogenic risk factors. However, these observations do
not directly address interventions to reduce this burden. A sine qua non of useful
screening programs is that interventions based on abnormal screening results must
be not only efficacious but also more efficacious than interventions that occur
later in the disease process. If earlier interventions are not more efficacious
than those initiated later in life, they incur cost and potential risk with no benefit.
Studies in adults indicate that treating prehypertension with medication or lifestyle
change is associated with a lower subsequent incidence of hypertension.[12],[13],[14]
In children with coarctation of the aorta, long-term followup studies demonstrate
that early surgical repair is associated with reduced incidence of subsequent CVD.[15]
The extent to which efficacy of intervention in youth extends to primary and less
severe hypertension is not known. For obesity and lipids, no studies directly address
the benefits (or risks) of early childhood treatment over treatment later in life
on clinical disease.
The case of heterozygous familial hypercholesterolemia (FH) provides partial proof
of the concept that early reduction of risk factors may reduce future disease rates.
Individuals with FH are at increased risk for early CVD because of elevated LDLC
levels from infancy. In natural history studies, 50 percent of males and 25 percent
of females with FH develop clinical CVD by age 50 years.[16],[17]
In RCTs among older children and adolescents with FH, statin treatment substantially
lowers LDLC levels and slows progression of atherosclerosis as assessed by
noninvasive testing.[18]
Although pediatric medication trials are of relatively short duration, they suggest
that sustained LDLC-lowering therapy in children with FH will lower the risk
of early clinical CVD. By extension and by analogy with adult treatment guidelines,
the Expert Panel recommends treating less extreme elevations of LDLC in childhood,
especially in the setting of multiple CV risk factors. The extent to which either
lifestyle change or medication treatment of lipids in youth reduces the atherosclerotic
burden or risk of CV events is not yet known, nor is the long-term safety of treatment
with statins beginning in childhood, although published trials do not show any adverse
impact on growth, pubertal maturation, or hormonal metabolism over several years
(see Section IX. Lipids and Lipoproteins).
It is vital to know how well primary care clinicians can incorporate recommended
screening programs, including testing and interventions, into routine practice.
Unfortunately, evidence is meager regarding repeated testing strategies or the effectiveness
or sustainability of interventions in real-world practices. Thus, the evidence review
almost exclusively identified studies that addressed intervention efficacy in a
research setting. Nevertheless, the Expert Panel anticipates that practices will
be better equipped to adopt its recommendations in the future than they are today.
The recommendations may very well effect changes in policy and systems, as well
as additional research that will assist clinicians in overcoming current time, space,
personnel, and reimbursement challenges.
Other Issues
A potential ancillary benefit of childhood CV risk factor screening is to alert
older family members of the need to have their own risk factors checked, especially
lipids. If they have not been screened previously, parents and grandparents of children
with abnormal lipid values should have their own lipid levels checked by their primary
care providers. Several studies have shown that first-degree relatives of children
with elevated LDLC levels have both higher LDLC levels themselves and
higher rates of CV events.[19],[20],[21]
It is also possible that knowledge of an abnormal risk factor in a child could spur
lifestyle changes for the whole family. Many clinicians can cite anecdotes of families
who made salutary changes after learning of a child's elevated cholesterol, BP,
or BMI. However, the extent to which this phenomenon occurs is unclear, nor is it
known whether this effect adds substantially to a concerted primordial prevention
approach aimed at all children and families. Certainly, the literature is clear
that, in general, knowledge alone is insufficient for effective behavior change.
FURTHER RESEARCH NEEDED
One of the most difficult issues to address in CV risk factor screening in childhood
is the potential value of slowing or reversing atherosclerosis in its early stages.
This issue is particularly important with respect to lipids. Statin treatment among
high-risk adults reduces CV events within months of initiation.[22]
However, statin therapy does not eliminate risk, and adult trials cannot include
those who have already died of very early CV events, of which sudden death is a
particular concern. Thus, an RCT of childhood CV risk factor screening and treatment
of elevated levels with a noninvasive measure of atherosclerosis as the primary
outcome is an appealing study design. This study design could apply to any of the
childhood CV risk factors, not just elevated LDLC. As described in the previous
section, one caveat of using measures of subclinical atherosclerosis as end points
is that, whether invasive or noninvasive, all of those measures are surrogates for
the true outcome of clinical CV events. Trials with surrogate end points have sometimes
led to misleading recommendations and harmful clinical practices.[23]
Through the assessment of benefit, risk, and cost, the science of clinical decisionmaking
offers an alternative to large long-term trials for the evaluation of CV risk factor
screening in children. A decision analytic framework allows modeling of the immediate
and downstream consequences of assessing childhood risk factors and intervening
among those with abnormal results. As with any analytic method, decision analysis
has strengths and weaknesses. The most important attribute of decision analysis
is to ask the right questions. In the case of childhood CV risk factor screening,
one strength of decision analysis is the ability to compare a number of strategies,
including no screening or intervention, primordial prevention approaches only, universal
or selective screening at specific ages, and the marginal benefit of screening at
younger versus older ages. If a model contains appropriate decision points and outcomes,
it can be a valuable method for assessing effects on long-term health outcomes without
having to wait decades. In such a model, empirical evidence drives the quantitative
comparison of one decision versus another. A good decision analysis not only will
point out where data are lacking but also, through sensitivity analysis, will identify
the most important new data to collect. The cost to society at large will likely
be a major factor in decisions regarding screening. When added to a decision analytic
model, the costs of screening, followup, and intervention can lead to estimation
of the cost-effectiveness of various screening strategies, and sensitivity analysis
can show where variability in costs is meaningful or irrelevant. Cost-effectiveness
can be a major driver of policy decisions to support prevention programs. For these
reasons, well-considered cost-effectiveness analyses of childhood CV risk factor
screening should be a priority for future research.
CONCLUSIONS
The Expert Panel recommends two complementary strategies to reduce future risk for
clinical CVD. Primordial prevention seeks to prevent the acquisition of risk factors
by optimizing CV health for all, beginning in infancy. Primary prevention requires
screening to identify children at increased risk for CVD. This section focused on
screening and reviewed the limitations of current knowledge within the requirements
for a useful screening program. Taking these factors into account, the Expert Panel
recommends routine measurement of length/height and weight beginning in infancy,
with calculation of BMI annually beginning at age 2 years to identify growth trends;
yearly assessment of BP from age 3 years; and universal screening for lipid abnormalities
by a nonfasting non-HDLC level at age 10 years. These screening strategies,
described in detail in the respective risk factor sections of these Guidelines,
will identify a relatively large number of children for whom the Expert Panel recommends
intensified lifestyle intervention. Only a small number of children will require
pharmacologic therapy. While they await the results of future research, the Expert
Panel members conclude that recommending these assessments, followed by interventions
as part of routine pediatric care, represents the best current primary prevention
strategy to lower lifetime risk of atherosclerotic vascular disease.
REFERENCES
[1]
Hagan JF, Shaw JS, Duncan PM eds. Bright Future: Guidelines for Health Supervision
of Infants, Children and Adolescents, Third Edition. Elk Grove Village, IL: American
Academy of Pediatrics.
[2]
Maron BJ, Humphries JO, Rowe RD, Mellitis EG. Prognosis of surgically corrected
coarctation of the aorta: A 20 year postoperative appraisal. Circulation 1973;47:119-126.
[3]
Kwiterovich PO Jr. Recognition and management of dyslipidemia in children and adolescents.
J Clin Endocrinol Metab 2008;93:4200-4209.
[4]
Gillman MW, Cupples LA, Moore LL, Ellison RC. Impact of within-person variability
on identifying children with hypercholesterolemia: Framingham Children's Study.
J Pediatr 1992;121:342-347.
[5]
Rosner BA, Cook NR, Evans DA, Keough ME, Taylor JO, Polk BF, Hennekens CH.
Reproducibility and predictive values of routine blood pressure measurements in
children. Am J Epidemiol 1987;126:1115-1125.
[6]
Tarini BA, Christakis DA, Welch HG. State newborn screening in the tandem mass spectrometry
era: more tests, more false-positive results. Pediatrics 2006;118:448-456.
[7]
Wald NJ, Hackshaw AK, Frost CD. When can a risk factor be used as a worthwhile screening
test? BMJ 1999;319:1562-1565.
[8]
NCEP Expert Panel of Blood Cholesterol Levels in Children and Adolescents. National
Cholesterol Education Program (NCEP): Highlights of the Report of the Expert Panel
on Blood Cholesterol Levels in Children and Adolescents. Pediatrics 1992;89:495-501.
(PM:1741227)
[9]
National High Blood Pressure Education Program Working Group on High Blood Pressure
in Children and Adolescents. The fourth report on the diagnosis, evaluation, and
treatment of high blood pressure in children and adolescents. Pediatrics 2004;114(2
Suppl 4th Report):555-576. (PM:15286277)
[10]
McMahan CA, Gidding SS, Malcom GT, Tracy RE, Strong JP, McGill HC, Jr. Pathobiological
determinants of atherosclerosis in youth risk scores are associated with early and
advanced atherosclerosis. Pediatrics 2006;118(4):1447-1455.) (PM:17015535)
[11]
Bachman RP, Schoen EJ, Stembridge A, Jurecki ER, Imagire RS. Compliance with childhood
cholesterol screening among members of a prepaid health plan. Arch Pediatr Adol
Med 1993;147:382-385.
[12]
Julius S, Nesbitt SD, Egan BM, Weber MA, Michelson EL, Kaciroti N, Black HR, Grimm
RH Jr, Messerli FH, Oparil S, Schork MA; Trial of Preventing Hypertension (TROPHY)
Investigators. Feasibility of treating prehypertension with an angiotensin-receptor
blocker. N Engl J Med 2006;354: 1685-1697.
[13]
Elmer PJ, Obarzanek E, Vollmer WM, Simons-Morton D, Stevens VJ, Young DR, Lin PH,
Champagne C, Harsha DW, Svetkey LP, Ard J, Brantley PJ, Proschan MA, Erlinger TP,
Appel LJ; PREMIER Collaborative Research Group. Effects of comprehensive lifestyle
modification on diet, weight, physical fitness, and blood pressure control: 18-month
results of a randomized trial. Ann Intern Med 2006;144(7):485-95.
[14]
Effects of weight loss and sodium reduction intervention on blood pressure and hypertension
incidence in overweight people with high-normal blood pressure. The Trials of Hypertension
Prevention, phase II. The Trials of Hypertension Prevention Collaborative Research
Group. Arch Intern Med 1997;157(6):657-67.
[15]
Toro-Salazar OH, Steinberger J, Thomas W, Rocchini AP, Carpenter B, Moller JH. Long-term
follow-up of patients after coarctation of the aorta repair. Amer J Cardiol 2002;89:541-547.
[17]
Slack J. Risks of ischaemic heart disease in familial hyperlipoproteinemia. Lancet
1969;2:1380-1382.
[18]
Rodenburg J, Vissers MN, Wiegman A, van Trotsenburg AS, van der Graaf A, de Groot
E, Wijburg FA, Kastelein JJ, Hutten BA, Statin treatment in children with familial
hypercholesterolemia: the younger, the better. Circulation 2007;116(6):664-668.
[19]
Schrott HG, Clarke WR, Wiebe DA, Connor WE, Lauer RM. Increased coronary mortality
in relatives of hypercholesterolemic school children: the Muscatine study. Circulation
1979;59(2):320-326. (PM:758999)
[20]
Moll PP, Sing CF, Weidman WH, Gordon H, Ellefson RD, Hodgson PA, Kottke BA. Total
cholesterol and lipoproteins in school children: prediction of coronary heart disease
in adult relatives. Circulation 1983;67:127-134.
[21]
Muratova VN, Islam SS, Demerath EW, Evans Minor V, Neal WA. Cholesterol screening
among children and their parents. Prev Med 2001;33:1-6.
[22]
Thavendirananthan P, Bagai A, Brookhart MA, Choudry NK. Primary prevention of cardiovascular
diseases with statin therapy: a meta-analysis of randomized controlled trials. Arch
Intern Med 2006;166:2307-2313.
[23]
CAST Investigators, DS Echt, PR Liebson, LB Mitchell, RW Peters, D Obias-Manno,
AH Barker, D Arensberg, A Baker, L Friedman, HL Greene, ML Huther and DW Richardson,
Mortality and morbidity in patients receiving encainide, flecainide or placebo.
The Cardiac Arrhythmias Suppression Trial. N Engl J Med 1991; 324: 781788.
Back to Top
Back to Table of Contents
|
|