The contribution of childhood obesity to adult carotid intima-media thickness: the Bogalusa Heart Study.

OBJECTIVE: Although obese children are at increased risk for coronary heart disease in later life, it is not clear if the association results from the persistence of childhood obesity into adulthood. We examined the relation of both childhood and adult levels of body mass index (BMI, kg m(-2)) to carotid intima-media thickness (IMT) measured at the (mean) age of 36 years. DESIGN AND SUBJECTS: Prior to the determination of adult IMT, the 1142 participants had been examined 7 (mean) times in the Bogalusa Heart Study. MEASUREMENTS: In addition to BMI, levels of lipids, lipoproteins and blood pressure were measured at each examination. Cumulative levels of each risk factor were based on the areas under the individual growth curves calculated using multilevel models for repeated (BMI) measurements. We then examined the relation of these cumulative levels to adult IMT. RESULTS: Carotid IMT was associated with cumulative levels of BMI in both childhood and adulthood (P<0.001 for each association). Furthermore, the association between childhood BMI and adult IMT persisted, but was reduced, after controlling for adult BMI. Although childhood levels of lipids, lipoproteins and blood pressure were also associated with adult IMT, these associations were not independent of adult levels of these risk factors. CONCLUSIONS: These results emphasize the adverse effects of elevated childhood BMI levels. In addition to the strong tracking of BMI levels from childhood to adulthood, there appears to be a modest, independent effect of childhood BMI on adult IMT. The prevention of childhood obesity should be emphasized.

Persistence of high diastolic blood pressure in thin children. The Bogalusa Heart Study.

Relationships between initial anthropometric variables and subsequent diastolic blood pressure (fourth phase) were examined in children identified as being in the upper quintile for diastolic blood pressure at Year 1. Of 156 white children, aged 10 to 14 years, with diastolic blood pressure levels in the upper age-race-sex-specific quintile at Year 1, 38% remained in the upper quintile at Year 4. However, there was a definite trend for leaner children, defined by ponderosity (weight/height3) to remain in the highest diastolic blood pressure quintile (p less than 0.001). Of white children originally identified in the highest quintile for diastolic blood pressure and the lowest quintile for ponderosity (lean group), 67% (18 of 27) remained in the upper quintile at Year 4. In contrast, only 21% (11 of 52) of white children identified as being in the highest quintile for both diastolic blood pressure and ponderosity (obese group) at Year 1 were in the upper diastolic blood pressure quintile at Year 4. Similar results were seen in children examined 5 years later. Pearson correlation coefficients and linear regression analyses confirmed the negative relationship between initial ponderosity and subsequent diastolic blood pressure, especially in older children. A similar relationship was noted in black children. Potential differences in the etiological process of obesity-related and non-obesity-related high blood pressure were examined. These observations indicate that characteristics other than obesity can contribute to high blood pressure in late childhood.

Body fat patterning and blood pressure in children and young adults. The Bogalusa Heart Study.

The relationship of central body fat (measured by subscapular skinfold) and peripheral body fat (measured by triceps skinfold) to blood pressure was investigated in 3784 subjects aged 5 to 24 years old from the biracial community of Bogalusa, Louisiana. After adjustment for height, age, sex, and race, significant relationships were found for both central body fat (r = 0.19 and 0.14, p less than 0.0001) and peripheral body fat (r = 0.15 and 0.12; p less than 0.0001) with systolic and diastolic (fourth phase) blood pressure, respectively. However, the relationship between peripheral body fat and blood pressure, after controlling for the level of central body fat, was negligible (r = 0.00 and 0.01 for systolic and diastolic blood pressure, respectively). In contrast, the central body fat-blood pressure relationship remained statistically significant even after controlling for the peripheral body fat level. For central body fat, the partial correlations with systolic blood pressure were highest in young children (r = 0.15), dropped slightly during adolescence (r = 0.12), and became nonsignificant only in 18- to 24-year-old female subjects; correlations remained high in both black and white 18- to 24-year-old male subjects (r = 0.18 and 0.16, respectively). Mean levels of systolic blood pressure from the lowest to the highest quartile of central body fat ranged from 100.4 to 108.9 mm Hg. The adult hypertension-central body fat relationship, which has been shown by others, appears to exist in children. Continued efforts at early identification and prevention of obesity in children are warranted.

Relation of circumferences and skinfold thicknesses to lipid and insulin concentrations in children and adolescents: the Bogalusa Heart Study.

BACKGROUND: Although body fat patterning has been related to adverse health outcomes in adults, its importance in children and adolescents is less certain. OBJECTIVE: We examined the relation of circumference (waist and hip) and skinfold-thickness (subscapular and triceps) measurements to lipid and insulin concentrations among 2996 children and adolescents aged 5-17 y. DESIGN: This was a community-based, cross-sectional study conducted in 1992-1994. RESULTS: A central or abdominal distribution of body fat was related to adverse concentrations of triacylglycerol, LDL cholesterol, HDL cholesterol, and insulin; these associations were independent of race, sex, age, weight, and height. These associations were observed whether fat patterning was characterized by using 1) waist circumference alone (after adjustment for weight and height), 2) waist-to-hip ratio, or 3) principal components analysis. Compared with a child at the 10th percentile of waist circumference, a child at the 90th percentile was estimated to have, on average, higher concentrations of LDL cholesterol (0.17 mmol/L), triacylglycerol (0.11 mmol/L), and insulin (6 pmol/L) and lower concentrations of HDL cholesterol (-0.07 mmol/L). These differences, which were independent of weight and height, were significant at the 0.001 level and were consistent across race-sex groups. CONCLUSIONS: These findings emphasize the importance of obtaining information on body fat distribution, waist circumference in particular, in children. Waist circumference, which is relatively easy to measure, may help to identify children likely to have adverse concentrations of lipids and insulin.

Relation of body fat patterning to lipid and lipoprotein concentrations in children and adolescents: the Bogalusa Heart Study.

Although a truncal distribution of adipose tissue in adults is associated with several metabolic complications, its importance in early life has received little attention. The relation of several anthropometric measures to serum concentrations of lipids, lipoproteins, and apolipoproteins was therefore examined in 361 children who were between ages 6 and 18 y. (Children had been selected previously because of extreme levels of very-low-density- and low-density-lipoprotein cholesterol.) Analyses revealed two groups of anthropometric variables: truncal measures (waist circumference and subscapular, subcostal, and suprailiac skinfold thicknesses) and thickness of peripheral skinfolds (femoral, triceps, calf, and biceps). After generalized obesity was adjusted for children with high concentrations of both cholesterol fractions had more truncal fat but less peripheral fat than did children with low lipoprotein cholesterol concentrations. A truncal fat pattern was also associated with decreased concentrations of high-density-lipoprotein cholesterol and apolipoprotein A-1. Knowledge of fat patterning may help identify persons prone to hyperlipidemia.

Relation of body fat distribution to hyperinsulinemia in children and adolescents: the Bogalusa Heart Study.

The relation of body fat distribution to plasma levels of glucose and insulin during an oral glucose tolerance test was examined in 355 Black and White school-age children. Both central and peripheral fat were similarly related to fasting, 30-min, and 1-h glucose. Unlike peripheral fat, central body fat was more strongly related to the 1-h insulin response (r = 0.35 vs 0.26); this association remained significant for central fat independent of peripheral fat (r = 0.18). The strong relation of central fat to insulin response was noted in both races and sexes but not in either sexually immature or relatively thin children. These findings indicate that, even in early life, a central body fat pattern relates positively to insulin response to glucose load. Thus, knowledge of body fat localization may help identify persons most susceptible to hyperinsulinemia in early life.

Levels and correlates of LDL and VLDL particle sizes among children: the Bogalusa heart study.

Levels of lipids and lipoproteins among children vary by sex and race/ethnicity, and are correlated with age, obesity, and other characteristics. There is, however, little information on the distribution and correlates of low-density lipoprotein (LDL) and very-low-density lipoprotein (VLDL) subclasses in early life. We used nuclear magnetic resonance (NMR) spectroscopy to determine mean LDL and VLDL particle sizes among 10- to 17-year-olds (n=918) who participated in the 1992-94 examination of the Bogalusa heart study. As compared with girls, boys had a smaller (0.1 nm) mean LDL particle size and a larger (0.9 nm) mean VLDL size; furthermore, the average size of VLDL particles increased with age among white boys but not among other children. Although there were also black/white differences in particle sizes, with black children having larger LDL and smaller VLDL particles, these racial contrasts could be attributed to differences in lipid levels. Levels of triglycerides, insulin, and relative weight were associated with the size of VLDL (positive) and LDL (negative) particles. These results suggest that the analysis of lipoprotein subclasses may provide a better understanding of the role of various risk factors in the development of coronary heart disease

The relation of atherosclerotic lesions to antemortem and postmortem lipid levels: the Bogalusa Heart Study.

Although postmortem lipid levels have been used as surrogates for levels during life, it is uncertain whether atherosclerotic lesions are related similarly to antemortem and postmortem lipid values. In a sample of 23 children and young adults who had been examined for cardiovascular disease risk factors and subsequently died from violent causes, we examined the relation of (a) postmortem lipid levels to values obtained 1 to 14 years earlier, and (b) atherosclerotic lesions to antemortem and postmortem lipid levels. Postmortem levels of triglycerides and very-low-density lipoprotein cholesterol (VLDLC) were higher than levels during life, but postmortem levels of total cholesterol (TC), low-density lipoprotein cholesterol (LDLC) and high-density lipoprotein cholesterol (HDLC) were related to antemortem levels (rs > 0.40). After excluding eight persons who likely received large volumes of intravenous fluids before death, the within-person variability between antemortem and postmortem levels of LDLC and HDLC was similar to the antemortem variability. Furthermore, the relation of atherosclerotic lesions to antemortem and postmortem lipid levels differed only slightly for TC, LDLC and HDLC. In contrast, lesions in the coronary arteries showed the strongest association with antemortem VLDLC levels, but were not associated with postmortem VLDLC levels. Despite the very small number of subjects, our results suggest that if intravenous fluids are not administered before death, postmortem levels of TC, LDLC and HDLC are fairly representative of levels during life. Postmortem levels of VLDLC or triglycerides, however, should not be used as surrogates for antemortem levels.

Risk factors and the anatomic distribution of coronary artery disease.

Differences in the importance of risk factors according to the anatomic location of coronary artery disease (CAD) were assessed in 4722 men and 1069 women who underwent arteriography. Examined characteristics included total and high-density lipoprotein (HDL)-cholesterol, triglycerides, obesity, smoking, alcohol consumption, diabetes, and hypertension. Of these risk factors, the ratio of total to HDL-cholesterol showed the highest correlation with the overall severity of CAD (r = 0.24, men; r = 0.38, women); in contrast, its relation to left main (LM) disease was much lower (r = 0.10, men; r = 0.08 women) than were correlations with stenotic disease in the left anterior descending, circumflex, and right coronary arteries. Other risk factors also showed weaker associations with LM disease than with stenoses in other vessels, and none was related to increased LM disease after controlling for disease in other vessels. For example, as compared with men who had no significant CAD, those with 1-, 2-, and 3-vessel disease had mean increases in total cholesterol of 12, 18, and 19 mg/dl, respectively. In contrast, after adjusting for disease in other vessels, LM disease (present in 293 men) was associated with only a 4 mg/dl increase in mean cholesterol levels (P = 0.20). These results indicate that the relation of risk factors to CAD differs according to the location of the stenotic disease, and that LM disease is poorly predicted by the standard risk factors.

Polymorphism in the 5'-flanking region of the insulin gene and its potential relation to cardiovascular disease risk: observations in a biracial community. The Bogalusa Heart Study.

The gene frequencies of a polymorphism in the 5'-flanking region of the insulin gene and its relationship to cardiovascular disease risk were studied in a well defined population of children (mean age 5.5 years) from a biracial community. The BglII endonuclease was used for digestion of the DNA around this polymorphic region. The risk factors studied included parental and grandparental self-reported histories of myocardial infarction and diabetes mellitus, fasting glucose and insulin levels, lipoprotein, cholesterol, and triglyceride levels, and skinfold thicknesses and weight. Four alleles were observed at this locus, with the class 2 allele being significantly more common among blacks than whites. Among white children, the class 3 allele was associated with increased risk for grandparental diabetes mellitus. White children with 2 copies of the class 3 allele had significantly higher levels of glucose. Black children with a copy of the class 3 allele had significantly higher levels of insulin. This study indicates that the class 3 allele is potentially associated with risk for diabetes mellitus and coronary heart disease that can be observed in childhood.

Cardiovascular risk factors from birth to 7 years of age: the Bogalusa Heart Study. Design and participation.

Cardiovascular risk factor variables were examined in a cohort of 440 infants from birth through 7 years of age. Anthropometric measures, BP, serum lipid and lipoprotein values, and dietary intake data were obtained according to detailed protocols. Various quality controls to ensure the collection of valid and reliable data were instituted. Participation remained high throughout the study with 80% of the children examined at 6 months and 60% at 7 years of age. Rates were slightly higher for black than for white children. Children born in the private hospital were more likely to continue in the study than children born in the charity hospital. Children of parents examined when the child was 2 years of age were more likely to be examined during the preschool phases, but parental examination was not related to child examination when the child was 7 years of age. As in studies of school-aged children, measurement errors were lowest for height, weight, and serum total cholesterol. Measurement errors for BP were highest at the younger ages, particularly for diastolic BP. Examination of a newborn cohort throughout time affords the opportunity to study early development of relationships and tracking of cardiovascular risk factors.

Value of childhood blood pressure measurements and family history in predicting future blood pressure status: results from 8 years of follow-up in the Bogalusa Heart Study.

The value of BP measurements and family history of cardiovascular disease in predicting future BP status was studied in 1,501 children, initially 2 to 14 years of age, who were examined four times during an 8-year period in the Bogalusa Heart Study. Correlation coefficients between year 1 and year 9 BPs were as follows for systolic and diastolic BPs, respectively: 0.41 and 0.35 (P less than .0001). These correlations were significant in all age groups. For children in the upper quartile of BP at any one prior examination, the percentage remaining in the year 9 upper quartile ranged from 41% to 52% for systolic BP and 35% to 44% for diastolic BP. Three serial BP measurements in the upper quartile increased the percentages remaining in the upper quartile to 68% for systolic BP and 62% for diastolic BP. Conversely, of those children not in the upper quartile of systolic BP at year 9, 96.8% did not have all three prior measurements in the upper quartile. Family history of hypertension was shown to independently predict year 9 systolic BP status. These results confirm the importance of serial BP measurements and family history of hypertension for the practicing physician.

Serum apolipoproteins A-I and B in 2,854 children from a biracial community: Bogalusa Heart Study.

Serum apolipoprotein A-I (apo A-I) and apolipoprotein B (apo B) profiles were examined in 2,854 children, 5 to 17 years of age, from a total biracial community. Black boys had higher apo A-I levels than white boys (P less than .001), whereas girls showed no such race-related difference. Black-white difference in apo A-I persisted among boys with similar triglyceride levels provided that triglyceride levels were high. The ratio of high-density lipoprotein cholesterol (HDL-C)/apo A-I was higher in black than in white children, irrespective of sex (P less than .001). Only black children showed sex-related differences for apo A-I (boys greater than girls, P less than .05). Sex-related differences were seen in white children for HDL-C/apo A-I ratio (boys greater than girls, P less than .001) and in children of both races for apoB (girls greater than boys, P less than .01). Age-related changes were more apparent for apo A-I and HDL-C/apo A-I ratio than for apo B. A progressive decrease in apo A-I was noted during sexual maturation only in white boys. The magnitude of inverse association of apo B to HDL-C was less strong in black children (P less than .01). Although apo A-I was inversely correlated with very low-density lipoprotein cholesterol and triglycerides in white children, no association was noted in black children. These findings are indicative of intrinsic metabolic differences among the race-sex groups, resulting in variability in lipoprotein composition and levels and atherogenic potential.

Tracking of serum cholesterol levels in a multiracial sample of preschool children.

The relation of an initial measurement of serum total cholesterol to subsequent levels over a (mean) 13-month interval was examined in a multiracial (white, Hispanic, American Indian, and black) sample of 1680 one- to four-year-olds. Although the relation of the initial level to the final measurement (r = .54) did not vary by race, sex, relative weight, or changes in relative weight, the association increased with age at the time of the initial measurement (eg, r = .64 among 4-year-olds). Based on the initial and final total cholesterol determinations, the within-person standard deviation was 21 mg/dL and the coefficient of variation was 13%. Although the final total cholesterol level was within 5 mg/dL of the initial level for 18% of the children, the two determinations differed by greater than or equal to 25 mg/dL for about 35% of the children and by greater than or equal to 50 mg/dL for about 8%. Of the 149 children who had an initial cholesterol level greater than or equal to 200 mg/dL, 34% (about five times the expected number) had a follow-up level that was similarly elevated whereas 25% had a subsequent measurement below 170 mg/dL. The results indicate that although an initial cholesterol level in early life is moderately predictive of subsequent levels, it may be difficult to interpret a single total cholesterol determination because of substantial within-person variability.

Cardiovascular risk factors from birth to 7 years of age: the Bogalusa Heart Study. Serum lipids and lipoproteins.

Serum lipids and lipoprotein cholesterol fractions were examined in a newborn cohort that was followed from birth to 7 years of age. Although white and female infants had higher cord blood levels of high-density lipoprotein cholesterol (HDL-C) than did black and male infants, respectively, these differences did not persist throughout early childhood. Mean levels of all serum lipids and lipoproteins increased greatly in the first 6 months of life, and by 2 years of age, levels approached those seen in adolescents. Infants consuming cow's milk had higher 6-month levels of serum total cholesterol and low-density lipoprotein cholesterol than did formula-fed infants. However, milk source in infancy did not significantly influence total cholesterol or low-density lipoprotein cholesterol levels at age 7 years. Serum lipid and lipoprotein levels at age 7 years were associated with previously measured levels as early as 6 months of age, and infants with unfavorable levels were likely to have similar adverse levels at 7 years of age. In addition, increases in obesity between 6 months and 7 years of age were positively associated with increases in levels of serum triglycerides. These results suggest that certain persons at increased risk for cardiovascular disease can be identified in infancy.

The relation of overweight to cardiovascular risk factors among children and adolescents: the Bogalusa Heart Study.

BACKGROUND: Although overweight and obesity in childhood are related to dyslipidemia, hyperinsulinemia, and hypertension, most studies have examined levels of these risk factors individually or have used internal cutpoints (eg, quintiles) to classify overweight and risk factors. OBJECTIVE: We used cutpoints derived from several national studies to examine the relation of overweight (Quetelet index, >95th percentile) to adverse risk factor levels and risk factor clustering. DESIGN: The sample consisted of 9167 5- to 17-year-olds examined in seven cross-sectional studies conducted by the Bogalusa Heart Study between 1973 and 1994. RESULTS: About 11% of examined schoolchildren were considered overweight. Although adverse lipid, insulin, and blood pressure levels did not vary substantially with the Quetelet index at levels <85th percentile, risk factor prevalences increased greatly at higher levels of the Quetelet index. Overweight schoolchildren were 2.4 times as likely as children with a Quetelet index <85th percentile to have an elevated level of total cholesterol. Odds ratios for other associations were 2.4 (diastolic blood pressure), 3.0 (low-density lipoprotein cholesterol), 3.4 (high-density lipoprotein cholesterol), 4.5 (systolic blood pressure), 7.1 (triglycerides), and 12.6 (fasting insulin). Several of these associations differed between whites and blacks, and by age. Of the 813 overweight schoolchildren, 475 (58%) were found to have at least one risk factor. Furthermore, the use of overweight as a screening tool could identify 50% of schoolchildren who had two or more risk factors. CONCLUSIONS: Because overweight is associated with various risk factors even among young children, it is possible that the successful prevention and treatment of obesity in childhood could reduce the adult incidence of cardiovascular disease.

Black/white differences in risk factors for arteriographically documented coronary artery disease in men.

Although the leading cause of death among black men in the United States is coronary artery disease (CAD), risk factors have not been well documented in black populations. Therefore, possible racial differences in the relation of several characteristics to the extent of CAD were assessed in 4,722 white and 169 black men who underwent arteriography. Associations between an occlusion score (ranging from 0 to 300), reflecting the severity of CAD, and levels of total and high-density lipoprotein (HDL) cholesterol, triglycerides, cigarette smoking, alcohol intake, relative weight, systemic hypertension and diabetes mellitus were examined. Most risk factors were significantly related to the extent of CAD in both races, but lipid levels showed stronger associations with CAD among blacks: correlations between CAD and total cholesterol were 0.16 (whites) vs 0.29 (blacks) and associations with HDL cholesterol were -0.22 (whites) vs -0.49 (blacks). In addition, at adverse levels of certain risk factors, blacks had more extensive CAD than did whites: mean occlusion scores were 148 (whites) and 238 (blacks) at HDL cholesterol levels less than 30 mg/dl. As assessed by multiple linear regression, however, only triglyceride levels were differentially related to CAD between whites (beta = 0) and blacks (beta = 0.47), p less than 0.01 for racial contrast. These results document the importance of risk factors in black men and indicate black/white differences in the relation of triglycerides to CAD.

Interpretation of linear regression models that include transformations or interaction terms.

In linear regression analyses, we must often transform the dependent variable to meet the statistical assumptions of normality, variance stability, or linearity. Transformations, however, can complicate the interpretation of results because they change the scale on which the dependent variable is measured. In this setting, the inclusion of product terms or the transformation of some independent (or predictor) variables may further complicate interpretation. In this article, we present some interpretations of linear models that include transformations or product terms. We illustrate these interpretations using regression analyses designed to study determinants of serum testosterone levels. These examples show how one can present results using simple measures, such as medians, and interpret regression parameters.

Correlates of leukocyte counts in men.

Because of previously reported associations between a high leukocyte count and risk of ischemic heart disease (IHD), we examined the relation of leukocyte counts to various characteristics among 3591 white and 506 black 31- to 45-year-old men. The mean leukocyte count was approximately 1000 cells/microL higher among whites than among blacks, and approximately 1900 cells/microL higher among current smokers than among nonsmokers. The leukocyte count was also higher among men who had recently stopped smoking and among men who reported their general health as poor or fair. Independent of these relations, the leukocyte count was associated positively with the platelet count (r = 0.29), triglyceride level (r = 0.21), heart rate (r = 0.15), and use of corticosteroids and beta-blockers; and inversely with alcohol consumption and prothrombin time (r = -0.10). The examined characteristics could together account for 37% of the variability in leukocyte counts. These relatively strong associations indicate that it may be difficult to disentangle the relation of the leukocyte count to IHD from that of other risk factors.

Cigarette smoking and leukocyte subpopulations in men.

Because of previously reported associations among the total leukocyte count, cigarette smoking, and risk of cardiovascular disease, we examined the relation of cigarette smoking to various leukocyte subpopulations among 3467 men aged 31 to 45 years. The median total leukocyte count was 36% higher (7840 vs. 5760 cells/mL) among current cigarette smokers than among men who had never smoked, and both stratification and regression analyses were used to examine independent associations with leukocyte subpopulations. At equivalent counts of other subpopulations, CD4+ lymphocytes and neutrophils were the cell types most strongly associated with cigarette smoking; each standard deviation change in counts of these subpopulations increased the odds of current (vs. never) smoking by approximately threefold. Furthermore, whereas 15% of the 238 men with relatively low (< 25 percentile) counts of both neutrophils and CD4+ lymphocytes were cigarette smokers, 96% of the 249 men with relatively high counts of both subpopulations were current smokers. Counts of T lymphocytes also tended to be higher among the 32 men with self-reported ischemic heart disease than among other men. These results, along with previous reports of immunologically active T lymphocytes in atherosclerotic plaques, suggest that this subpopulation may be of particular interest in studies examining the relation of leukocytes to cardiovascular disease.