Negative relationships between growth in height and levels of cholesterol in puberty: a 3-year follow-up study

Transcription

1 IJE vol.32 no.6 International Epidemiological Association 2003; all rights reserved. International Journal of Epidemiology 2003;32: DOI: /ije/dyg207 Negative relationships between growth in height and levels of cholesterol in puberty: a 3-year follow-up study Katsuyasu Kouda, 1,2 Harunobu Nakamura, 1 WenYing Fan 1 and Hiroichi Takeuchi 1 Accepted 23 April 2003 Background Previously, there were only a few reports on the negative relationship between pubertal growth in height and levels of serum lipid in boys. Detailed information on both genders is needed. Methods Results Conclusion Keywords We investigated the relationship between pubertal growth in height and serum lipid. Subjects were 1442 boys and 1350 girls followed up from age years (the fifth grade level of elementary school) to age years (the second year of junior high school). Anthropometric variables and serum lipids were measured by the same protocol at both ages. From cross-sectional analysis, at both ages negative relationships between total cholesterol levels and height were found in both genders. On longitudinal analysis, height at age years was one of the factors predicting the level of total cholesterol at age years. In addition, negative relationships between increase in height and change in serum lipids (total cholesterol and high density lipoprotein cholesterol) over the 3-year period were obtained in both genders. Thus, pubertal children who experience a large increase in height tended to show a decrease in serum lipids, and children who experience a small increase in height tended to show an increase in serum lipids. In both genders, total cholesterol level in pubertal children is negatively associated with height. Height velocity is inversely associated with dynamic changes in serum lipids during puberty. Body height, cardiovascular diseases, child development, growth, hyperlipidaemia, Japan, lipids, longitudinal studies, puberty, schools Tracking of serum lipids and lipoproteins from childhood to adulthood has been reported. 1 3 In the wake of previous findings, it is considered that serum lipid measurements in childhood and adolescence can predict adult values. 1 3 In contrast, there are some cases whose serum lipid levels are high in childhood and normal in adulthood. 4 Serum lipids levels in schoolchildren are affected by race and sex Thus, detailed analysis of serum lipids in child development is needed. 1 Department of Public Health, Hamamatsu University School of Medicine, Hamamatsu, Japan. 2 Department of Hygiene, Kansai Medical University, Moriguchi, Japan. Correspondence: Dr Katsuyasu Kouda, Department of Hygiene, Kansai Medical University, 10 15, Fumizono-cho, Moriguchi, Osaka , Japan. koudak@takii.kmu.ac.jp Total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) levels of newborns are very low compared with those of adults, and increase with growth. 5 In schoolchildren, dynamic changes in serum lipids are observed during puberty. Serum lipids values peak at 9 10 years of age, and decrease thereafter. In adolescence, serum lipids begin to increase again. These changes in serum lipids during puberty are more remarkable in males than in females The decrease in serum lipids levels during puberty are associated with the process of sexual maturation, 6,9,13,14 and related to an increase in the plasma testosterone concentration in boys or oestradiol concentration in girls. 9,15,16 Growth spurt in height, which is affected by the secretion level of growth hormone (GH), also occurs in this period. It has been reported that basal GH level is positively correlated with high-density 1105

2 1106 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY lipoprotein cholesterol (HDL-C), 17,18 and the administration of GH decreases LDL-C and increases HDL-C. 19 There have been only a few reports on the relationship between growth in height and levels of serum lipids. Freedman et al. 20 conducted a longitudinal study with 78 boys, and reported that the relationships of change in LDL-C with change in weight were positive. In contrast, the relationship of change in LDL-C with change in height was negative. Chiang et al. 21 also reported finding a negative association between change in TC and change in height in 208 white boys and 189 black boys. From these reports, it is postulated that children who demonstrate a relatively large increase in height might also show a large decrease in serum lipids during puberty. However, these previous reports, which discussed the relationship between change in serum lipids and pubertal growth in height, arose from investigations including only boys. Detailed information on both genders is needed. In this study, we conducted a 3-year longitudinal study in Japanese schoolchildren (1442 boys and 1350 girls), and examined the relationship between pubertal growth in height and levels of serum lipids in both genders. Methods Population A total of 3219 children aged years, who were in the fifth grade of elementary school located in Iwata City in Shizuoka prefecture, Japan, were initially eligible. This group included 569 males and 567 females in 1994, 552 males and 480 females in 1996, and 512 males and 539 females in We could examine 99.9% of the initially eligible children aged years, and could follow a total of 2792 children (1442 males and 1350 females) for over 3 years, until they reached the age of years, the second year of junior high school. Children who graduated or moved outside the school district, and children with the same family and given names were lost to follow-up. The retention rate was 86.7%. Concerning the pubertal development in this study population, we have reported previously that the 3.8% of the girls aged years in 1996 and 1997 experienced menarche. 22 It was reported that peak height velocity (cm/year) was shown to occur around age 9 10 years in girls and years in boys in Japanese children in Examinations The same protocol was used in all examinations of children aged years and years. At each examination, height to 0.1 cm and weight to 0.1 kg were measured. 24 Measurements of height and body weight were made by Yogo teachers who have a Japanese national educational license, and play a role in health education and healthcare in elementary and junior high schools. Body mass index (BMI, kg/m 2 ) was calculated by dividing the weight (kg) by the height (m) squared. For measurements of systolic blood pressure (SBP) and diastolic blood pressure (DBP), an automated device (BP-103N II, Colin Corporation, Komaki, Japan) was used. Serum TC and serum HDL-C were also measured, and atherogenic index [AI = (TC HDL-C)/HDL-C] was calculated at each examination. TC was determined enzymatically (Pureauto CHO-N, Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan) using a Hitachi 7350 automatic chemistry analyser. HDL-C was measured by a direct method (Cholestest HDL, Daiichi Pure Chemicals Co., Ltd., Tokyo, Japan) on the same analyser. Statistical analysis For analysis of baseline and follow-up levels of the variables measured, 95% CI was used to evaluate the difference between the result of each examination at age years and the result of each examination at age years. Pearson s correlation test was employed for expressing a 3-year tracking of serum lipids and anthropometric variables. Simple regression analysis and multiple regression analysis were employed for analysis of cross-sectional association between TC and variables at the two ages. These regression analyses were also employed for analysis of factors predicting the levels of TC at the age of years from height, weight, SBP, and DBP at age years. Simple and multiple regression analyses were also employed for analysis of relationships between changes in serum lipids and changes in anthropometric variables over a 3-year period. Pearson s correlation test was also employed for expressing correlation coefficients between changes in TC and changes in height over a 3-year period. All statistical analyses were performed using StatView (version 5.0, SAS institute Inc., Cary, NC, USA). Results Table 1 shows the longitudinal change and relationship over a 3-year follow-up period. In boys, TC level decreased from age years to years. In addition, there were correlations between baseline and follow-up variables. Table 2 shows the cross-sectional associations between TC and variables at two ages, years and years. At both ages, negative regression coefficients of height and TC were found in both boys and girls. Table 3 shows the results of regression analyses for predicting the level of TC in children at age years from height, weight, SBP, and DBP at age years. Height at years was the predicting factor of TC level at years, and regression coefficients of height were negative in both genders. Table 4 shows the results of regression analyses between change in anthropometric variables and change in serum lipids over a 3-year period. In both genders, a negative association was observed between change in height and change in TC. A negative association was also observed between change in height and change in serum HDL-C in both genders. Figure 1 shows the relationship between increase in height and change in TC. Pubertal children who experience a large increase in height tended to show a decrease in serum lipids, and children who experience a small increase in height tended to show an increase in serum lipids in both genders. Discussion In boys, TC level decreased from age years to years. It is acknowledged that serum TC and HDL-C values peak at 9 10 years old, and decrease between 10 and 14 years old in boys. 5,6,8,9,12,14,25 28 In addition, there were strong correlations between baseline and follow-up levels of anthropometric variables and serum lipids.

3 HEIGHT AND CHOLESTEROL IN PUBERTY 1107 Table 1 Longitudinal change and relationship over a 3-year follow-up period Correlation years years Difference coefficient Mean ± SE Mean ± SE Mean (95% CI) (95% CI) Height (cm) Boys ± ± (20.6, 21.0) 0.87 (0.85, 0.88) Girls ± ± (15.1, 15.6) 0.78 (0.76, 0.80) Weight (kg) Boys 33.4 ± ± (15.1, 15.6) 0.89 (0.87, 0.90) Girls 33.6 ± ± (13.4, 13.8) 0.86 (0.84, 0.87) Body mass index (kg/m 2 ) Boys 17.4 ± ± (1.7, 1.8) 0.87 (0.86, 0.88) Girls 17.3 ± ± (2.4, 2.6) 0.84 (0.83, 0.86) Total cholesterol (mmol/l) Boys 4.48 ± ± ( 0.19, 0.13) 0.73 (0.71, 0.76) Girls 4.51 ± ± (0.07, 0.13) 0.70 (0.67, 0.73) High density lipoprotein cholesterol (mmol/l) Boys 1.72 ± ± ( 0.12, 0.09) 0.71 (0.68, 0.73) Girls 1.64 ± ± (0.02, 0.05) 0.68 (0.65, 0.71) Atherogenic index Boys 1.64 ± ± (0.05, 0.09) 0.75 (0.72, 0.77) Girls 1.79 ± ± ( 0.04, 0.01) 0.74 (0.72, 0.77) Systolic blood pressure (mmhg) Boys ± ± (5.6, 6.9) 0.39 (0.35, 0.43) Girls ± ± (1.6, 2.9) 0.39 (0.34, 0.43) Diastolic blood pressure (mmhg) Boys 60.9 ± ± (0.8, 1.8) 0.27 (0.22, 0.31) Girls 62.6 ± ± (0.3, 1.3) 0.34 (0.29, 0.38) Pearson s correlation test boys and 1350 girls. Table 2 Cross-sectional associations between total cholesterol and variables at the age of years and the age of years Dependent Explanatory Regression coefficient (95% CI) Sex variable variable Unadjusted Adjusted Boys years TC a years height 0.40 ( 0.63, 0.16) 1.15 ( 1.46, 0.84) years weight 0.28 (0.07, 0.49) 0.87 (0.58, 1.15) years SBP b 0.25 (0.13, 0.36) 0.17 (0.03, 0.32) years DBP c 0.29 (0.12, 0.45) 0.15 ( 0.05, 0.35) years TC years height 0.66 ( 0.83, 0.49) 1.19 ( 1.42, 0.96) years weight 0.03 ( 0.18, 0.11) 0.54 (0.35, 0.72) years SBP 0.14 (0.02, 0.26) 0.11 ( 0.03, 0.26) years DBP 0.25 (0.09, 0.41) 0.24 (0.05, 0.43) Girls years TC years height 0.78 ( 0.99, 0.56) 1.07 ( 1.38, 0.76) years weight 0.38 ( 0.60, 0.16) 0.33 (0.02, 0.65) years SBP 0.07 ( 0.05, 0.19) 0.09 ( 0.07, 0.25) years DBP 0.15 ( 0.03, 0.32) 0.13 ( 0.09, 0.35) years TC years height 0.38 ( 0.66, 0.09) 0.62 ( 0.94, 0.29) years weight 0.11 ( 0.09, 0.32) 0.28 (0.05, 0.52) years SBP 0.30 (0.17, 0.44) 0.22 (0.05, 0.39) years DBP 0.36 (0.17, 0.55) 0.17 ( 0.07, 0.42) Unadjusted, univariate model for each variable, simple regression analysis; Adjusted, simultaneously adjusted for all variables in the Table, multiple regression analysis boys and 1350 girls. a Total cholesterol. b Systolic blood pressure. c Diastolic blood pressure. Previously, Orchard et al. 29 and Chu et al. 30 investigated children aged years in cross-sectional studies and reported finding a negative correlation between serum lipid and height in boys. However, they did not demonstrate significant relationships in girls. In contrast, in cross-sectional studies, we demonstrated negative associations between height and TC in not only boys but also girls. In Japanese children, in 1997, the peak height velocity (cm/year) was shown to occur at age 9 10 years in girls and years in boys. 23 Present subjects aged years and years were around the peak height velocity. In contrast, subjects in previous reports were older than subjects in the present study. This might be one of the reasons why previous studies could not demonstrate the relationship in girls.

4 1108 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY Table 3 Factors predicting the levels of total cholesterol at the age of years Regression coefficient (95% CI) Sex Explanatory variable Unadjusted Adjusted Boys years height 0.56 ( 0.79, 0.33) 1.17 ( 1.48, 0.86) years weight 0.09 ( 0.12, 0.31) 0.75 (0.46, 1.03) years SBP a 0.13 (0.01, 0.25) 0.14 ( 0.00, 0.29) years DBP b 0.09 ( 0.08, 0.25) 0.00 ( 0.21, 0.20) Girls years height 0.23 ( 0.46, 0.00) 0.51 ( 0.85, 0.17) years weight 0.02 ( 0.22, 0.25) 0.33 ( 0.01, 0.68) years SBP 0.10 ( 0.03, 0.23) 0.12 ( 0.05, 0.30) years DBP 0.07 ( 0.11, 0.26) 0.03 ( 0.27, 0.21) Unadjusted, univariate model for each variable, simple regression analysis; Adjusted, simultaneously adjusted for all variables in the Table, multiple regression analysis boys and 1350 girls. a Systolic blood pressure. b Diastolic blood pressure. Table 4 Analysis of relationship between change in anthropometric variables and change in serum lipids over a 3-year period Dependent Explanatory Regression coefficient (95% CI) Sex Variable variable Unadjusted Adjusted Boys Change in TC a Change in height 1.18 ( 1.42, 0.93) 1.66 ( 1.96, 1.35) Change in weight 0.14 ( 0.34, 0.06) 0.64 (0.40, 0.88) Change in HDL-C b Change in height 0.37 ( 0.50, 0.23) 0.19 ( 0.36, 0.02) Change in weight 0.33 ( 0.44, 0.22) 0.24 ( 0.38, 0.10) Change in AI c Change in height 0.01 ( 0.01, 0.00) 0.02 ( 0.03, 0.02) Change in weight 0.02 (0.01, 0.02) 0.03 (0.02, 0.03) Girls Change in TC Change in height 1.11 ( 1.39, 0.84) 1.33 ( 1.62, 1.04) Change in weight 0.18 ( 0.12, 0.48) 0.67 (0.36, 0.98) Change in HDL-C Change in height 0.42 ( 0.56, 0.29) 0.33 ( 0.48, 0.19) Change in weight 0.40 ( 0.54, 0.25) 0.28 ( 0.43, 0.12) Change in AI Change in height 0.00 ( 0.00, 0.01) 0.01 ( 0.01, 0.00) Change in weight 0.02 (0.02, 0.03) 0.03 (0.02, 0.03) Unadjusted, univariate model for each variable, simple regression analysis; Adjusted, simultaneously adjusted for all variables in the Table, multiple regression analysis boys and 1350 girls. a Total cholesterol. b High-density lipoprotein cholesterol. c Atherogenic index. Looking at predictive factors for TC level in children at age years, height at age years was one of these. In addition, the regression coefficients were negative in both genders. Further, we investigated the changes in anthropometric variables and serum lipids over a 3-year period, and demonstrated that the relationship of change in TC with change in weight was positive, and that the relationship of change in TC with change in height was negative. These results were consistent with previous reports. Freedman et al. 20 conducted a longitudinal study in 78 white boys from age years to years in the Bogalusa Heart Study, and reported finding an inverse correlation of changes in serum lipids with increase in height. Chiang et al. 21 also followed 208 white boys and 189 black boys from age 8 12 years to years in the Bogalusa Heart Study, and reported finding a negative correlation between changes in serum lipids and increase in height. However, few previous publications have reported any such an association in girls. In the present result, negative relationships between changes in height and changes in serum lipids were shown in both genders. Thus, pubertal children who experience a large increase in height tended to show a decrease in serum lipids, and children who experience a small increase in height tended to show an increase in serum lipids. In both genders, TC level in pubertal children is negatively associated with height. Height velocity is inversely associated with dynamic changes in serum lipids during puberty. Acknowledgements This work was supported in part by a grant-in-aid Research on the establishment of a healthy lifestyle from childhood from the Ministry of Health and Welfare, Japan. The authors thank the school administration in Iwata City for their help. KEY MESSAGES Cholesterol levels in pubertal children are affected by their development. A negative relationship between height and total cholesterol is shown around growth spurt in height. Height velocity is inversely associated with dynamic changes in serum lipids during puberty.

5 HEIGHT AND CHOLESTEROL IN PUBERTY 1109 Figure 1 Relationship between change in total cholesterol (TC) and increase in height. Open square and dotted line, boys (r = 0.24, P 0.001); closed triangle and straight line, girls (r = 0.21, P 0.001). Data are from 1442 boys and 1350 girls, and are presented as Pearson s correlation test References 1 Porkka KVK, Viikari JSA, Åkerblom HK. Tracking of serum HDLcholesterol and other lipids in children and adolescents: the Cardiovascular Risk in Young Finns Study. Prev Med 1991;20: 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: Porkka KVK, Viikari JSA, Taimela S, Dahl M, Åkerblom HK. Tracking and predictiveness of serum lipid and lipoprotein measurements in childhood: a 12-year follow-up. The Cardiovascular Risk in Young Finns study. Am J Epidemiol 1994;140: Lauer RM, Clarke WR. Use of cholesterol measurements in childhood for the prediction of adult hypercholesterolemia. The Muscatine Study. JAMA 1990;264: Farinaro E, Giumetti D, Coraggio S, Panico S, Mancini M. Serum lipid and lipoprotein profiles in childhood. Prev Med 1983;12: Berenson GS, Srinivasan SR, Cresanta JL, Foster TA, Webber LS. Dynamic changes of serum lipoproteins in children during adolescence and sexual maturation. Am J Epidemiol 1981;113: Freedman DS, Bowman BA, Srinivasan SR, Berenson GS, Otvos JD. Distribution and correlates of high-density lipoprotein subclasses among children and adolescents. Metabolism 2001;50: Hickman TB, Briefel RR, Carroll MD et al. Distributions and trends of serum lipid levels among United States children and adolescents ages 4 19 years: data from the Third National Health and Nutrition Examination Survey. Prev Med 1998;27: Stožicky F, Slaby P, Volenikova L. Longitudinal study of serum cholesterol, apolipoproteins and sex hormones during puberty. Acta Paediatr Scand 1991;80: Srinivasan SR, Wattigney W, Webber LS, Berenson GS. Race and gender differences in serum lipoproteins of children, adolescents, and young adults emergence of an adverse lipoprotein pattern in white males: the Bogalusa Heart Study. Prev Med 1991;20: Rifkind BM, Tamir I, Heiss G, Wallace RB, Tyroler HA. Distribution of high density and other lipoproteins in selected LRC prevalence study populations: a brief survey. Lipids 1979;14: Labarthe DR, Nichaman MZ, Harrist RB, Grunbaum JA, Dai S. Development of cardiovascular risk factors from ages 8 to 18 in Project HeartBeat! Study design and patterns of change in plasma total cholesterol concentration. Circulation 1997;95: Morrison JA, Laskarzewski PM, Rauh JL et al. Lipids, lipoproteins, and sexual maturation during adolescence: the Princeton Maturation Study. Metabolism 1979;28: Tell GS, Mittelmark MB, Vellar OD. Cholesterol, high density lipoprotein cholesterol and triglycerides during puberty: the Oslo Youth Study. Am J Epidemiol 1985;122: Laskarzewski PM, Morrison JA, Gutai J, Khoury PR, Glueck CJ. Longitudinal relationships among endogenous testosterone, estradiol,

6 1110 INTERNATIONAL JOURNAL OF EPIDEMIOLOGY and Quetelet index with high and low density lipoprotein cholesterols in adolescent boys. Pediatr Res 1983;17: Kirkland RT, Keenan BS, Probstfield JL et al. Decrease in plasma highdensity lipoprotein cholesterol levels at puberty in boys with delayed adolescence. Correlation with plasma testosterone levels. JAMA 1987; 257: Luoma PV, Myllylä VV, Sotaniemi EA, Hokkanen TEJ. Plasma HDL cholesterol and growth hormone in epileptics treated with anticonvulsants. Acta Pharmacol Toxicol (Copenh) 1980;47: Bänsch D, Dirkes-Kersting A, Schulte H, Assmann G, von Eckardstein A. Basal growth hormone levels are positively correlated with highdensity lipoprotein cholesterol levels in women. Metabolism 1997; 46: I Allemand D, Eiholzer U, Schlumpf M, Steinert H, Riesen W. Cardiovascular risk factors improve during 3 years of growth hormone therapy in Prader-Willi syndrome. Eur J Pediatr 2000;159: Freedman DS, Cresanta JL, Srinivasan SR, Webber LS, Berenson GS. Longitudinal serum lipoprotein changes in white males during adolescence: the Bogalusa Heart Study. Metabolism 1985;34: Chiang YK, Srinivasan SR, Webber LS, Berenson GS. Relationship between change in height and changes in serum lipid and lipoprotein levels in adolescent males: the Bogalusa Heart Study. J Clin Epidemiol 1989;42: Kouda K, Nakamura H, Miyahara T et al. Comparative study of hypercholesterolemia and obesity; daily physical activity and dietary behavior of years old children. J Child Health 1998;57: [In Japanese.] High cholesterol is an important risk factor for coronary heart disease (CHD). It is involved in the development of atherosclerotic changes of the arterial wall, which have been shown to begin at an early age, even in apparently healthy children and adolescents. 1 Levels of cholesterol and other cardiovascular risk factors in childhood are also known to track into adult life and influence subsequent subclinical and clinical cardiovascular disease. 2 Studies of predictors of childhood cholesterol are therefore of importance for primary prevention and may help to identify those particularly susceptible to later hyperlipidaemia, atherosclerosis, and cardiovascular disease at an early age. A paper in 23 Yoshizaki M, Sasaki O, Sato T et al. KOKUMIN EISEI NO DOKO. J Health Welfare Stat 1998;45(9 Suppl.):472. [In Japanese.] 24 Foster TA, Voors AW, Webber LS, Frerichs RR, Berenson GS. Anthropometric and maturation measurements of children, ages 5 to 14 years, in a biracial community the Bogalusa Heart Study. Am J Clin Nutr 1977;30: Yamamoto A, Horibe H, Sawada S et al. Serum lipid levels in elementary and junior high school children and their relationship to relative weight. Prev Med 1988;17: Fukushige J, Igarashi H, Ueda K, Akazawa K, Nose Y. Serum cholesterol levels in school-aged Japanese children: the Hisayama study. Acta Paediatr Jpn 1996;38: Okuni M, Hayashi K, Kiryu S, Yamauchi K. Risk factors of arteriosclerosis in Japanese children. Jpn Circ J 1980;44: Zahavi I, Goldbourt U, Cohen-Mandelzweig L et al. Distributions of total cholesterol, triglycerides, and high-density lipoprotein cholesterol in Israeli Jewish children of different geographic-ethnic origins, ages 9 17 years. Prev Med 1987;16: Orchard TJ, Rodgers M, Hedley AJ, Mitchell JRA. Changes in blood lipids and blood pressure during adolescence. BMJ 1980;28: Chu NF, Rimm EB, Wang DJ, Liou HS, Shieh SM. Relationship between anthropometric variables and lipid levels among school children: the Taipei Children Heart Study. Int J Obes Relat Metab Disord 1998;22: IJE vol.32 no.6 International Epidemiological Association 2003; all rights reserved. International Journal of Epidemiology 2003;32: DOI: /ije/dyg315 Commentary: The association between height growth and cholesterol levels during puberty: implications for adult health Rebecca Hardy and Claudia Langenberg MRC National Survey of Health and Development, Department of Epidemiology and Public Health, Royal Free & University College London Medical School, 1 19 Torrington Place, London WC1E 6BT, UK. rebecca.hardy@ucl.ac.uk this issue of the International Journal of Epidemiology reports negative associations between height and total cholesterol in Japanese boys and girls around the time of puberty. 3 As well as cross-sectional associations the authors find that changes in height were negatively associated with concurrent changes in total cholesterol. An important question is what are the implications of these findings for later health? Short stature in adulthood has been consistently linked with increased CHD morbidity and mortality of CHD and also with unfavourable levels of CHD risk factors, including cholesterol. 4 Adult height is determined by genetic factors as well as early environmental influences on growth. Which particular period of growth may be crucial for the initiation and progression of cardiovascular disease remains to be determined. Few studies have considered the association between childhood growth and later cholesterol levels. A previous study in