OBJECTIVE. This study sought to examine the usefulness of non–high-density lipoprotein cholesterol levels in predicting future dyslipidemia and other cardiovascular risk in adulthood.
METHODS. The study sample consisted of a longitudinal cohort of subjects (n = 1163; 30.1% black and 55.4% female) who participated in the Bogalusa Heart Study both as children at 5 to 14 years of age and as adults 27 years later.
RESULTS. The childhood level of non–high-density lipoprotein cholesterol, like low-density lipoprotein cholesterol, was the best predictor of the adulthood level; the next best predictor for both variables was the change in BMI from childhood to adulthood. Furthermore, those in the age-, race-, and gender-specific top quartile, compared with those in the bottom quartile, of non–high-density lipoprotein cholesterol and low-density lipoprotein cholesterol levels in childhood were 4.5 and 3.5 times more likely, respectively, to develop adult dyslipidemia, independent of baseline BMI and BMI change after 27 years. Although, at equivalent cutoff points, childhood high-risk versus acceptable-risk status for both lipid measures was associated significantly with increased prevalence of obesity and adverse levels of low-density lipoprotein cholesterol and triglycerides in adulthood, only childhood non–high-density lipoprotein cholesterol high-risk status was associated with increased prevalence of low high-density lipoprotein cholesterol levels, hyperinsulinemia, and hyperglycemia (marginal).
CONCLUSIONS. Adverse levels of non–high-density lipoprotein cholesterol versus low-density lipoprotein cholesterol in childhood not only equally persist over time and better predict adult dyslipidemia but also are related to nonlipid cardiovascular risk factors in adulthood.
- non–high-density lipoprotein cholesterol
- low-density lipoprotein cholesterol
- coronary artery disease risk
Adverse levels of serum lipoprotein cholesterols have long been recognized as an important risk factor for coronary artery disease (CAD) among adults.1 Furthermore, adverse levels among youths are associated with early subclinical coronary atherosclerosis.2–4 Accordingly, the National Cholesterol Education Programs for adults5 and children,6 the American Academy of Pediatrics Committee on Nutrition,7 and the American Heart Association guidelines for primary prevention of atherosclerotic cardiovascular disease beginning in childhood8 have provided guidelines for the detection, evaluation, and treatment of adverse levels of lipoproteins.
With respect to lipid profiling for CAD risk assessment, low-density lipoprotein (LDL) cholesterol levels, estimated with the equation described by Friedewald et al,9 are targeted widely for primary prevention and intervention. However, there are limitations to the use of estimated LDL cholesterol levels.10–12 In particular, the estimated values include cholesterol derived from lipoprotein(a) and intermediate-density lipoproteins, to varying degrees, and values are artificially low with elevated triglyceride levels associated with a nonfasting state or metabolic disorders. To obviate these limitations, measurements of non–high-density lipoprotein (HDL) cholesterol, which are obtained directly by subtracting HDL cholesterol levels from total cholesterol levels and include quantitatively all potentially atherogenic apolipoprotein B-containing lipoproteins, are considered a simpler and better screening tool for the assessment of CAD risk. Consequently, non-HDL cholesterol is being used increasingly in clinical research involving adult populations,3,13–15 and Adult Treatment Panel III of the National Cholesterol Education Program recommended non-HDL cholesterol as a secondary target for therapy among patients with metabolic syndrome or diabetes mellitus.5 Furthermore, population frequency distributions and clinically useful cutoff points for this variable are currently available for US populations16
The Bogalusa Heart Study, a biracial (black/white), community-based investigation of the evolution of cardiovascular risk since childhood,17 reported the distribution and correlates of non-HDL cholesterol among children.18 The purposes of this study were to determine whether non-HDL cholesterol level versus LDL cholesterol level could be useful, in terms of long-term (27 years) tracking of adverse levels of these factors from childhood to adulthood, and to determine the effect of childhood adverse levels of these factors on adult dyslipidemia and other cardiovascular risk factors, such as obesity, hyperglycemia, hyperinsulinemia, and hypertension.
The study cohort was derived from 2 cross-sectional surveys, 1 performed in 1973 to 1974 among children (n = 3446; 37% black and 47% female) and the other performed in 2001 to 2002 among those who remained in the community as adults (n = 1203), conducted in the biracial (65% white and 35% black) community of Bogalusa, Louisiana. Subjects (n = 1163; 30% black and 55% female) for whom fasting blood samples were obtained at both examinations were included in the study. These subjects were 5 to 14 years of age at baseline and were monitored for an average period of 27 years. With respect to age, race, gender, and lipid profile, the baseline childhood characteristics of the study cohort, which represented 34% of the originally ascertained childhood population, were similar to the characteristics of the subjects who did not participate in the follow-up survey as adults. For example, the childhood mean differences in lipid levels between participants and nonparticipants varied from 0.2 mg/dL for HDL cholesterol to −2.1 mg/dL for non-HDL cholesterol.
Standardized protocols were used by trained examiners.19 Participants were instructed to fast for 12 hours before venipuncture, and compliance was determined by interview on the morning of the examination. Height and weight were measured in triplicate. BMI, calculated as weight in kilograms divided by the square of the height in meters, was used as a measure of obesity. Right upper arm length and circumference were used to select the cuff size for blood pressure measurements with mercury sphygmomanometers. Two randomly assigned nurses measured blood pressure (3 replicates each) while the subjects were in a relaxed sitting position. Subjects were asked to complete questionnaires regarding personal health history (eg, hypertension, hypercholesterolemia, or diabetes mellitus and medical treatment for these conditions).
In 1973 to 1974, cholesterol and triglyceride levels were measured with chemical procedures (Technicon Auto Analyzer II; Technicon Instrument Corp, Tarrytown, NY), according to the protocol developed by the Lipid Research Clinics Program.20 In 2001 to 2002, these levels were measured with enzymatic procedures (Hitachi 902 automatic analyzer; Roche Diagnostics, Indianapolis, IN). Lipoprotein cholesterol levels were measured with a combination of heparin/calcium precipitation and agar/agarose gel electrophoresis.21 Both chemical and enzymatic procedures met the performance requirements of the Lipid Standardization Program of the Centers for Disease Control and Prevention (Atlanta, GA). The laboratory has been monitored for precision and accuracy of total cholesterol, triglyceride, and HDL cholesterol measurements by this agency's surveillance program since 1973.
A commercial radioimmunoassay kit was used for measurement of plasma immunoreactive insulin levels (Padebas Pharmacia, Piscataway, NJ). Glucose levels were measured as part of a multiple chemistry profile, with a glucose oxidase method.
All data analyses were performed with SAS 9.1 (SAS Institute Inc, Cary, NC). To evaluate the persistence or tracking of elevated levels of non-HDL cholesterol versus LDL cholesterol from childhood to adulthood, the baseline (childhood) age-, race-, and gender-specific top quintile for each of these lipoprotein variables was used as a cutoff point to classify children as dyslipidemic and to examine the distribution of such children among the corresponding adulthood quintiles 27 years later. In a multivariate linear-regression analysis, the baseline (childhood) level of each lipoprotein variable was evaluated as an independent predictor of follow-up (adulthood) level, controlling for age, race, gender, baseline BMI, and changes in BMI from baseline to the follow-up assessment. A stepwise logistic-regression analysis was then used to determine the odds ratio of developing any dyslipidemia in adulthood, on the basis of extreme age-, race-, and gender-specific quartiles of childhood non-HDL cholesterol or LDL cholesterol levels. The odds ratios were adjusted for baseline BMI and BMI changes from childhood to adulthood. Colinearity was checked in the fixed model. Because there was no interaction effect between childhood race (or gender) and non-HDL cholesterol (or LDL cholesterol) levels, results were presented without stratification according to race or gender. Adverse levels of LDL cholesterol, non-HDL cholesterol, HDL cholesterol, or triglycerides in adulthood were determined as defined by the National Cholesterol Education Program guidelines for adults,5 and the presence of any of these conditions was termed adult dyslipidemia. Subjects were classified as having dyslipidemia if their LDL cholesterol levels were ≥160 mg/dL, non-HDL cholesterol levels were ≥190 mg/dL, HDL cholesterol levels were <40 mg/dL, or triglyceride levels were ≥150 mg/dL.
Finally, the prevalence of cardiovascular risk factors in adulthood was examined according to childhood levels of non-HDL cholesterol and LDL cholesterol. With respect to LDL cholesterol, children were classified, on the basis of pediatric guidelines,6,7 as being at acceptable (<110 mg/dL) or higher (≥130 mg/dL) risk. With respect to non-HDL cholesterol, values equivalent to the cutoff points for LDL cholesterol mentioned above were derived from a regression analysis of non-HDL cholesterol levels versus LDL cholesterol levels in the Bogalusa Heart Study pediatric population reported previously.18 Accordingly, levels of non-HDL cholesterol of <123 mg/dL or ≥144 mg/dL were considered acceptable or high, respectively. Clinical abnormalities regarding cardiovascular risk factor variables in adulthood were defined as follows: BMI of ≥30 kg/m2; LDL cholesterol level of ≥160 mg/dL; triglyceride level of ≥150 mg/dL; HDL cholesterol level of <40 mg/dL; glucose level of ≥126 mg/dL; insulin level of >18 μU/mL; systolic blood pressure of ≥140 mm Hg and/or diastolic blood pressure of ≥90 mm Hg; or receiving medication for dyslipidemia, hyperglycemia, or hypertension. Fasting insulin levels (a reasonably good indicator of insulin resistance) of ≥18 μU/mL were found to be associated with insulin resistance for all normoglycemic individuals.22
The persistence (tracking) of serum non-HDL cholesterol and LDL cholesterol levels from childhood to adulthood was examined in terms of persistence of ranking in highest quintile of the distribution over a 27-year period. If there was no tracking, then 20% of those in a given quintile at baseline would persist in that ranking at the follow-up assessment through chance alone. As shown in Fig 1, 38.5% of individuals who ranked highest (in the top quintile) with respect to non-HDL cholesterol levels in childhood also did so in adulthood; another 27.7% remained in the next highest (fourth) quintile. In other words, 66.2% of individuals who ranked highest in childhood tended to maintain their high ranks by being above the 60th percentile in adulthood. LDL cholesterol levels showed a similar trend for tracking over time.
The association between childhood and adulthood non-HDL cholesterol or LDL cholesterol levels was found to be independent in a stepwise, multivariate, regression analysis. For example, as shown in Table 1, a childhood non-HDL cholesterol level 1 mg/dL higher predicted independently an adulthood level 0.52 mg/dL higher; an increase in BMI of 1 kg/m2 from childhood to adulthood predicted an adulthood non-HDL cholesterol level 1.93 mg/dL higher. As shown by the standardized regression coefficients, the best predictor for adulthood non-HDL cholesterol or LDL cholesterol level was the corresponding childhood level. The next best predictor for both variables was the change in BMI from childhood to adulthood, followed by race (white more than black), gender (male more than female), and age, in that order. Of note, overall these variables accounted for relatively lower percentages of variability for non-HDL cholesterol (R2 = 26%) and LDL cholesterol (R2 = 23%).
In a stepwise logistic-regression analysis, children in the age-, race-, and gender-specific top quartile of non-HDL cholesterol and LDL cholesterol levels, compared with those in the bottom quartile, were 4.5 and 3.5 times more likely, respectively, to develop dyslipidemia, in terms of adverse levels of LDL cholesterol, non-HDL cholesterol, triglycerides, or HDL cholesterol, independent of baseline BMI and the change in BMI after 27 years (Table 2). If the adverse levels of triglycerides were eliminated from the definition of dyslipidemia, however, then the odds of developing dyslipidemia on the basis of childhood non-HDL cholesterol levels (odds ratio: 5.23; P < .0001) versus LDL cholesterol levels (odds ratio: 5.59; P < .0001) were somewhat similar.
On the basis of non-HDL cholesterol and LDL cholesterol levels, children were classified into acceptable-risk and higher-risk groups. The prevalence rates of obesity, dyslipidemia, hypertension, hyperinsulinemia, and hyperglycemia in adulthood, after 27 years, were compared for the 2 groups (Table 3). On the basis of the childhood non-HDL cholesterol level criterion, the prevalence of adulthood CAD risk factors in the higher-risk group, compared with the acceptable-risk group, was significantly greater with respect to obesity, high LDL cholesterol level, high triglyceride level, low HDL cholesterol level, hyperinsulinemia, and hyperglycemia (marginal significance). With respect to childhood LDL cholesterol levels, adulthood prevalence was significantly greater only for conditions of obesity, high LDL cholesterol level, and high triglyceride level.
This community-based study demonstrated that elevated levels of non-HDL cholesterol, like levels of LDL cholesterol, in childhood tracked (persisted) reasonably in ranking over a 27-year period into adulthood, predicted adult dyslipidemia, and were associated significantly with increased prevalence of obesity and adverse levels of LDL cholesterol and triglycerides in adulthood. In addition, childhood higher-risk status based on non-HDL cholesterol levels differed from that based on LDL cholesterol levels in that it was associated with increased prevalence of hyperglycemia, hyperinsulinemia, and low HDL cholesterol levels in adulthood.
The observed tracking of adverse levels of non-HDL cholesterol from childhood to adulthood and the related predictability of adult dyslipidemia are consistent with observations from previous studies, including our own.23–28 As might be expected on the basis of tracking data, the best predictor of the adulthood non-HDL cholesterol level was the baseline (childhood) level. The findings support the idea that cholesterol screening in pediatric populations is likely to identify a group of children who are at risk for developing dyslipidemia as adults. Furthermore, the current observation that a gain in adiposity (BMI) from childhood to adulthood (which is a modifiable risk factor variable) was the next best predictor of adverse levels of adulthood non-HDL cholesterol and LDL cholesterol underscores the importance of controlling obesity during the early developmental period to prevent the emergence of dyslipidemia.
The higher prevalence of low HDL cholesterol levels, hyperglycemia, and hyperinsulinemia in adulthood among those with non-HDL cholesterol higher-risk status in childhood indicates that this lipid measure could also be a marker for certain nonlipid risk factors characteristic of metabolic syndrome,29,30 as suggested previously.31 It should be noted, however, that the prevalence of adult hypertension, which is a component of metabolic syndrome, showed no significant trend between acceptable-risk and higher-risk groups. It is likely that non-HDL cholesterol levels, as a measure of lipid triad or atherogenic dyslipidemia, reflect an imbalance in the metabolism of carbohydrates and lipids.31
As a limitation, this study lacks comparative data on apolipoprotein B, which is a better indicator of the number of atherogenic particles and the attendant CAD risk than is LDL cholesterol.32–34 Although measurement of apolipoprotein B has improved greatly, the cost and lack of standardization analogous to the lipid profile are still impediments to its wider use. It was reported that the non-HDL cholesterol level is the best surrogate measure of the apolipoprotein B level because it correlated better with the apolipoprotein B level than did the LDL cholesterol level across a wide range of triglyceride levels among adults.14 Moreover, studies found the non-HDL cholesterol level to be a better predictor of CAD-related morbidity and death than the LDL cholesterol level.35–39
The present findings indicate the usefulness of childhood non-HDL cholesterol levels in predicting adult dyslipidemia and other nonlipid cardiovascular risks for a cohort of subjects who were still young to experience coronary events. Moreover, because overnight fasting is not required for measurement of non-HDL cholesterol levels (as opposed to LDL cholesterol levels), regardless of triglyceride levels, this measure could be especially advantageous for CAD risk evaluation in pediatric populations. However, because the present results were obtained with fasting samples, corresponding data are needed for nonfasting samples. Additional population-based studies are obviously needed to validate the current findings and to develop consensus regarding childhood cutoff points to define risk status.
This work was supported by grant AG16592 from the National Institute on Aging, grant HL38855 from the National Heart, Lung, and Blood Institute, and grant HD043820 from the National Institute of Child Health and Human Development.
The Bogalusa Heart Study is a joint effort of many investigators and staff members, whose contributions are acknowledged gratefully. We especially thank the Bogalusa School System, teachers, parents, and, most importantly, participants.
- Accepted January 17, 2006.
- Address correspondence to Gerald S. Berenson, MD, Tulane Center for Cardiovascular Health, 1440 Canal St, Suite 1829, New Orleans, LA 70112. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA.2001;285 :2486– 2497
- ↵National Cholesterol Education Program. Report of the Expert Panel on Blood Cholesterol Levels in Children and Adolescents. Pediatrics.1992;89 :525– 584
- ↵American Academy of Pediatrics, Committee on Nutrition. Cholesterol in children. Pediatrics.1998;101 :141– 147
- ↵Kavey RW, Daniels SR, Lauer RM, Atkins DL, Hayman LL, Taubert K. American Heart Association guidelines for primary prevention of atherosclerotic cardiovascular disease beginning in childhood. Circulation.2003;107 :1562– 1566
- ↵Friedewald W, Levy R, Fredrickson DS. Estimation of the concentration of low-density lipoprotein cholesterol in plasma, with use of the preparative ultracentrifuge. Clin Chem.1972;18 :499– 502
- ↵Frost PH, Davis BR, Burlando AJ, et al. Serum lipids and incidence of coronary heart disease: findings from the Systolic Hypertension in the Elderly Program (SHEP). Circulation.1996;94 :2381– 2388
- ↵Ballantyne CM, Andrews TC, Hsia JA, Kramer JH, Shear C, ACCESS Study Group. Correlation of non-high-density lipoprotein cholesterol with apolipoprotein B: effect of 5-hydroxymethylglutaryl-coenzyme A reductase inhibitors on non-high-density lipoprotein cholesterol levels. Am J Cardiol.2001;88 :265– 269
- ↵The Bogalusa Heart Study 20th Anniversary Symposium. Am J Med Sci.1995;310 (suppl 1):S1–S138
- ↵Berenson GS, McMahan CA, Voors AW, et al. Cardiovascular Risk Factors in Children: The Early Natural History of Atherosclerosis and Essential Hypertension New York, NY: Oxford University Press; 1980
- ↵Lipid Research Clinics Program. Manual of Laboratory Operations, Vol 1, Lipid and Lipoprotein Analysis Bethesda, MD: National Institutes of Health; 1974. Department of Health, Education, and Welfare publication NIH 75–628
- ↵Srinivasan SR, Berenson GS. Serum lipoprotein in children and methods for study. In: Lewis LA, ed. CRC Handbook of Electrophoresis: Vol III, Lipoprotein Methodology and Human Studies Boca Raton, FL: CRC Press; 1983:185– 204
- ↵Laakso M. How good a marker is insulin level for insulin resistance? Am J Epidemiol.1993;137 :959– 965
- ↵Lauer RM, Lee J, Clarke WR. Factors affecting the relationship between childhood and adult cholesterol levels: the Muscatine Study. Pediatrics.1988;82 :309– 318
- Webber LS, Srinivasan SR, Wattigney WA, Berenson GS. Tracking of serum lipids and lipoproteins from childhood to adulthood: the Bogalusa Heart Study. Am J Epidemiol.1991;133 :884– 899
- ↵DeFronzo RA, Ferrannini E. Insulin resistance: a multifaceted syndrome responsible for NIDDM, obesity, hypertension, dyslipidemia and atherosclerotic cardiovascular disease. Diabetes Care.1991;14 :173– 194
- ↵Lamarche B, Moorjani S, Lupien PJ, et al. Apolipoprotein A-I and B levels and the risk of ischemic heart disease during a five-year follow-up of men in the Quebec Cardiovascular Study. Circulation.1996;94 :273– 278
- ↵Schaefer EJ, Lamon-Fava S, Cohn SD, et al. Effects of age, gender, and menopausal statins on plasma low density lipoproteins cholesterol and apolipoprotein B levels in the Framingham Offspring Study. J Lipid Res.1994;35 :779– 792
- Bittner V, Hardison R, Kelsey SF, Weiner BH, Jacobs AK, Sopko G. Non-high-density lipoprotein cholesterol levels predict five-year outcome in the Bypass Angioplasty Revascularization Investigation (BARI). Circulation.2002;106 :2537– 2542
- Lu W, Resnick HE, Jablonski KA, et al. Non-HDL cholesterol as a predictor of cardiovascular disease in type 2 diabetes. Diabetes Care.2003;26 :16– 23
- Copyright © 2006 by the American Academy of Pediatrics