0021-972X/91/7304-0907$03.00/0 Journal of Clinical Endocrinology and Metabolism Copyright (p*) 1991 by The Endocrine Society
Vol. 73, No. 4 Printed in U.S.A.
Insulin-Like Growth Factor-I as a Reflection of Body Composition, Nutrition, and Puberty in Sixth and Seventh Grade Girls* DARRELL M. WILSON, JOEL D. KILLEN, LAWRENCE D. HAMMER, IRIS F. LITT, CARRIE VOSTI, BEVERLY MINER, CHRIS HAYWARD, AND C. BARR TAYLOR Departments of Pediatrics (D.M.W., L.D.H., I.F.L.), Medicine (J.D.K., C.V., B.M.), and Psychiatry (C.B.T., C.H.), Stanford University School of Medicine, Stanford, California 94305
ABSTRACT. Large variations in nutritional intake have profound effects on the GH-insulin-like growth factor-I (IGF-I) axis in children and adults, but the effect of normal variations in nutrition on IGF-I concentrations is largely unstudied, particularly during puberty. We measured serum IGF-I concentrations in 325 sixth and seventh grade girls (12.4 ± 0.7 yr) at the beginning of a multisite school-based health curriculum. The mean serum IGF-I level among the 243 girls with complete data was 573 ± 244 fig/L. Pubertal stage was significantly associated with IGF-I (P < 0.0001, by analysis of variance). Mean concentrations rose from 427 ± 198 Mg/L among those at the earliest pubertal stages to 639 ±219 jig/L among the mature girls. After adjusting for the association with the stage of pubertal
N
development, serum IGF-I was not significantly associated with measures of body composition (body mass index, triceps skin fold thickness, waist/hip ratio, height, and weight). Additionally, IGF-I concentrations were not associated with nutritional intake (total calories, total protein, total fat, and total carbohydrate) or such measures of nutrition as serum iron, hemoglobin, red cell mean corpuscular volume, white cell count, and cholesterol. IGF-I concentrations, however, were significantly correlated with transferrin concentrations, another possible index of nutritional status (r = 0.29; P < 0.0001). IGF-I is not a clinically useful index of nutritional status among normal pubertal girls. (J Clin Endocrinol Metab 73: 907-912,1991)
matched controls. These data imply that serum IGF-I reflects nutritional status, and some have suggested that serum IGF-I be used to monitor nutritional intervention in acutely ill adults (20, 21). There are few data, however, regarding any association between IGF-I concentrations and nutrition among normal adolescents. The major question addressed in this report is whether IGF-I concentrations could be used as an index of nutritional status among normal pubertal girls.
UTRITION has profound effects on the GH-insulin-like growth factor-I (IGF-I) axis (1). Studies of severe malnutrition in children demonstrated low IGFI concentrations in these subjects (2-4). Children with chronic inflammatory bowel disease have low IGF-I concentrations which rise as nutritional intake improves (5). Similar decreases in IGF-I have been detected in fasting normal and obese adults (6, 7). As fasting subjects are refed, IGF-I levels rise toward normal in proportion to the adequacy of the caloric replacement (8, 9). Decreased IGF-I levels also occur in females with anorexia nervosa (10-12). In contrast, moderate caloric restriction of obese adults does not result in a significant decrease in serum IGF-I concentrations (13, 14). The effect of obesity on IGF-I concentrations is much less clear, with different investigators reporting increased (15, 16), normal (17), and even low (18, 19) levels compared to those in ageReceived December 21, 1990. Address all correspondence and requests for reprints to: Dr. Darrell M. Wilson, Department of Pediatrics, Endocrine Division, Room S322, Stanford Medical Center, Stanford, California 94305. * This work was supported by Research Grants HD-24240 and HD24779 from the NIH. These data were submitted in part at the Second International Symposium on Insulin-Like Growth Factors/Somatomedins, Jan 12-16,1991, San Francisco, CA.
Materials and Methods Overview The data presented in this report were collected before the implementation of a multisite school-based health curriculum to promote healthy weight regulation. Subjects All sixth and seventh grade girls from 4 schools in Northern California were invited to participate. Of 971 eligible students, 939 girls, 10.7-14.8 yr of age, agreed to participate in the project, and 325 (35%) agreed to have their blood drawn. Only the 243 subjects who had complete baseline data (IGF-I concentrations, 907
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
WILSON ET AL.
908
height, weight, pubertal stage, and nutritional assessment) were used in subsequent data analysis (Table 1). Ethnic origin was determined by self-identification (Table 2); groups representing 5% or less of the total study group were combined as "other." Pubertal self-staging Subjects were instructed to select from line drawings (with written descriptions) representing the five Tanner stages of breast development (frontal and lateral) and the five Tanner stages of pubic hair development (22) that most accurately reflected their own stage of pubertal development. Self-evaluations correlate closely with those of physicians (23, 24). Pubic hair and breast stages were closely correlated (Spearman r = 0.62; P < 0.0001), and the median (round to the next integer) of the two stages, defined as the sexual maturity index (SMI) (25), was used for data analysis. Since only 6 subjects reported no pubertal development, SMI stages 1 and 2 were combined (Table 3).
JCE&M-1991 Vol 73 • No 4
ton, IL) with the subjects wearing light indoor clothing without shoes or coats. The body mass index (BMI), a common measure of adiposity, was calculated as weight (kilograms) divided by height (meters) squared. The waist and hip circumferences were measured in light indoor clothing using calibrated metal tapes and standard techniques. Triceps skin-fold thickness was measured to the nearest 0.2 mm according to established guidelines (26). Dietary assessment A semiquantitative food frequency assessment (27) was used to estimate dietary intake. While each instrument was reviewed to ensure completion and accuracy, preliminary analysis revealed sharply skewed distribution, with some subjects making clearly erroneous responses. To minimize the effects of these outlying subjects, nutritional data were trimmed, such that responses below the 10th percentile were set at the 10th percentile, while those above the 90th percentile were set at the 90th.
Physical measurements Standing height was measured twice to the nearest millimeter using a portable direct reading Harpenden stadiometer and standard techniques. Weight was determined to the nearest 0.1 kg with a calibrated electronic scale (Scale-tronix, Whea-
Assays Blood samples were obtained between 0900-1300 h by trained personnel using standard venipuncture techniques and sterile disposable equipment. Cholesterol, iron, transferrin, he-
TABLE 1. Characteristics of the study population Total study group (n = 939)
Grade placement 6th grade 7th grade Missing
Data analysis group (n = 243)
Difference
No.
%
538 375 26
150 93
28 25
NSa
102 249 361 114 113 12.4 ± 0.7 3.9 ± 1.9 20.3 ± 3.8 n = 675 3040 ± 1930 109 ± 70 104 ± 67 421 ± 272
34 66 115 28
33 27 32 25
NSa
SMI
Stages 1 and 2 Stage 3 Stage 4 Stage 5 Missing Age (yr) Parental education (arbitrary units)c BMI (kg/m2) Nutrition intake Total calories (kCal/day) Total protein (g/day) Total fat (g/day) Total carbohydrate (g/day) Biochemical measures IGF-I fczg/L) Cholesterol (mmol/L) Iron (jtmol/L) Transferrin (g/L) Hemoglobin (g/L) Mean corpuscular vol (nm3) White blood cell count
12.3 ± 0.7 4.0 ± 1.8 20.4 ± 4.1 n = 243 2863 ± 1775 102 ± 64 97 ± 6 1 402 ± 253 573 ± 4.30 ± 15.5 ± 3.44 ± 134 ± 83 ± 6.2 ±
0.026 NS* NS* NS" NS" NS" NS6
244 0.63 5.5 0.51 7.7 4.1 1.4
Values are the mean ± SD. Significance was determined between those bled vs. not bled. 0 By x2, excluding missing categories. 6 By Wilcoxon rank sum test. c Six-point scale based on the number of years of education of the most educated parent.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
IGF-I AND NUTRITION IN PUBERTAL GIRLS TABLE 2. Ethnic distribution of the study population Total study Ethnic origin n Caucasian African American Asian Hispanic Pacific Islander Native American Other Missing/unknown
392
Total 0
Analysis
Data analysis group
gruup % of total
n
909
% of analysis group
34 175 239 49 25 13 12
42 4 19 25 5 3 1 1
125 7° 41 48 13° 2° 4° 3°
51 3 17 20 5 1 2 1
939
100
243
100
Combined for subsequent data analysis.
The data were analyzed using the Statistical Analysis System for the PC (SAS Institute, Cary, NC). The two-tailed Wilcoxon rank sum test was used to compare groups on scaled responses, and the x2 test was used for categorical responses. The Spearman rank correlation method was used. While none of the variables except the waist/hip ratio was significantly associated with ethnicity [analysis of variance (ANOVA) within each of the four SMI groups], many of the variables were associated with SMI (by ANOVA; see Table 4). A z-score (SD score, a value of zero represents the mean for a given pubertal stage) was calculated for each variable to reduce the confounding effects of pubertal development (see Table 4). Since IGF-I concentrations have a log-normal distribution, the z-score for IGF-I was calculated from the log of the IGF-I concentration.
TABLE 3. IGF-I concentrations by SMI and ethnicity
4. Correlations between log-normalized IGF z-scores and body composition and nutritional variable z-scores (first column), and correlations between raw IGF-I concentrations and body composition and nutritional variables (second column)
TABLE
SMI stage Ethnicity Asians Mean SD
No. Hispanics Mean SD
No.
3
4
465 177 9
508 209 15
700 352 17
292 152
519 211 12
618 171 24
695 214 9
642 238
620
221
449 212
20
30
59
16
3
Caucasians Mean SD
No. Other Mean SD
No. Total Mean SD
2
430
435
606
630
42 2
247 9
260 15
427 119 34
496 215 66
644 247 115
5
z-Scores" for each variable
BMP
237
Skin fold thickness6 Waist/hip ratio6 Ht6 Wt6
0.10 (0.13) 0.02
Raw data BMI Skin fold thickness
(0.75) -0.01
(0.84) 0.11 (0.10) 0.11
Waist/hip ratio Ht Wt
(0.08)
639 147 3
Total calories
639 219
Total fat
28 No. Pubertal development (SMI) was significantly associated with IGFI concentrations, while ethnicity was not (by ANOVA; see Results).
IGF-I z-scores
Total protein
Total carbohydrate Transferrin6
moglobin, red cell mean corpuscular volume, and white cell count were determined by standard clinical laboratory methods (Allied Laboratories, Cupertino, CA).
Iron
IGF-I
Red cell MCV6
IGF-binding proteins were removed by acid chromatography before RIA (28) using IGF-I (Amgen, Inc., Thousand Oaks, CA) and the anti-IGF-I antibody UBK487, a gift of Drs. L. Underwood and J. J. Van Wyk, University of North Carolina, distributed by the National Hormone and Pituitary Program (29).
White cell count
Hemoglobin
0.04 (0.52) 0.05 (0.43) 0.08 (0.22) 0.02 (0.75) 0.23 (0.0005) 0.01 (0.85)
Total calories
0.11 (0.08) 0.00
Hemoglobin
Total protein Total fat Total carbohydrate Transferrin Iron
Red cell MCV
(0.93)
Consent Signed consent from a parent and assent from each student were obtained before venipuncture. All studies were approved by the Stanford University Medical Committee for the Protection of Human Subjects in Research.
Cholesterol
0.01 (0.90)
-0.05 (0.4)
White cell count Cholesterol
IGF-I 0.25 (0.0001) 0.11
(0.09) -0.07 (0.27) 0.28 (0.0001) 0.29 (0.0001)
0.10 (0.14) 0.09 (0.13) 0.13 (0.04) 0.06 (0.33) 0.29 (0.0001) 0.01 (0.92) 0.18 (0.005) 0.07 (0.26) 0.02 (0.75) -0.06 (0.32)
The use of the z-scores removes the confounding effects of pubertal stage. Note that a z-score of zero represents the mean value for a given pubertal stage. Spearman rho values are given, with P values in parentheses. MCV, Mean corpuscular volume. 0 z-Scores are normalized for the stage of pubertal development (SMI; see Materials and Methods). 6 Implies that the variable was associated with SMI (P < 0.05, by ANOVA).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
WILSON ET AL.
910
JCE & M • 1991 Vol73«No4
male students with complete data was 573 ± 244 Pubertal stage was significantly associated with IGF-I concentration (P < 0.0001, by ANOVA). IGF-I concentrations rose from 427 ± 198 ng/L among those with early or no pubertal development to 639 ± 2 1 9 fig/L among the fully mature girls (Fig. 1). When SMI was included in the model, neither age nor self-reported ethnicity was associated with IGF-I concentrations (P = 0.46 and 0.61, respectively, by ANOVA; Table 3). Socioeconomic status was not associated with IGF-I concentrations (P = 0.13, by ANOVA).
1,000 800
§ 600 LL
(3 400 200
Effect of body composition on IGF-I concentrations 1 &2
3
4
SMI STAGE FlG. 1. IGF-I concentrations through pubertal development among 243 sixth and seventh grade girls. The enclosed box extends from the 25th to the 75th percentile; the horizonal line within the box represents the 50th percentile. Extremes of the lines above and below each box indicate the 90th and 10th percentiles, respectively. Normal adult concentrations of IGF-I range from 135-450 fig/L.
z rr
DC LU LJ_ CO
After controlling for the association between pubertal development and IGF-I concentrations, ANOVA revealed that none of the various measures of body composition (BMI, triceps skin fold thickness, waist/hip ratio, height, or weight) was significantly associated with IGF-I concentration. As expected, height, weight, and BMI, which increase as the SMI increases, were correlated with absolute IGF-I concentrations (Table 4). When the effect of SMI was removed for calculating zscores (based on SMI) for IGF-I and the other variables, no independent associations between these variables and IGF-I z-scores could be detected. The mean IGF-I z-score in girls with BMI z-scores below the 10th percentile was similar to that in girls with BMI z-scores above the 90th percentile (-0.12 ± 0.73 vs. 0.08 ± 1.13, respectively; P > 0.48). Effect of nutrition on IGF-I concentrations
600
800
1000
1200
1400
IGF-I (UG/L) FIG. 2. IGF-I concentrations us. transferrin concentrations. There a small but significant correlation between these two variables (r = 0.29; P = 0.0005; Spearman rank correlation). P < 0.05 was considered significant. Data are presented as the mean ± SD.
Results The general characteristics of the total study population and of those subjects who completed all phases required for subsequent data analysis were quite similar (Table 1). The ethnic composition of the total study population and that of subjects who completed all phases required for subsequent data analysis were significantly different. The mean concentration of IGF-I among the 243 fe-
After controlling for the association between pubertal development and IGF-I concentrations, ANOVA revealed that none of the other measures, i.e. nutrition, total calories, total protein, total fat, total carbohydrate, serum iron, hemoglobin, red cell mean corpuscular volume, or cholesterol, was significantly associated with the IGF-I concentration. Moreover, none of these measures was significantly correlated with the IGF-I z-score. Serum transferrin concentrations, however, were significantly associated with IGF-I concentrations (P < 0.0002, by ANOVA, included SMI and transferrin). Additionally, transferrin and IGF-I were positively correlated with each other (Fig. 2). When data analysis was restricted to only those 115 students at the modal stage of puberty (SMI = 4), IGFI was significantly correlated only with transferrin (r = 0.22; P = 0.018) and with no other variables associated with either body composition or nutrition.
Discussion The association we found in these subjects between IGF-I concentrations and puberty development agrees
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
IGF-I AND NUTRITION IN PUBERTAL GIRLS with previous reports (30, 31). As sex steroid concentrations rise during puberty, spontaneous GH secretion increases (32). This increase in GH is reflected by the pubertal rise in IGF-I concentrations. It is clear that the degree of pubertal development is of greater importance than age when examining IGF-I concentrations during adolescence. The collection of nutritional intake data is problematic during childhood and adolescence. However, despite a broad range of reported caloric intake we found no association between IGF-I concentrations and multiple measures of nutrition in these normal adolescent females. Since IGF-I concentrations are clearly decreased in severe malnutrition and fasting (2-9), there is a threshold below which inadequate nutrition begins to alter IGF-I concentrations. Among obese adults, restriction of daily caloric intakes as low as 24 Cal/kg ideal BW for 20 days (13) or 18 Cal/kg ideal BW for 90 days (14) does not alter IGF-I concentrations despite significant weight loss. Moreover, adolescents with cystic fibrosis have normal IGF-I levels despite a high incidence of essential fatty acid deficiency (33). Our data suggest that the broad range of dietary intakes found in normal girls during puberty is sufficient to support normal serum IGF-I. We were able to obtain accurate anthropometric measurements on these subjects. After controlling for the obvious confounding effects of pubertal progression, IGF-I concentrations have no significant association with any of our physical measures. While other investigators have found altered IGF-I concentrations in clinical groups selected for specific diagnosis such as anorexia nervosa (10-12) or obesity (15-19), there are few data regarding the association between body composition and IGF-I concentrations, particularly in adolescence. Our data demonstrate little association between body composition or adiposity and serum IGF-I concentrations among the general population of pubertal girls. IGF-I concentrations are strongly associated with the stage of pubertal development. Changes in nutrition, in turn, can alter both the age at which puberty begins and the rate of progression through puberty. It is, thus, possible that we were unable to detect a small but significant effect of altered nutrition on IGF-I concentration if the subject's nutritional status also altered the course of pubertal development. We found a small but statistically significant association only between serum transferrin (a potential index of nutritional status) and IGF-I. It is possible that the correlation between these two potential indices of nutrition reflects variations in nutrition that were not detected by our assessment of the dietary intake. While our study could have missed subtle associations between nutrition or body composition and IGF-I con-
911
centrations, our data demonstrate that serum IGF-I is not a clinically useful index of nutritional status among normal pubertal girls. Acknowledgments The authors thank Yun Kee, M.S., and Michael Powers, B.S., for accurately performing the IGF-I assays, Helena C. Kraemer, Ph.D., and Ann Varady for their statistical advice, the teachers and staff of the Santa Clara Unified and FranklinMcKinley school districts, and the skilled nurses who obtained the blood samples.
References 1. Phillips LS. Nutrition, somatomedins, and the brain. Metabolism. 1986;35:78. 2. Hintz RL, Suskind R, Amatayakul K, Thanangkul 0, Olson R. Plasma somatomedin and growth hormone values in children with protein-calorie malnutrition. J Pediatr. 1978;92:153-6. 3. Grant DB, Hambley J, Becker D, Pimstone BL. Reduced sulfation factor in undernourished children. Arch Dis Child. 1973;48:597. 4. Smith IF, Latham MC, Azubuike JA, et al. Blood plasma levels of cortisol, insulin, growth hormone, and somatomedin in children with marasmus, kwashiorkor, and intermediate forms of proteinenergy malnutrition. Proc Soc Exp Biol Med 1981;167:607-ll. 5. Kirschner BS, Sutton MM. Somatomedin-C levels in growthimpaired children and adolescents with chronic inflammatory bowel disease. Gastroenterology. 1986;91:830-6. 6. Clemmons DR, Kibanski A, Underwood LE, et al. Reduction of plasma immunoreactive somatomedin-C during fasting in humans. J Clin Endocrinol Metab. 1981;153:1247-50. 7. Caufriez A, Golstein J, Lebrun P, Herchuelz A, Furlanetto R, Copinschi G. Relation between immunoreactive somatomedin-C, insulin and T3 patterns during fasting in obese subjects. Clin Endocrinol (Oxf). 1984;20:65-70. 8. Isley WL, Underwood LE, Clemmons DR. Dietary components that regulate serum somatomedin-C concentrations in humans. J Clin Invest. 1983;71:175-82. 9. Isley WL, Underwood LE, Clemmons DR. Changes in plasma somatomedin-C in response to ingestion of diets with variable protein and energy content. J Parent Ent Nutr. 1984;8:407-ll. 10. Rappaport R, Prevot C, Czemichow P. Somatomedin activity and growth hormone secretion. Changes related to body weight in anorexia nervosa. Acta Paediatr Scand. 1980;67:37-41. 11. Winter J, Gwirstman HE, George DT, Kaye WH, Loriaux DL, Cutler Jr GB. Adrenocorticotropin-stimulated adrenal androgen secretion in anorexia nervosa: impaired secretion at low weight with normalization after long-term weight recovery. J Clin Endocrinol Metab. 1985;61:693-7. 12. Tanaka T, Maesaka H, Suwa S. Changes in somatomedin activity in anorexia nervosa. Endocrinol Jpn. 1985;32:891-7. 13. Clemmons DR, Snyder DK, Williams R, Underwood LE. Growth hormone administration conserves lean body mass during dietary restriction in obese subjects. J Clin Endocrinol Metab. 1987;64:878-83. 14. Snyder DK, Clemmons DR, Underwood LE. Treatment of obese, diet-restricted subjects with growth hormone for 11 weeks: effects on anabolism, lipolysis, and body composition. J Clin Endocrinol Metab. 1988;67:54-61. 15. Van Vliet G, Bosson D, Rummens E, Robyn C, Wolter R. Evidence against growth hormone-releasing factor deficiency in children with idiopathic obesity. Acta Endocrinol (Copenh). 1986; 279(Suppl):403-10. 16. Loche S, Cappa M, Borrelli P, et al. Reduced growth hormone response to growth hormone-releasing hormone in children with simple obesity: evidence for somatomedin-C mediated inhibition. Clin Endocrinol (Oxf). 1987;27:145-53. 17. Caufriez A, Goldstein J, Leburn P, Herchuelz A, Furlanetto R,
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
912
18. 19.
20.
21. 22. 23. 24. 25.
WILSON ET AL. Copinschi G. Relations between immunoreactive somatomedin-C, insulin and T3 patterns during fasting in obese subjects. Clin Endocrinol (Oxf). 1984;20:65-70. Gama R, Teale JD, Marks V. The effect of synthetic very low calorie diets on the GH-IGF-I axis in obese subjects. Clin Chim Acta. l990;188:31-8. Minuto F, Barreca A, Del Monte P, et al. Spontaneous growth hormone and somatomedin-C/insulin-like growth factor secretion in obese subjects during puberty. J Endocrinol Invest. 1988;11:48995. Clemmons DR, Underwood LE, Dickerson RN, et al. Use of plasma somatomedin-C/insulin-like growth factor I measurements to monitor the response to nutritional repletion in malnourished patients. Am J Clin Nutr. 1985;41:191-8. Unterman TG, Vazquez RM, Slas AJ, Martyn PA, Phillips LS. Nutrition and somatomedin. XIII. Usefulness of somatomedin-C in nutritional assessment. Am J Med. 1985;78:228-34. Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in girls. Arch Dis Child. 1969;44:291. Duke PM, Litt IF, Gross RT. Adolescents' self-assessment of sexual maturation. Pediatrics. 1980;66:918. Frankowski F, Duke-Duncan P, Guillot A, et al. Young adolescents' self-assessment of sexual maturation [Abstract]. Am J Dis Child. 1987;141:385. Wilson DM, Kraemer HC, Ritter PL, Hammer LD. Growth curves and adult height estimation for adolescents with early and late
JCE&M«1991 Vol 73 • No 4
puberty. Am J Dis Child. 1987;141:565-70. 26. US DHEW. Skinfold thickness of youths 12-17 years. In: Vital and health statistics, ser 11, no. 32. Rockville: DHEW; 1974. 27. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a semi quantitative food frequency questionnaire. Am J Epidemiol. 1985;122:51-65. 28. Homer JM, Liu F, Hintz RL. Comparison of 125-1 somatomedin A and 125-1 somatomedin C radioreceptor assays for somatomedin peptide content in whole and acid-chromatographed plasma. J Clin Endocrinol Metab. 1978;47:1287. 29. Wilson DM, Bennett A, Adamson GD, et al. Somatomedins in pregnancy: a cross sectional study of insulin-like growth factors I and II and somatomedin peptide content in normal human pregnancies. J Clin Endocrinol Metab. 1982;56:858. 30. Luna AM, Wilson DM, Wibbelsman CJ, et al. Somatomedins in adolescence: a cross sectional study of the effect of puberty on plasma insulin-like growth factor I and II levels. J Clin Endocrinol Metab. 1987;57:268. 31. Rosenfield RI, Furlanetto R, Bock D. Relationship of somatomedin-C concentrations to pubertal changes. J Pediatr. 1983;103:723. 32. Costin G, Kaufman FR, Brasel JA. Growth hormone secretory dynamics in subjects with normal stature. J Pediatr. 1989; 115:53744. 33. Rosenfeld RG, Landon C, Lewiston N, Nagashima R, Hintz RL. Demonstration of plasma somatomedin concentrations in cystic fibrosis. J Pediatr. 1981;99:252-4.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
Vol. 73, No. 4 Printed in U.S.A.
Insulin-Like Growth Factor-I as a Reflection of Body Composition, Nutrition, and Puberty in Sixth and Seventh Grade Girls* DARRELL M. WILSON, JOEL D. KILLEN, LAWRENCE D. HAMMER, IRIS F. LITT, CARRIE VOSTI, BEVERLY MINER, CHRIS HAYWARD, AND C. BARR TAYLOR Departments of Pediatrics (D.M.W., L.D.H., I.F.L.), Medicine (J.D.K., C.V., B.M.), and Psychiatry (C.B.T., C.H.), Stanford University School of Medicine, Stanford, California 94305
ABSTRACT. Large variations in nutritional intake have profound effects on the GH-insulin-like growth factor-I (IGF-I) axis in children and adults, but the effect of normal variations in nutrition on IGF-I concentrations is largely unstudied, particularly during puberty. We measured serum IGF-I concentrations in 325 sixth and seventh grade girls (12.4 ± 0.7 yr) at the beginning of a multisite school-based health curriculum. The mean serum IGF-I level among the 243 girls with complete data was 573 ± 244 fig/L. Pubertal stage was significantly associated with IGF-I (P < 0.0001, by analysis of variance). Mean concentrations rose from 427 ± 198 Mg/L among those at the earliest pubertal stages to 639 ±219 jig/L among the mature girls. After adjusting for the association with the stage of pubertal
N
development, serum IGF-I was not significantly associated with measures of body composition (body mass index, triceps skin fold thickness, waist/hip ratio, height, and weight). Additionally, IGF-I concentrations were not associated with nutritional intake (total calories, total protein, total fat, and total carbohydrate) or such measures of nutrition as serum iron, hemoglobin, red cell mean corpuscular volume, white cell count, and cholesterol. IGF-I concentrations, however, were significantly correlated with transferrin concentrations, another possible index of nutritional status (r = 0.29; P < 0.0001). IGF-I is not a clinically useful index of nutritional status among normal pubertal girls. (J Clin Endocrinol Metab 73: 907-912,1991)
matched controls. These data imply that serum IGF-I reflects nutritional status, and some have suggested that serum IGF-I be used to monitor nutritional intervention in acutely ill adults (20, 21). There are few data, however, regarding any association between IGF-I concentrations and nutrition among normal adolescents. The major question addressed in this report is whether IGF-I concentrations could be used as an index of nutritional status among normal pubertal girls.
UTRITION has profound effects on the GH-insulin-like growth factor-I (IGF-I) axis (1). Studies of severe malnutrition in children demonstrated low IGFI concentrations in these subjects (2-4). Children with chronic inflammatory bowel disease have low IGF-I concentrations which rise as nutritional intake improves (5). Similar decreases in IGF-I have been detected in fasting normal and obese adults (6, 7). As fasting subjects are refed, IGF-I levels rise toward normal in proportion to the adequacy of the caloric replacement (8, 9). Decreased IGF-I levels also occur in females with anorexia nervosa (10-12). In contrast, moderate caloric restriction of obese adults does not result in a significant decrease in serum IGF-I concentrations (13, 14). The effect of obesity on IGF-I concentrations is much less clear, with different investigators reporting increased (15, 16), normal (17), and even low (18, 19) levels compared to those in ageReceived December 21, 1990. Address all correspondence and requests for reprints to: Dr. Darrell M. Wilson, Department of Pediatrics, Endocrine Division, Room S322, Stanford Medical Center, Stanford, California 94305. * This work was supported by Research Grants HD-24240 and HD24779 from the NIH. These data were submitted in part at the Second International Symposium on Insulin-Like Growth Factors/Somatomedins, Jan 12-16,1991, San Francisco, CA.
Materials and Methods Overview The data presented in this report were collected before the implementation of a multisite school-based health curriculum to promote healthy weight regulation. Subjects All sixth and seventh grade girls from 4 schools in Northern California were invited to participate. Of 971 eligible students, 939 girls, 10.7-14.8 yr of age, agreed to participate in the project, and 325 (35%) agreed to have their blood drawn. Only the 243 subjects who had complete baseline data (IGF-I concentrations, 907
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
WILSON ET AL.
908
height, weight, pubertal stage, and nutritional assessment) were used in subsequent data analysis (Table 1). Ethnic origin was determined by self-identification (Table 2); groups representing 5% or less of the total study group were combined as "other." Pubertal self-staging Subjects were instructed to select from line drawings (with written descriptions) representing the five Tanner stages of breast development (frontal and lateral) and the five Tanner stages of pubic hair development (22) that most accurately reflected their own stage of pubertal development. Self-evaluations correlate closely with those of physicians (23, 24). Pubic hair and breast stages were closely correlated (Spearman r = 0.62; P < 0.0001), and the median (round to the next integer) of the two stages, defined as the sexual maturity index (SMI) (25), was used for data analysis. Since only 6 subjects reported no pubertal development, SMI stages 1 and 2 were combined (Table 3).
JCE&M-1991 Vol 73 • No 4
ton, IL) with the subjects wearing light indoor clothing without shoes or coats. The body mass index (BMI), a common measure of adiposity, was calculated as weight (kilograms) divided by height (meters) squared. The waist and hip circumferences were measured in light indoor clothing using calibrated metal tapes and standard techniques. Triceps skin-fold thickness was measured to the nearest 0.2 mm according to established guidelines (26). Dietary assessment A semiquantitative food frequency assessment (27) was used to estimate dietary intake. While each instrument was reviewed to ensure completion and accuracy, preliminary analysis revealed sharply skewed distribution, with some subjects making clearly erroneous responses. To minimize the effects of these outlying subjects, nutritional data were trimmed, such that responses below the 10th percentile were set at the 10th percentile, while those above the 90th percentile were set at the 90th.
Physical measurements Standing height was measured twice to the nearest millimeter using a portable direct reading Harpenden stadiometer and standard techniques. Weight was determined to the nearest 0.1 kg with a calibrated electronic scale (Scale-tronix, Whea-
Assays Blood samples were obtained between 0900-1300 h by trained personnel using standard venipuncture techniques and sterile disposable equipment. Cholesterol, iron, transferrin, he-
TABLE 1. Characteristics of the study population Total study group (n = 939)
Grade placement 6th grade 7th grade Missing
Data analysis group (n = 243)
Difference
No.
%
538 375 26
150 93
28 25
NSa
102 249 361 114 113 12.4 ± 0.7 3.9 ± 1.9 20.3 ± 3.8 n = 675 3040 ± 1930 109 ± 70 104 ± 67 421 ± 272
34 66 115 28
33 27 32 25
NSa
SMI
Stages 1 and 2 Stage 3 Stage 4 Stage 5 Missing Age (yr) Parental education (arbitrary units)c BMI (kg/m2) Nutrition intake Total calories (kCal/day) Total protein (g/day) Total fat (g/day) Total carbohydrate (g/day) Biochemical measures IGF-I fczg/L) Cholesterol (mmol/L) Iron (jtmol/L) Transferrin (g/L) Hemoglobin (g/L) Mean corpuscular vol (nm3) White blood cell count
12.3 ± 0.7 4.0 ± 1.8 20.4 ± 4.1 n = 243 2863 ± 1775 102 ± 64 97 ± 6 1 402 ± 253 573 ± 4.30 ± 15.5 ± 3.44 ± 134 ± 83 ± 6.2 ±
0.026 NS* NS* NS" NS" NS" NS6
244 0.63 5.5 0.51 7.7 4.1 1.4
Values are the mean ± SD. Significance was determined between those bled vs. not bled. 0 By x2, excluding missing categories. 6 By Wilcoxon rank sum test. c Six-point scale based on the number of years of education of the most educated parent.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
IGF-I AND NUTRITION IN PUBERTAL GIRLS TABLE 2. Ethnic distribution of the study population Total study Ethnic origin n Caucasian African American Asian Hispanic Pacific Islander Native American Other Missing/unknown
392
Total 0
Analysis
Data analysis group
gruup % of total
n
909
% of analysis group
34 175 239 49 25 13 12
42 4 19 25 5 3 1 1
125 7° 41 48 13° 2° 4° 3°
51 3 17 20 5 1 2 1
939
100
243
100
Combined for subsequent data analysis.
The data were analyzed using the Statistical Analysis System for the PC (SAS Institute, Cary, NC). The two-tailed Wilcoxon rank sum test was used to compare groups on scaled responses, and the x2 test was used for categorical responses. The Spearman rank correlation method was used. While none of the variables except the waist/hip ratio was significantly associated with ethnicity [analysis of variance (ANOVA) within each of the four SMI groups], many of the variables were associated with SMI (by ANOVA; see Table 4). A z-score (SD score, a value of zero represents the mean for a given pubertal stage) was calculated for each variable to reduce the confounding effects of pubertal development (see Table 4). Since IGF-I concentrations have a log-normal distribution, the z-score for IGF-I was calculated from the log of the IGF-I concentration.
TABLE 3. IGF-I concentrations by SMI and ethnicity
4. Correlations between log-normalized IGF z-scores and body composition and nutritional variable z-scores (first column), and correlations between raw IGF-I concentrations and body composition and nutritional variables (second column)
TABLE
SMI stage Ethnicity Asians Mean SD
No. Hispanics Mean SD
No.
3
4
465 177 9
508 209 15
700 352 17
292 152
519 211 12
618 171 24
695 214 9
642 238
620
221
449 212
20
30
59
16
3
Caucasians Mean SD
No. Other Mean SD
No. Total Mean SD
2
430
435
606
630
42 2
247 9
260 15
427 119 34
496 215 66
644 247 115
5
z-Scores" for each variable
BMP
237
Skin fold thickness6 Waist/hip ratio6 Ht6 Wt6
0.10 (0.13) 0.02
Raw data BMI Skin fold thickness
(0.75) -0.01
(0.84) 0.11 (0.10) 0.11
Waist/hip ratio Ht Wt
(0.08)
639 147 3
Total calories
639 219
Total fat
28 No. Pubertal development (SMI) was significantly associated with IGFI concentrations, while ethnicity was not (by ANOVA; see Results).
IGF-I z-scores
Total protein
Total carbohydrate Transferrin6
moglobin, red cell mean corpuscular volume, and white cell count were determined by standard clinical laboratory methods (Allied Laboratories, Cupertino, CA).
Iron
IGF-I
Red cell MCV6
IGF-binding proteins were removed by acid chromatography before RIA (28) using IGF-I (Amgen, Inc., Thousand Oaks, CA) and the anti-IGF-I antibody UBK487, a gift of Drs. L. Underwood and J. J. Van Wyk, University of North Carolina, distributed by the National Hormone and Pituitary Program (29).
White cell count
Hemoglobin
0.04 (0.52) 0.05 (0.43) 0.08 (0.22) 0.02 (0.75) 0.23 (0.0005) 0.01 (0.85)
Total calories
0.11 (0.08) 0.00
Hemoglobin
Total protein Total fat Total carbohydrate Transferrin Iron
Red cell MCV
(0.93)
Consent Signed consent from a parent and assent from each student were obtained before venipuncture. All studies were approved by the Stanford University Medical Committee for the Protection of Human Subjects in Research.
Cholesterol
0.01 (0.90)
-0.05 (0.4)
White cell count Cholesterol
IGF-I 0.25 (0.0001) 0.11
(0.09) -0.07 (0.27) 0.28 (0.0001) 0.29 (0.0001)
0.10 (0.14) 0.09 (0.13) 0.13 (0.04) 0.06 (0.33) 0.29 (0.0001) 0.01 (0.92) 0.18 (0.005) 0.07 (0.26) 0.02 (0.75) -0.06 (0.32)
The use of the z-scores removes the confounding effects of pubertal stage. Note that a z-score of zero represents the mean value for a given pubertal stage. Spearman rho values are given, with P values in parentheses. MCV, Mean corpuscular volume. 0 z-Scores are normalized for the stage of pubertal development (SMI; see Materials and Methods). 6 Implies that the variable was associated with SMI (P < 0.05, by ANOVA).
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
WILSON ET AL.
910
JCE & M • 1991 Vol73«No4
male students with complete data was 573 ± 244 Pubertal stage was significantly associated with IGF-I concentration (P < 0.0001, by ANOVA). IGF-I concentrations rose from 427 ± 198 ng/L among those with early or no pubertal development to 639 ± 2 1 9 fig/L among the fully mature girls (Fig. 1). When SMI was included in the model, neither age nor self-reported ethnicity was associated with IGF-I concentrations (P = 0.46 and 0.61, respectively, by ANOVA; Table 3). Socioeconomic status was not associated with IGF-I concentrations (P = 0.13, by ANOVA).
1,000 800
§ 600 LL
(3 400 200
Effect of body composition on IGF-I concentrations 1 &2
3
4
SMI STAGE FlG. 1. IGF-I concentrations through pubertal development among 243 sixth and seventh grade girls. The enclosed box extends from the 25th to the 75th percentile; the horizonal line within the box represents the 50th percentile. Extremes of the lines above and below each box indicate the 90th and 10th percentiles, respectively. Normal adult concentrations of IGF-I range from 135-450 fig/L.
z rr
DC LU LJ_ CO
After controlling for the association between pubertal development and IGF-I concentrations, ANOVA revealed that none of the various measures of body composition (BMI, triceps skin fold thickness, waist/hip ratio, height, or weight) was significantly associated with IGF-I concentration. As expected, height, weight, and BMI, which increase as the SMI increases, were correlated with absolute IGF-I concentrations (Table 4). When the effect of SMI was removed for calculating zscores (based on SMI) for IGF-I and the other variables, no independent associations between these variables and IGF-I z-scores could be detected. The mean IGF-I z-score in girls with BMI z-scores below the 10th percentile was similar to that in girls with BMI z-scores above the 90th percentile (-0.12 ± 0.73 vs. 0.08 ± 1.13, respectively; P > 0.48). Effect of nutrition on IGF-I concentrations
600
800
1000
1200
1400
IGF-I (UG/L) FIG. 2. IGF-I concentrations us. transferrin concentrations. There a small but significant correlation between these two variables (r = 0.29; P = 0.0005; Spearman rank correlation). P < 0.05 was considered significant. Data are presented as the mean ± SD.
Results The general characteristics of the total study population and of those subjects who completed all phases required for subsequent data analysis were quite similar (Table 1). The ethnic composition of the total study population and that of subjects who completed all phases required for subsequent data analysis were significantly different. The mean concentration of IGF-I among the 243 fe-
After controlling for the association between pubertal development and IGF-I concentrations, ANOVA revealed that none of the other measures, i.e. nutrition, total calories, total protein, total fat, total carbohydrate, serum iron, hemoglobin, red cell mean corpuscular volume, or cholesterol, was significantly associated with the IGF-I concentration. Moreover, none of these measures was significantly correlated with the IGF-I z-score. Serum transferrin concentrations, however, were significantly associated with IGF-I concentrations (P < 0.0002, by ANOVA, included SMI and transferrin). Additionally, transferrin and IGF-I were positively correlated with each other (Fig. 2). When data analysis was restricted to only those 115 students at the modal stage of puberty (SMI = 4), IGFI was significantly correlated only with transferrin (r = 0.22; P = 0.018) and with no other variables associated with either body composition or nutrition.
Discussion The association we found in these subjects between IGF-I concentrations and puberty development agrees
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
IGF-I AND NUTRITION IN PUBERTAL GIRLS with previous reports (30, 31). As sex steroid concentrations rise during puberty, spontaneous GH secretion increases (32). This increase in GH is reflected by the pubertal rise in IGF-I concentrations. It is clear that the degree of pubertal development is of greater importance than age when examining IGF-I concentrations during adolescence. The collection of nutritional intake data is problematic during childhood and adolescence. However, despite a broad range of reported caloric intake we found no association between IGF-I concentrations and multiple measures of nutrition in these normal adolescent females. Since IGF-I concentrations are clearly decreased in severe malnutrition and fasting (2-9), there is a threshold below which inadequate nutrition begins to alter IGF-I concentrations. Among obese adults, restriction of daily caloric intakes as low as 24 Cal/kg ideal BW for 20 days (13) or 18 Cal/kg ideal BW for 90 days (14) does not alter IGF-I concentrations despite significant weight loss. Moreover, adolescents with cystic fibrosis have normal IGF-I levels despite a high incidence of essential fatty acid deficiency (33). Our data suggest that the broad range of dietary intakes found in normal girls during puberty is sufficient to support normal serum IGF-I. We were able to obtain accurate anthropometric measurements on these subjects. After controlling for the obvious confounding effects of pubertal progression, IGF-I concentrations have no significant association with any of our physical measures. While other investigators have found altered IGF-I concentrations in clinical groups selected for specific diagnosis such as anorexia nervosa (10-12) or obesity (15-19), there are few data regarding the association between body composition and IGF-I concentrations, particularly in adolescence. Our data demonstrate little association between body composition or adiposity and serum IGF-I concentrations among the general population of pubertal girls. IGF-I concentrations are strongly associated with the stage of pubertal development. Changes in nutrition, in turn, can alter both the age at which puberty begins and the rate of progression through puberty. It is, thus, possible that we were unable to detect a small but significant effect of altered nutrition on IGF-I concentration if the subject's nutritional status also altered the course of pubertal development. We found a small but statistically significant association only between serum transferrin (a potential index of nutritional status) and IGF-I. It is possible that the correlation between these two potential indices of nutrition reflects variations in nutrition that were not detected by our assessment of the dietary intake. While our study could have missed subtle associations between nutrition or body composition and IGF-I con-
911
centrations, our data demonstrate that serum IGF-I is not a clinically useful index of nutritional status among normal pubertal girls. Acknowledgments The authors thank Yun Kee, M.S., and Michael Powers, B.S., for accurately performing the IGF-I assays, Helena C. Kraemer, Ph.D., and Ann Varady for their statistical advice, the teachers and staff of the Santa Clara Unified and FranklinMcKinley school districts, and the skilled nurses who obtained the blood samples.
References 1. Phillips LS. Nutrition, somatomedins, and the brain. Metabolism. 1986;35:78. 2. Hintz RL, Suskind R, Amatayakul K, Thanangkul 0, Olson R. Plasma somatomedin and growth hormone values in children with protein-calorie malnutrition. J Pediatr. 1978;92:153-6. 3. Grant DB, Hambley J, Becker D, Pimstone BL. Reduced sulfation factor in undernourished children. Arch Dis Child. 1973;48:597. 4. Smith IF, Latham MC, Azubuike JA, et al. Blood plasma levels of cortisol, insulin, growth hormone, and somatomedin in children with marasmus, kwashiorkor, and intermediate forms of proteinenergy malnutrition. Proc Soc Exp Biol Med 1981;167:607-ll. 5. Kirschner BS, Sutton MM. Somatomedin-C levels in growthimpaired children and adolescents with chronic inflammatory bowel disease. Gastroenterology. 1986;91:830-6. 6. Clemmons DR, Kibanski A, Underwood LE, et al. Reduction of plasma immunoreactive somatomedin-C during fasting in humans. J Clin Endocrinol Metab. 1981;153:1247-50. 7. Caufriez A, Golstein J, Lebrun P, Herchuelz A, Furlanetto R, Copinschi G. Relation between immunoreactive somatomedin-C, insulin and T3 patterns during fasting in obese subjects. Clin Endocrinol (Oxf). 1984;20:65-70. 8. Isley WL, Underwood LE, Clemmons DR. Dietary components that regulate serum somatomedin-C concentrations in humans. J Clin Invest. 1983;71:175-82. 9. Isley WL, Underwood LE, Clemmons DR. Changes in plasma somatomedin-C in response to ingestion of diets with variable protein and energy content. J Parent Ent Nutr. 1984;8:407-ll. 10. Rappaport R, Prevot C, Czemichow P. Somatomedin activity and growth hormone secretion. Changes related to body weight in anorexia nervosa. Acta Paediatr Scand. 1980;67:37-41. 11. Winter J, Gwirstman HE, George DT, Kaye WH, Loriaux DL, Cutler Jr GB. Adrenocorticotropin-stimulated adrenal androgen secretion in anorexia nervosa: impaired secretion at low weight with normalization after long-term weight recovery. J Clin Endocrinol Metab. 1985;61:693-7. 12. Tanaka T, Maesaka H, Suwa S. Changes in somatomedin activity in anorexia nervosa. Endocrinol Jpn. 1985;32:891-7. 13. Clemmons DR, Snyder DK, Williams R, Underwood LE. Growth hormone administration conserves lean body mass during dietary restriction in obese subjects. J Clin Endocrinol Metab. 1987;64:878-83. 14. Snyder DK, Clemmons DR, Underwood LE. Treatment of obese, diet-restricted subjects with growth hormone for 11 weeks: effects on anabolism, lipolysis, and body composition. J Clin Endocrinol Metab. 1988;67:54-61. 15. Van Vliet G, Bosson D, Rummens E, Robyn C, Wolter R. Evidence against growth hormone-releasing factor deficiency in children with idiopathic obesity. Acta Endocrinol (Copenh). 1986; 279(Suppl):403-10. 16. Loche S, Cappa M, Borrelli P, et al. Reduced growth hormone response to growth hormone-releasing hormone in children with simple obesity: evidence for somatomedin-C mediated inhibition. Clin Endocrinol (Oxf). 1987;27:145-53. 17. Caufriez A, Goldstein J, Leburn P, Herchuelz A, Furlanetto R,
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
912
18. 19.
20.
21. 22. 23. 24. 25.
WILSON ET AL. Copinschi G. Relations between immunoreactive somatomedin-C, insulin and T3 patterns during fasting in obese subjects. Clin Endocrinol (Oxf). 1984;20:65-70. Gama R, Teale JD, Marks V. The effect of synthetic very low calorie diets on the GH-IGF-I axis in obese subjects. Clin Chim Acta. l990;188:31-8. Minuto F, Barreca A, Del Monte P, et al. Spontaneous growth hormone and somatomedin-C/insulin-like growth factor secretion in obese subjects during puberty. J Endocrinol Invest. 1988;11:48995. Clemmons DR, Underwood LE, Dickerson RN, et al. Use of plasma somatomedin-C/insulin-like growth factor I measurements to monitor the response to nutritional repletion in malnourished patients. Am J Clin Nutr. 1985;41:191-8. Unterman TG, Vazquez RM, Slas AJ, Martyn PA, Phillips LS. Nutrition and somatomedin. XIII. Usefulness of somatomedin-C in nutritional assessment. Am J Med. 1985;78:228-34. Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in girls. Arch Dis Child. 1969;44:291. Duke PM, Litt IF, Gross RT. Adolescents' self-assessment of sexual maturation. Pediatrics. 1980;66:918. Frankowski F, Duke-Duncan P, Guillot A, et al. Young adolescents' self-assessment of sexual maturation [Abstract]. Am J Dis Child. 1987;141:385. Wilson DM, Kraemer HC, Ritter PL, Hammer LD. Growth curves and adult height estimation for adolescents with early and late
JCE&M«1991 Vol 73 • No 4
puberty. Am J Dis Child. 1987;141:565-70. 26. US DHEW. Skinfold thickness of youths 12-17 years. In: Vital and health statistics, ser 11, no. 32. Rockville: DHEW; 1974. 27. Willett WC, Sampson L, Stampfer MJ, et al. Reproducibility and validity of a semi quantitative food frequency questionnaire. Am J Epidemiol. 1985;122:51-65. 28. Homer JM, Liu F, Hintz RL. Comparison of 125-1 somatomedin A and 125-1 somatomedin C radioreceptor assays for somatomedin peptide content in whole and acid-chromatographed plasma. J Clin Endocrinol Metab. 1978;47:1287. 29. Wilson DM, Bennett A, Adamson GD, et al. Somatomedins in pregnancy: a cross sectional study of insulin-like growth factors I and II and somatomedin peptide content in normal human pregnancies. J Clin Endocrinol Metab. 1982;56:858. 30. Luna AM, Wilson DM, Wibbelsman CJ, et al. Somatomedins in adolescence: a cross sectional study of the effect of puberty on plasma insulin-like growth factor I and II levels. J Clin Endocrinol Metab. 1987;57:268. 31. Rosenfield RI, Furlanetto R, Bock D. Relationship of somatomedin-C concentrations to pubertal changes. J Pediatr. 1983;103:723. 32. Costin G, Kaufman FR, Brasel JA. Growth hormone secretory dynamics in subjects with normal stature. J Pediatr. 1989; 115:53744. 33. Rosenfeld RG, Landon C, Lewiston N, Nagashima R, Hintz RL. Demonstration of plasma somatomedin concentrations in cystic fibrosis. J Pediatr. 1981;99:252-4.
The Endocrine Society. Downloaded from press.endocrine.org by [${individualUser.displayName}] on 14 November 2015. at 01:47 For personal use only. No other uses without permission. . All rights reserved.
