Differences in HDL Cholesterol Concentrations in Japanese, American, and Australian Children
Background Mortality from coronary heart disease is relatively low in Japan compared with other developed countries and has remained low despite an increasing standard of living and an apparent increase in mean plasma cholesterol concentration in adults over the past three decades. Important differences in childhood plasma lipoprotein profile might contribute to some of the difference in coronary heart disease mortality seen between Japan and both Australia and North America.
Methods and Results Plasma HDL cholesterol and total cholesterol were surveyed in representative populations of schoolchildren in Australia, Japan, and Bogalusa, La. The mean concentration of plasma HDL cholesterol (but not total cholesterol) was higher for Japanese schoolchildren than for Australian or US schoolchildren (P<.001). In addition, the difference in plasma HDL cholesterol between the ages of 8 to 10 years and 12 to 15 years was much greater for Australian (boys, 15.2%; girls, 2.6%) and US (boys, 9.1%; girls, 2.7%) children than for their Japanese counterparts (boys, 4.2%; girls, 1.9%). An examination of potential explanatory factors revealed little difference in body mass index between samples, higher physical activity levels for the Japanese compared with the Australians, and substantial differences in dietary intake between Japanese and Australian schoolchildren.
Conclusions The relatively high ratio of plasma HDL cholesterol to total cholesterol in Japanese schoolchildren and the relatively small negative difference of plasma HDL cholesterol with age may help to explain why the coronary heart disease mortality rate in Japan is low compared with that in other developed countries.
Mortality from CHD in Japan is very much lower than in most developed countries. CHD mortality for both men and women in Australia and the United States is more than five times higher than in Japan.1 Table 1⇓ provides a comparison of the age-standardized rates.
Furthermore, CHD mortality in Japan has remained low as the standard of living has increased, despite an increase in mean plasma TC concentration over the past three decades. Studies conducted in the 1950s reported a mean cholesterol concentration among adults in Japan of ≈4.0 mmol/L,2 whereas a national survey conducted in 1989 found that mean concentrations had increased to 5.26 mmol/L for men and 5.14 mmol/L for women 40 to 49 years old,3 yet age-adjusted CHD mortality for the age range 30 to 69 years fell by 24% for men and 37% for women over the period 1968 to 1978.4 This has occurred in a setting in which blood pressure for both sexes is higher than in most other developed countries5 and the proportion of male smokers is much higher than in the US and Australia.3
A number of hypotheses have been put forward to explain the apparent inconsistency between increasing westernization of lifestyle in Japan and the maintenance of low CHD mortality levels. One possibility is that the age cohort whose lifestyle has changed following economic development are not yet old enough to contribute to the CHD mortality statistics. This seems unlikely, because advances in the material standard of living have occurred over a period of at least four decades.6 That changes in lifestyle have occurred in the “at-risk” age group is reflected in the increase in serum cholesterol in those >40 as well as <40 years old.3 Furthermore, in other Asian nations whose economic development occurred later than in Japan, CHD mortality has already risen substantially.7
It could be that the high fish intake of Japanese, for example, acting via a pathway involving coagulation, may have protected them against the increase in CHD that would have been expected to follow the rise in plasma cholesterol.8 However, studies in other populations provide estimates for potential risk reduction for consumption of fish that are far short of the current difference between Japanese and American CHD mortality,9 and risk reduction resulting from fish intake does not appear to increase linearly with increased consumption.10
We propose here that an important difference in childhood lipoprotein concentration may explain some of the remaining difference in CHD mortality. Evidence from laboratory and epidemiological studies supports the hypothesis that the concentration of either plasma TC or LDL-C increases risk of CHD, whereas the concentration of HDL, usually measured as HDL-C, reduces risk in adults.11 In addition, it was shown in the Bogalusa Heart Study12 and the PDAY study13 that a higher concentration of both LDL-C and VLDL-C and a lower concentration of HDL-C in children and young adults are associated with a higher risk of premature atherosclerosis. In this article, we compare data on TC and HDL-C in representative samples of the school-age population from Japan and Australia with data from the Bogalusa study on subjects of the same age in Louisiana. In addition, measures of relevant lifestyle factors, such as diet, and anthropometric factors that might be expected to influence lipoprotein concentrations are also reported.
This article makes use of data collected in five surveys: (1) the Australian Schools Health and Fitness Survey, 198514 ; (2) the Australian National Dietary Survey of Schoolchildren, 198515 ; (3) the Japanese Society of School Health Survey, 199316 ; (4) the Japanese Physical Education and School Health Centre Dietary Survey, 1991/9217 ; and (5) data from the Bogalusa Heart Study from the period 1987 to 1993.18 19
Australian Data Collection: The Australian Schools Health and Fitness Survey, 1985
This survey was conducted on 9000 schoolchildren between the ages of 7 and 15 years (500 in each age/sex stratum). These children were selected by use of a two-stage probability sample. The first stage involved the selection of schools, in which a response rate of 90.1% was attained. The second stage of sampling consisted of the random selection of 10 boys and 10 girls in each age group from each school. Consent from both parents and child was required for a student to be included in the study and was obtained for 77.5% of students in this initial sample. Further details of the sample have been described previously.20
In addition to undergoing the same field tests as the children in the sample of 9000, a subsample of 2400 children was selected to undergo measurements of blood pressure, body fat, and endurance fitness and to have fasting blood taken for lipid analysis. This subsample consisted of 400 subjects randomly selected from each of six strata defined by age (9, 12, and 15 years) and sex. Of this subsample of 2400, 481 refused venipuncture.
Measurements were undertaken between May and October 1985. Serum lipids were measured in subjects after a 12-hour fast according to procedures of the Lipid Research Clinics.21 Blood samples were analyzed at the Flinders Medical Centre (South Australia), which throughout the study period met the criteria specified by the World Health Organization Collaborating Center for Reference and Research in Blood Lipids (Centers for Disease Control and Prevention, Atlanta, Ga). Plasma TC was determined with a Technicon Autoanalyzer 11, and HDL-C measurement was conducted after heparin manganese precipitation of apolipoprotein B–containing lipoproteins.22
Students were asked by questionnaire to record the types of physical activity and sports played during the previous week. The students were required to list the activities (including physical education classes and travel to and from school) in terms of frequency, duration, and perceived effort. Height was measured to the nearest 0.1 cm with a stadiometer with the subject in bare feet. Weight was measured to the nearest 0.5 kg with bathroom scales that were calibrated daily against a standard weight. BMI was calculated as weight in kilograms divided by the square of height in meters.
Australian National Dietary Survey of Schoolchildren, 1985
This survey was conducted in conjunction with the Australian Schools Health and Fitness Study. Its primary aim was to determine food and nutrient intake of a representative cohort of Australian schoolchildren 10 to 15 years old by use of a 24-hour dietary record.
In small groups, children were instructed in the use of the dietary record form by trained data collectors. After its completion, each student was interviewed individually, and record books were checked to confirm that all food and drink consumed had been recorded and to clarify incomplete and illegible information. British tables of food composition23 were used to estimate individual nutrient intake for the 24-hour period, with substitution of Australian food composition data for foods for which they were available. The overall response rate for the National Dietary Survey of Schoolchildren was 65.5%. Other details of the dietary survey are available elsewhere.24
Japanese Data Collection: Japanese Society of School Health Survey, 1993
This survey, supported by the Ministry of Education, was conducted by the Japanese Society of School Health to examine the health status of schoolchildren between the ages of 8 and 15 years. The study sample consisted of 9075 schoolchildren from 55 schools in 13 prefectures (not including Hokkaido).
The study was conducted over the period from April 1993 to January 1994. The survey questionnaire included sections related to physical activity and food consumption. Elementary-school children were able to take the questionnaire home to obtain the assistance of their parents in answering it. Junior high and high school children completed the questionnaire alone during school time. The questions sought information on behavior on weekdays rather than weekends.
In relation to usual physical activity, students were asked how much time they spent doing various activities in a usual week. Possible activities were listed and classified according to intensity (ie, vigorous, moderate, and light). The students were requested to record the time spent on each of the three types of activity according to their own perception of the associated exertion experienced, although the investigators suggested sports that should be placed in each category as a guide.
For food intake, subjects were asked whether or not they had consumed any food from a list of food categories (covering the major food categories in the Japanese diet) in the 24 hours before the survey.
Blood samples were collected by each school and sent to laboratories in each prefecture. Plasma TC was determined by enzymatic methods, and HDL-C was analyzed with the prevailing methodology of the particular regional laboratory.
Height was measured with subjects barefoot, and weight was measured with subjects clothed only in underwear. BMI (kg/m2) was calculated.
Japanese Physical Education and School Health Centre Dietary Survey, 1991/1992
This survey involved 1923 schoolchildren from 58 schools throughout Japan. Schools were randomly selected from within each committed area, with the number selected being proportional to the total number of schools in the area. From each selected school, one sixth grade class (students 11 to 12 years old) was selected from the elementary schools and one second grade class (students 13 to 14 years old) from the junior high schools.
Overall, the response rate was 79.3%, with rates of 77.7% (n=793) and 81.2% (n=732) for the elementary schoolchildren and the junior high school students, respectively.
The survey was conducted during the period from October to December 1991.
Each child completed a 24-hour dietary record for 2 consecutive days, Friday and Saturday. The record books were distributed to the participants within each school, who completed the survey with the help of their parents.
US Data Collection: The Bogalusa Study
The seventh cross-sectional survey of the Bogalusa Heart Study was conducted throughout the 1992-93 and 1993-94 school years. The eligible population included all school-aged children living in Bogalusa, La, a biracial (65% white, 35% black), semirural community. Overall participation was ≈80%, a total of 2530 children 8 to 18 years old.
Food intake data (24-hour dietary recall) were collected for 284 subjects 10 years old during 1987-88, representing 90% of the 1977-1978 birth cohort. The mean age of this sample was 10.6 years, with 50% girls, 50% boys, 42% black, and 58% white.
Study Measurements: General Examination
Trained examiners (nurses and technicians) collected data according to specific written protocols.25 Participants were instructed to fast for 12 hours before examination, and compliance was determined from a direct interview.
Height was measured to the nearest 0.1 cm with a manual height board. Weight was measured to the nearest 0.1 kg with a balance-beam scale. For each height and weight, replicate measurements were obtained by the same examiner. An observation consists of the average of two readings.
Serum Lipids and Lipoproteins
Plasma TC level was determined with an Abbott VP System (Abbott Laboratories) by enzymatic procedures according to the manufacturer’s protocol. Serum levels of VLDL-C, LDL-C, and HDL-C were analyzed by a combination of heparin-calcium precipitation and agar–agarose gel electrophoresis procedures.26 The Core Lipid Laboratory was standardized by the Centers for Disease Control and Prevention in Atlanta, Ga, and monitored by a surveillance program. A 10% random sample of the study population was selected for blind, duplicate determinations to assess reproducibility of the laboratory analyses.27 Coefficients of variation were 4.3% for HDL-C and 2.0% for plasma TC.
Dietary Intake Measurement Methodology
The 24-hour dietary-recall method was adapted for use in interviewing children.28 29 30 The method included a standardized protocol and used graduated food models31 and a product identification notebook. The extended table of nutrient values was used to estimate nutrient composition.32
The numbers of subjects in each of the age and sex groups in the principal surveys in each of the three countries are provided in Table 2⇓.
One-way ANOVA was used to test the equality of means of independent samples of Australian, US, and Japanese children. Separate analyses were undertaken for each age group (8 to 10 years, 10 to 12 years, and 12 to 15 years). The effect of age was assessed in two-way ANOVA restricted to the nonoverlapping age groups (8 to 10 years and 12 to 15 years). The χ2 test was used to assess the difference of proportions in the independent samples.
The mean concentrations of HDL-C for children in each of the three studies are displayed in Fig 1⇓.
Confining attention to the nonoverlapping age groups, concentrations of HDL-C were higher for Japanese than for US or Australian children (P<.001). Australian girls had higher HDL-C than US girls (P<.001). The difference across increasing age groups in HDL-C was much greater (P<.001) for boys (Australia, −15.2%; US, −9.1%; Japan, −4.2%) than for girls (Australia, −2.6%; US, −2.7%; Japan, −1.9%). In addition, there was an intercountry difference in the extent to which HDL-C differed between the ages of 8 to 10 years and 12 to 15 years: the difference was much greater for Australian and US boys (P<.001) than for Japanese boys.
Within the Bogalusa sample, the mean plasma HDL-C was greater for black children than for white children (0.19 mmol/L for girls, 0.20 mmol/L for boys). The difference in HDL-C between the ages of 8 to 10 and 12 to 15 years was greater in white boys (−9.4%) than black boys (−7.6%).
The mean concentration of plasma TC and the ratio of mean HDL-C to mean TC for the children of each of the three studies are shown in Table 3⇓.
The plasma TC of children in the studies was higher for 8- to 10-year-olds than for 12- to 15-year-olds (P<.001), but that of Japanese children was no higher than that of Australian and US children of each sex. Between the two ages, the difference was greater for US and Australian children than for Japanese children (P<.001). The ratio of HDL-C to mean TC was higher for Japanese children than for Australian or US children at every age. For boys, the size of the difference increased with age.
HDL-C in adults and children is influenced principally by levels of body fat,20 physical activity,33 and hormone status.34 Information was available for this comparison on height and weight and physical activity.
The height of Japanese children was similar to that of Australian and American children until age 15, at which time the Australians and Americans were taller. BMI was calculated from the height and weight data. These results are displayed in Fig 2⇓.
BMI was very similar for Japanese and Australian children of both sexes within each age group, with higher levels in older age groups. In each age group, the mean BMI levels for blacks and whites in the Bogalusa sample were similar to each other and higher than Japanese and Australian children for both boys and girls.
Information about physical activity patterns was available only for Australia and Japan. The total levels of activity are compared in Fig 3⇓.
The Japanese children at all ages and in both sexes were considerably more active than their Australian counterparts. Although some data were available from both countries on whether recorded activity was moderate or vigorous, the questions were not sufficiently comparable to justify presentation of a direct comparison between the samples. However, examination of the data for the three categories of activity (vigorous, moderate, and light) did show higher activity for Japanese in all categories.
In adults and young children, modification of dietary fat intake leads to parallel changes in HDL-C and non–HDL-C.35 36 With the exception of alcohol intake, no food has been shown to increase HDL-C while leaving non–HDL-C unaffected. However, the effects of many specific individual foods (rather than specific nutrients) on a variety of health outcomes are yet to be tested, and it seems plausible that foods other than those containing high concentrations of alcohol might affect HDL-C independently of non–HDL-C. With this in mind, we examined sample differences in food and nutrient intake.
The mean total daily intakes for energy, protein, fat, and carbohydrate for Japanese and Australian subjects and those from Bogalusa are presented in Table 4⇓. The Japanese data are from the Japanese Dietary Survey Report (1991-92).
Less total fat was consumed by Japanese children. Japanese children in both age groups consumed 27% of their energy from total fat, compared with 37% for both age groups of Australian children and 36% for children in the younger US sample. Data on fatty acid composition were not available for comparison. Japanese children consumed more total energy than Australian children, consistent with their higher level of activity.
Data on consumption of specific food categories were available for the Japanese and the Australian samples. The proportions of subjects who consumed at least one food in each of the generic groups in the two locations are presented in Table 5⇓. The results were very similar for boys and girls, so combined data only are presented.
The percentages consuming some food types, eg, cereals, vegetables, and milk or milk products, were similar. Several foods, such as eggs, soybean products, mushrooms, and fish, were consumed much more frequently by Japanese than Australian children. These differences were present in both the younger and older age groups.
The finding of a higher HDL-C concentration in Japanese children than in US or Australian children is not surprising, given the lower CHD mortality in the Japanese population. However, it has not previously been reported, nor has the striking difference in the extent to which HDL-C falls in boys after puberty between Japanese and US or Australian boys, despite plasma TC concentration remaining at a similar level in all three populations. This observation is in contrast to that obtained by Knuiman et al,35 37 who compared 7- to 8-year-olds in 16 countries and found higher HDL-C concentrations in those locations in which plasma TC was high. The ratio of mean HDL-C to mean TC for Japanese children 8 to 10 years old (0.35 to 0.36) was generally higher than for the Australian or North American children and much greater than the range reported for Philippine and Pakistani boys 7 to 8 years old (0.24 to 0.27).35
Selection bias is unlikely to explain the differences, because the samples were selected from defined populations and response rates were relatively high in each of the three locations. Although there were differences in methodology for determination of HDL-C between countries and within regions of Japan, these are unlikely to account for all the variation in HDL-C observed between the locations. In particular, it is unlikely that differences in the method of precipitation of non-HDL particles before HDL-C was measured could account for the negative difference in HDL-C observed after puberty in young men in the Australian and American samples and the failure to see such a trend in Japan. In support of this conclusion, a separate Japanese study38 of 2315 students 5 to 17 years old who were residents of Tokyo, which used a single method for HDL-C analysis (precipitation using phosphotungstate and manganese chloride), also failed to find a noticeable negative difference in HDL-C with age among teenage boys. In that study, plasma TC levels in boys were 4.2, 4.0, and 3.9 mmol/L at ages 9, 12, and 15 years, respectively. For HDL-C, the respective concentrations were 1.6, 1.5, and 1.5 mmol/L.
The lesser difference in HDL-C between the ages of 12 and 15 years in Japan might be explained by a later onset of puberty in Japan. However, other data from the same 1993 Japanese survey refute this proposition. The mean HDL-C even at age 18 years for Japanese male subjects was 1.43 mmol/L—much higher than the values at 15 years for subjects from Bogalusa or Australia.
One possibility is that these differences are genetically based. We are unable to address this directly with the data available for this study. The fact that white and black male adolescents in the Bogalusa data on which this study was based had a different mean HDL-C concentration overall, for example, supports this possibility. However, an argument against a genetic explanation for our principal finding here is that in the Bogalusa data, black male adolescents showed a difference in HDL-C (−7.6%) between ages 10 and 15 years similar to that of white male adolescents (−9.4%).
The limited data available in the samples on possible determinants of HDL-C suggest that a higher level of physical activity in Japanese children may be a contributor. Unfortunately, these data were not collected by the same methodology in the two countries. However, the estimates of mean total energy intake in age-matched Australian and Japanese subjects support the conclusion that Japanese children have a higher level of physical activity. Nonetheless, an objective validation of the questionnaire information would be desirable.
In children, the strongest determinant of HDL-C has generally been reported to be obesity.39 In this comparison, however, there were not great differences in BMI between Australian and Japanese children, although the US means at each age were higher.
The dietary differences between the samples were much greater. Although diet has not been shown to be a major determinant of HDL-C in adults or children within countries, others have speculated that it may be responsible for between-country variation.35 37 The differences noted in the macronutrient intake estimates between the three countries could conceivably reflect differences in dietary methodologies and the databases used for nutrient analysis. This seems unlikely, however, if we consider the magnitude of the difference observed. The general pattern of composition of Japanese and US diets is similar to that reported in the Seven Countries Study.40 The current thinking among researchers is that the 24-hour dietary recall method is acceptable for use to characterize intakes of groups of individuals.41
Lower total fat intake in the diet is usually associated with lower serum TC and lower HDL-C.35 The fact that in this study it was not could reflect genetic differences in the response to diet, as alluded to previously, or it may be due to other dietary influences, such as the effect of individual foods or food components, or alternatively to nondietary influences such as physical activity.
The dietary data available to us do not allow a comparison between countries of the amount of different foods consumed. Comparison of the proportion of the sample consuming food from a particular category may be somewhat misleading because there are variations in the frequency and amounts of individual foods consumed within a category between countries (eg, the cereal staple in Japan is rice, and in Australia it is wheat), despite the similar percentage of each population eating some foods from each category. For some food categories, however, there is a large contrast between countries that deserves further study. For example, in Japan the consumption of soybean products, including tofu, may be important. Tofu has been shown to contain phytoestrogens, which in the amounts consumed by the Japanese children might be responsible for their higher HDL-C.42
It is conceivable that the national differences observed among children in HDL-C concentration accounts for an important fraction of the CHD risk difference between Japan and the United States and Australia. However, it is not possible to quantify that fraction. To do this would require dose-response estimates for the effect of lipoprotein concentrations from a model that included terms for either duration or age at initial exposure. Such models have been developed for the prediction of risk for CHD due to serum TC for middle-aged adults43 but not for younger age groups. The finding of such a substantial difference in HDL-C in childhood between these three nations with such discrepant CHD rates in adulthood emphasizes the need to further explore the contribution of exposure in childhood to the adult incidence of CHD. Foremost among the hypotheses to be tested is that childhood differences in food intake and physical activity may importantly affect later risk of CHD death via a pathway that involves adolescent HDL-C concentration.
Selected Abbreviations and Acronyms
|BMI||=||body mass index|
|CHD||=||coronary heart disease|
We thank Professor Iwane (now deceased) particularly for his contributions to the conception of this work and the obtaining of the Japanese data. We also wish to acknowledge the work of the data collection team in Bogalusa and both the dietary and fitness survey groups involved with assembling the Australian data.
- Received December 31, 1996.
- Revision received May 16, 1997.
- Accepted June 6, 1997.
- Copyright © 1997 by American Heart Association
National Heart Foundation. Heart and Stroke Facts: A Report by the National Heart Foundation of Australia. 1995.
Hatano S. Changing CHD mortality and its causes in Japan during 1955-1985. Int J Epidemiol. 1989;18(suppl 1):149-158.
Okayama A, Ueshima H, Marmot MG, Nakamura M, Kita Y, Yamakawa M. Changes in total serum cholesterol and other risk factors for cardiovascular disease in Japan, 1980-1989. Int J Epidemiol.. 1993;22:1038-1047.
Ueshima H, Tatara K, Asakura S. Declining mortality from ischemic heart disease and changes in coronary risk factors in Japan, 1956-1980. Am J Epidemiol.. 1987;125:62-72.
The Economic Planning Agency, Japan. White Paper on Economics: Gross National Products. Appendix p. 65. 1994.
Hughes K. Trends in mortality from ischaemic heart disease in Singapore, 1959 to 1983. Int J Epidemiol.. 1986;15:44-50.
Shekelle RB, Stamler J. Fish and coronary heart disease: the epidemiologic evidence. Nutr Metab Cardiovasc Dis.. 1993;3:46-51.
Australian Health and Fitness Survey 1985. The Fitness, Health and Physical Performance of Australian School Students Aged 7-15 Years. ACHPER Publications. 1985.
Department of Community Services and Health. National Dietary Survey of Schoolchildren (Aged 10-15 years): 1985. No 2 Nutrients Consumed. Canberra, Australia: Australian Government Publication Service; 1985.
Japanese Society of School Health. Report on a Survey of Health Status of Schoolchildren (in Japanese). 1993.
The Japanese Physical Education and School Health Centre. Promotion of School Feeding through Cooperation of School, Family and Community: Report on Dietary Lifestyle Survey (in Japanese). 1994.
Berenson GS, ed. Causation of Cardiovascular Risk Factors in Children: Perspectives on Cardiovascular Risk in Early Life. New York, NY: Raven Press; 1986.
Nicklas T, Webber LS, Srinivasan SR, Berenson GS. Secular trends in dietary intakes and cardiovascular risk factors of 10-y-old children: the Bogalusa Heart Study (1973-1988). Am J Clin Nutr.. 1993;57:930-937.
Dwyer T, Gibbons LE. The Australian Schools Health and Fitness Survey: physical fitness related to blood pressure but not lipoproteins. Circulation.. 1994;89:1539-1544.
Lipid Research Clinics Program. Manual of Laboratory Operations: Lipids and Lipoprotein Analysis. Bethesda, Md: National Institutes of Health; 1974. US Dept of Health, Education, and Welfare publication NIH 75:628.
Wornick DF, Albers JJ. A comprehensive evaluation of the heparin-manganese precipitation procedure for estimating high density lipoprotein cholesterol. J Lipid Res.. 1978;19:65-76.
Paul AA, Southgate DAT. McCance and Widdowson’s The Composition of Foods. 4th ed. London, UK: Her Majesty’s Stationery Office; 1988.
Berenson GS, McMahan CA, Voors AW, Webber LS, Srinivasan SR, Frank GC, Foster TA, Blonde CV. Cardiovascular Risk Factors in Children: The Early Natural History of Atherosclerosis and Essential Hypertension. New York, NY: Oxford University Press Inc; 1980.
Srinavasan SR. Serum lipoproteins in children and methods for study. In: Lewis LA, ed. CRC Handbook of Electrophoresis, Vol 3: Lipoprotein Methodology and Human Studies. Boca Raton, Fla: CRC Press; 1983:185-204.
Farris RP, Nicklas TA. Characterizing children’s eating behaviour. In: Suskind RM, Suskind LL, eds. Textbook of Pediatric Nutrition. New York, NY: Raven Press; 1993:505-516.
Bogalusa Heart Study Nutrition Staff. In-House Dietary Studies Methodology. Eds 1-7. New Orleans, La: Louisiana State University Medical Center Specialized Center of Research, Arteriosclerosis; 1984-1990.
Moore MC, Goodloe MH. Extended Table⇑ of Nutrient Values (ENTV). New Orleans, La: Louisiana State University Medical Center; 1982:1-2.
Srinivasan S. Biologic determinants (or correlates) of serum lipoproteins in children. In: Berenson G, ed. Causation of Cardiovascular Risk Factors in Children: Perspectives on Cardiovascular Risk in Early Life. New York, NY: Raven Press; 1986:82-130.
Knuiman JT, West CE, Katan MB, Hautvast JGAJ. Total cholesterol and high density lipoprotein cholesterol levels in populations differing in fat and carbohydrate intake. Arteriosclerosis.. 1987;7:612-619.
Niinikoski H, Viikari J, Ronnemaa T, Lapinleimu H, Jokinen E, Salo P, Seppanen R, Leino A, Tuominen J, Valimaki I, Simell O. Prospective randomized trial of low-saturated-fat, low-cholesterol diet during the first 3 years of life. Circulation.. 1996;94:1386-1393.
Murata M. Hyperlipidemia in children: the current status and significant influences. Lipid.. 1993;4:416-423 (in Japanese).
Keys A. Seven Countries: A Multivariate Analysis of Death and Coronary Heart Disease. Cambridge, Mass: Harvard University Press; 1980.
Block G. A review of validations of dietary assessment methods. Am J Epidemiol.. 1982;115:492-505.
Law MR, Wald NJ, Thompson SG. By how much and how quickly does reduction in serum cholesterol concentration lower risk of ischaemic heart disease? BMJ.. 1994;308:367-373.