Usefulness of Childhood Non–High Density Lipoprotein Cholesterol Levels Versus Other Lipoprotein Measures in Predicting Adult Subclinical Atherosclerosis: The Bogalusa Heart Study
OBJECTIVE. This study sought to examine the usefulness of childhood non–high-density lipoprotein cholesterol level versus low-density lipoprotein cholesterol level, high-density lipoprotein cholesterol level, triglyceride level, apolipoprotein B level, apolipoprotein A-I level, total cholesterol/high-density lipoprotein cholesterol ratio, and apolipoprotein B/apolipoprotein A-I ratio in predicting adult excess carotid intima-media thickness, an indicator of subclinical atherosclerosis.
METHODS. This retrospective cohort study included 437 black and white subjects (70% white and 40% male) who participated in the Bogalusa Heart Study as children 5 to 17 years of age and as adults 16 to 19 years later.
RESULTS. In analyses of each lipoprotein measure as a risk factor for predicting excess carotid intima-media thickness in adulthood, non–high-density lipoprotein cholesterol level, low-density lipoprotein cholesterol level, total cholesterol/high-density lipoprotein cholesterol ratio, apolipoprotein B level, and apolipoprotein B/apolipoprotein A-I ratio emerged as significant predictors, with respective odds ratios of 2.60, 2.95, 1.78, 1.44, and 1.69, after adjustment for childhood BMI, systolic blood pressure, other lipoprotein measures, and follow-up years; the odds ratios for high-density lipoprotein cholesterol, triglyceride, and apolipoprotein A-I levels were not significant. Regarding the discriminating value of different childhood lipoprotein measures in predicting excess carotid intima-media thickness in adulthood, analyses of the area under the receiver operating characteristic curve for each lipoprotein measure, adjusted for the aforementioned nonlipoprotein covariates, indicated that the value of 0.65 for the non–high-density lipoprotein cholesterol level was similar in magnitude to those for other lipoprotein measures, with values ranging from 0.62 to 0.66.
CONCLUSIONS. Childhood non–high-density lipoprotein cholesterol levels are as good as other lipoprotein measures in predicting subclinical atherosclerosis in adulthood, which has practical implications for coronary artery disease risk assessment and intervention in pediatric populations.
- non–high-density lipoprotein cholesterol
- lipoprotein measures
- carotid intima-media thickness
- subclinical atherosclerosis
The fact that adverse levels of serum lipoprotein cholesterols are important risk factors for coronary artery disease (CAD) is well recognized.1 Although clinical manifestations of CAD do not usually emerge before middle age, an adverse lipoprotein profile is associated with subclinical coronary atherosclerosis even among youths.2–4 Consequently, guidelines have been provided to identify and to treat dyslipidemia in adults as well as children.5,6
Measurements of total cholesterol, low-density lipoprotein (LDL) cholesterol, and high-density lipoprotein (HDL) cholesterol are widely recommended for CAD risk assessment. However, studies in adults have shown the superior prognostic value of apolipoprotein B and apolipoprotein A-I levels7–9 and specific ratios of lipoprotein variables, such as the total cholesterol/HDL cholesterol ratio and the apolipoprotein B/apolipoprotein A-I ratio.10–12 Furthermore, it is increasingly being recognized that the non–HDL cholesterol level, calculated as the total cholesterol level minus the HDL cholesterol level and including all atherogenic apolipoprotein B-containing lipoproteins, is a simpler and better screening tool for assessment of CAD risk in adults.12–15 Because CAD risk factors, including lipoprotein variables, persist or track from childhood to adulthood,16–18 it is imperative to examine simultaneously the utility of childhood non–HDL cholesterol levels versus the aforementioned lipoprotein measures in predicting subclinical atherosclerosis in adulthood. The present study examines this aspect, using adult carotid intima-media thickness (IMT) as an indicator of subclinical atherosclerosis and future CAD risk,19–22 in the Bogalusa Heart Study cohort.
The Bogalusa Heart Study is a long-term investigation of the natural history of cardiovascular disease beginning in childhood, in the biracial (white: 65%; black: 35%) community of Bogalusa, Louisiana.23 The current study is based on data for 1061 subjects (white: 71%; male: 43%) 24 to 43 years of age (mean age: 31.9 years; SD: 2.9 years) who were examined in 2000–2001 as part of a long-term follow-up survey and underwent bilateral, far-wall, carotid IMT measurements in all 3 carotid artery segments. Of those subjects, 437 adult subjects (mean age: 31.9 years; SD: 3.0 years; white: 70%; male: 40%) who participated 16 to 19 years earlier, as children 5 to 17 years of age, in 1981–1982 (mean age: 10.8 years; SD: 3.4 years) or 1984–1985 (mean age: 12.4 years; SD: 2.5 years) surveys and had data on both apolipoprotein A-I and apolipoprotein B levels, as well as routinely measured lipoprotein levels and other risk factor variables, were selected for this study. In childhood, the selected subjects fasted for 12 hours before screening.
Standardized protocols were used by trained examiners. Height and weight were measured twice, and the mean values were used to calculate BMI, as a measure of overall adiposity. Blood pressure was measured in triplicate, with mercury sphygmomanometers, by 2 nurses, while subjects were relaxed and seated; the mean of 6 replicate readings was used.
Cholesterol and triglycerides levels were measured with a Technicon AutoAnalyzer II (Technicon Instrument Corp, Terrytown, NY), according to the laboratory manual of the Lipid Research Clinics Program.24 Serum HDL cholesterol levels were analyzed with the heparin-calcium precipitation procedure,25 LDL cholesterol levels were estimated with the equation described by Friedewald et al,26 and non–HDL cholesterol levels were obtained by subtracting HDL cholesterol levels from total cholesterol levels. The laboratory is being monitored for precision and accuracy by the Lipid Standardization Program of the Centers for Disease Control and Prevention (Atlanta, GA). Serum apolipoprotein A-I and apolipoprotein B levels were assayed with the electroimmunoassay procedure of Laurell,27 as described previously.28 The reproducibility of directly measured lipoprotein variables was evaluated on the basis of 10% randomly assigned pairs of blind duplicate analyses; intraclass correlation coefficients were 0.96 for total cholesterol levels, 0.99 for triglyceride levels, 0.94 for HDL cholesterol levels, 0.96 for apolipoprotein B levels, and 0.84 for apolipoprotein A-I levels.
As described previously,29 images of common, bulb, and internal carotid artery segments were recorded bilaterally by using a Toshiba Sonolayer SSH160A scanner (Toshiba Medical, Tokyo, Japan) and a 7.5-MHz linear array transducer, according to the protocol developed for the Atherosclerosis Risk in Communities Study.30 The maximal IMT readings of 3 right and 3 left far walls for common, bulb, and internal segments were averaged to obtain the overall composite IMT. With respect to reproducibility, the absolute value of the mean difference between original and repeat measurements performed after 10 to 12 days for the average carotid IMT of all segments was 0.05 mm (SD: 0.03 mm; n = 75), which is comparable to that noted previously.31
Statistical analyses were performed by using SAS 9.1 (SAS, Carey, NC). Non–HDL cholesterol level, triglyceride level, HDL cholesterol level, apolipoprotein A-I level, apolipoprotein B level, total cholesterol/HDL cholesterol ratio, and carotid IMT values were logarithmically transformed to approach normality. Analysis of covariance, controlling for age, was used to test the race and gender differences in study variables.
The age-, race-, and gender-specific top decile (versus the lower 9 deciles) was chosen to classify individuals as having excess carotid IMT in adulthood. It was shown that an average IMT of all carotid artery segments of ≥1.0 mm was associated prospectively with increased CAD risk19; this cutoff point corresponds approximately to the 90th percentile distribution noted earlier in the Bogalusa Heart Study cohort of younger adults.29 With respect to childhood lipoprotein measures, age-, race-, and gender-specific top quartile (versus the lower 3 quartiles) cutoff points were used to classify children at considerably higher risk for CAD.32
Logistic regression models were used to estimate the odds ratio and 95% confidence interval (CI) for each childhood lipoprotein measure as a risk factor for excess carotid IMT in adulthood. Model 1 measured the effect of each childhood lipoprotein measure, as a main risk factor, with adjustment for nonlipoprotein covariates (follow-up years, childhood BMI, and childhood systolic blood pressure); model 2 measured the effect of each childhood lipoprotein measure with adjustment for nonlipoprotein covariates and 7 other childhood lipoprotein measures.
The discriminative value of each childhood lipoprotein measure in predicting excess carotid IMT in adulthood was evaluated by using a multivariate c statistic representing the area under the receiver operating characteristic curve. Each model contained 1 childhood lipoprotein measure as a predictor, with adjustment for the aforementioned nonlipoprotein covariates. The areas under the curve (c values) were tested for equality, followed by pairwise comparison of each model with the other models. Areas under the curve of >0.5 indicated increased predictive ability (null hypothesis, true area: 0.5).
Mean values for childhood age, BMI, systolic blood pressure, and lipoprotein measures, as well as adulthood carotid IMT, for the study cohort are given in Table 1. Black subjects, compared with white subjects, had significantly higher HDL cholesterol levels, apolipoprotein A-I levels (male subjects only), and carotid IMT (male subjects only) and lower triglyceride levels. Male subjects, compared with female subjects, displayed significantly higher systolic blood pressure, apolipoprotein B/apolipoprotein A-I ratios (white subjects only), and carotid IMT and lower apolipoprotein A-I levels (white subjects only).
Mean adulthood carotid IMT values, according to quartiles of childhood lipoprotein measures, in the cohort are presented in Fig 1. The adulthood carotid IMT showed a significant increasing trend across childhood quartiles of non–HDL cholesterol levels (P = .02), LDL cholesterol levels (P = .03), apolipoprotein B levels (P = .01), apolipoprotein B/apolipoprotein A-I ratios (P = .01), and total cholesterol/HDL cholesterol ratios (P = .02).
Figure 2 shows the results of multivariate logistic regression analyses of each childhood lipoprotein measure (top quartile versus lower 3 quartiles) as a risk factor for predicting excess IMT (top decile versus lower 9 deciles) in adulthood. When each lipoprotein measure was adjusted for nonlipoprotein covariates only (model 1) or non–HDL cholesterol covariates and other 7 lipoprotein measures (model 2), non–HDL cholesterol level, LDL cholesterol level, total cholesterol/HDL cholesterol ratio, apolipoprotein B level, and apolipoprotein B/apolipoprotein A-I ratio emerged as significant risk factors in both models, with respective odds ratios of 1.84 (95% CI: 1.10–3.14), 2.20 (95% CI: 1.29–3.74), 1.69 (95% CI: 1.0–2.90), 1.86 (95% CI: 1.1–3.19), and 2.06 (95% CI: 1.21–3.52) for model 1 and 2.60 (95% CI: 1.41–4.78), 2.95 (95% CI: 1.66–5.25), 1.78 (95% CI: 1.0–3.51), 1.44 (95% CI: 1.0–2.80), and 1.69 (95% CI: 1.0–3.3) for model 2. The odds ratios for other lipoprotein measures were not significant in either model.
In terms of discriminative value of different childhood lipoprotein measures assessed with c statistics (Table 2), none of the other lipoprotein measures was significantly better than non–HDL cholesterol levels in predicting excess carotid IMT in adulthood. The predictive models produced c values ranging from 0.62 to 0.66 for these lipoproteins measures, which were similar to each other in magnitude.
This community-based longitudinal cohort study provides direct comparative data on the utility of childhood non–HDL cholesterol levels versus other lipoprotein measures in predicting subclinical atherosclerosis, measured as carotid IMT, in adulthood 16 to 19 years later. In this cohort, childhood non–HDL cholesterol levels, LDL cholesterol levels, total cholesterol/HDL cholesterol ratios, apolipoprotein B levels, and apolipoprotein B/apolipoprotein A-I ratios emerged as significant predictors of excess carotid IMT in adulthood. It is noteworthy that none of the childhood lipoprotein measures was superior to non–HDL cholesterol levels in this regard.
Earlier studies, including our own, showed independent associations between carotid IMT values for young adults and their childhood LDL cholesterol and total cholesterol levels.33–35 Importantly, with respect to early subclinical coronary atherosclerosis, autopsy studies of children and young adults showed significant consistent associations of adverse levels of LDL cholesterol and non–HDL cholesterol with the extent of lesions in the coronary arteries.2,3 Current findings with respect to childhood non–HDL cholesterol and LDL cholesterol levels are consistent with those earlier reports. However, no corresponding prospective data regarding other childhood lipoprotein measures are, to our knowledge, available in the literature for comparison.
As a CAD risk assessment lipoprotein measure in childhood, the non–HDL cholesterol level has certain advantages. It is a composite measure of LDL, lipoprotein(a), very low-density lipoprotein, very low-density lipoprotein remnant, and intermediate-density lipoprotein levels36,37 and consequently is a strong correlate and surrogate measure of apolipoprotein B levels.13 It has also been suggested that the non–HDL cholesterol level could be a marker for certain nonlipid risk factors for the metabolic syndrome.38 Earlier we found that elevated non–HDL cholesterol levels, like LDL cholesterol levels, in childhood not only tracked (persisted) reasonably in ranking over a 27-year period into adulthood and predicted adult dyslipidemia but also, unlike LDL cholesterol levels, were associated with nonlipid risk factors characteristic of the metabolic syndrome.39
As a limitation, this study did not compare the discriminatory value of non–HDL cholesterol relative to lipoprotein classes measured on the basis of particle size, density, or apolipoprotein composition.40–42 It should be mentioned that these specialized lipoprotein testing procedures are expensive and are not readily available for routine screening.
The present findings indicate the prognostic value of childhood non–HDL cholesterol levels in predicting adult CAD risk, assessed in terms of carotid IMT. Because overnight fasting is not required for measurements of total cholesterol and HDL cholesterol levels, from which the non–HDL cholesterol level is calculated, and these variables are well standardized, routinely analyzed, and less expensive to measure, this lipoprotein measure could be especially advantageous for CAD risk evaluation in pediatric populations. Furthermore, pediatric population-based frequency distribution data and clinically useful (but not yet validated) values equivalent to selected LDL cholesterol cutoff points recommended by the National Cholesterol Education Program pediatric panel6 are currently available.43 Additional studies are obviously needed in this direction.
This work was supported by grant AG 16592 from the National Institute on Aging, grant HL 38855 from the National Heart, Lung, and Blood Institute, and grant HD 043820 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 gratefully acknowledged. We especially thank the study participants.
- Accepted September 6, 2007.
- 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.
What's Known on This Subject
Non–HDL cholesterol levels are related strongly to the extent of lesions in coronary arteries. Cross-sectional studies suggested that non–HDL cholesterol levels are as good as other lipoprotein measures in the identification of preclinical atherosclerosis.
What This Study Adds
This longitudinal study adds evidence on the utility of childhood non–HDL cholesterol measurements in the prediction of atherosclerotic changes in the carotid artery in adulthood.
- ↵Rainwater DL, McMahan CA, Malcom GT, et al. Lipid and apolipoprotein predictors of atherosclerosis in youth: apolipoprotein concentrations do not materially improve prediction of arterial lesions in PDAY subjects. Arterioscler Thromb Vasc Biol.1999;19 (3):753– 761
- ↵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 (19):2486– 2497
- ↵National Cholesterol Education Program. Report of the Expert Panel on Blood Cholesterol Levels in Children and Adolescents. Pediatrics.1992;89 (3):525– 584
- ↵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 (3):273– 278
- ↵Castelli WP, Abbott RD, McNamara PM. Summary estimates of cholesterol used to predict coronary heart disease. Circulation.1983;67 (4):730– 734
- ↵Bittner V, Hardison R, Kelsey SF, Weiner BH, Jacobs AK, Superko G. Non–high-density lipoprotein cholesterol levels predict five year outcome in the Bypass Angioplasty Revascularization Investigation (BARI). Circulation.2002;106 (20):2537– 2542
- ↵Lauer RM, Lee J, Clarke WR. Factors affecting the relationship between childhood and adult cholesterol levels: the Muscatine Study. Pediatrics.1988;82 (3):309– 318
- ↵Chambless LE, Heiss G, Folsom AR, Rosamod W, Szklo M, Clegg X. Association of coronary heart disease incidence with carotid arterial wall thickness and major risk factors: the Atherosclerosis Risk in Communities (ARIC) Study, 1987–1993. Am J Epidemiol.1997;146 (6):483– 494
- Hulthe J, Wikstrand J, Emmanuelsson H, Wiklund O, de Feyter PJ, Wendelhag I. Atherosclerosis changes in the carotid artery bulb as measured by B-mode ultrasound are associated with extent of coronary atherosclerosis. Stroke.1997;28 (6):1189– 1194
- ↵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
- ↵Friedewald W, Levy R, Frederickson DS. Estimation of the concentration of low density lipoprotein cholesterol in plasma, with the use of the preparative ultracentrifuge. Clin Chem.1972;18 (6):499– 502
- ↵Laurell CB. Electroimmunoassay. Scand J Clin Lab Invest.1972;2 (suppl 124):21– 37
- ↵Srinivasan SR, Freedman DS, Sharma C, Webber LS, Berenson GS. Serum apolipoproteins A-I and B in 2,854 children from a biracial community: the Bogalusa Heart Study. Pediatrics.1986;78 (2):189– 200
- ↵Riley WA, Barnes RW, Evans GW, Burke GL. Ultrasonic measurement of the elastic modulus of the common carotid artery: the Atherosclerosis Risk in Communities (ARIC) Study. Stroke.1992;23 (8):1062– 1068
- ↵Davis PH, Dawson JD, Riley WA, Lauer RM. Carotid intima-media thickness is related to cardiovascular risk factors measured from childhood through middle age: the Muscatine Study. Circulation.2001;104 (23):2815– 2819
- ↵Grundy SM. Low-density lipoprotein, non–high-density lipoprotein, and apolipoprotein B as targets of lipid lowering therapy. Circulation.2002;106 (20):2526– 2529
- ↵Srinivasan SR, Frontini MG, Xu J, Berenson GS. Utility of childhood non–high-density lipoprotein cholesterol levels in predicting adult dyslipidemia and other cardiovascular risks: the Bogalusa Heart Study. Pediatrics.2006;118 (1):201– 206
- Freedman DS, Otvos JD, Jeyarajah EJ, Barboriak JJ, Anderson AJ, Walker JA. Relation of lipoprotein subclasses as measured by proton nuclear magnetic resonance spectroscopy to coronary artery disease. Arterioscler Thromb Vasc Biol.1998;18 (7):1046– 1053
- ↵Srinivasan SR, Myers L, Berenson GS. Distribution and correlates of non–high-density lipoprotein cholesterol in children: the Bogalusa Heart Study. Pediatrics.2002;110 (3). Available at: www.pediatrics.org/cgi/content/full/110/3/e29
- Copyright © 2008 by the American Academy of Pediatrics