- Email: [email protected]
Puberty decreases insulin sensitivity
Puberty decreases insulin sensitivity Puberty is commonly associated with an increase in insulin requirement In patients with insulin-dependent diabetes. To investigate whether this pubertal Increase in insulin requirement is confined to diabetic subjects, we examined insulin responses during oral glucose tolerance testing with glucose loads per unit weight (1.75 g / k g ) or unit surface area (55 g/m2), and insulin sensitivlty via euglycemlc-hyperinsulinemic clamp in prepubertal and pubertal children without diabetes. Irrespective of glucose dose, glucose tolerance testing elicited a threefold greater insulin response, but equivalent euglycemia, in pubertal versus prepubertal children (P <0.05). As assessed by the clamp procedure, prepubertal children were approximately 30% more sensitive than their pubertal counterparts (P <0.04). Insulin sensitlvity correlated Inversely with body mass index (r = -0.49, P <0.02), serum dehydroepiandrosterone sulphate concentration (r = - 0 . 5 7 , P <0.01), and log somatomedin C/insulinlike growth factor I (r = -0.45, P <0.05). We conclude that puberty is associated with decreased sensitivity to insulin that normally is compensated for by increased insulin secretion. Thus, in patients with insulin-dependent diabetes, an approximately 30% Increase in insulin dosage should be anticipated with the onset of puberty. (J PEDIATR1987;110:484-7) Clifford A. Bloch, F.C.P.(Paed) S.A., Peter Clemons, M.D., a n d Mark A. Sperling, M.D. From the Division of Endocrinology, Department of Pediatrics, University of Cincinnati
Exogenous insulin requirements tend to increase during pubertal maturation in children with insulin-dependent diabetes. 1"~This increase in insulin requirement in diabetic adolescents may reflect poor compliance, abnormal eating behavior that disrupts metabolic control) increased insulin requirement for pubertal growth, or a physiologic change in insulin sensitivity. A real change in insulin sensitivity has been invoked because in normal children undergoing oral glucose tolerance testing, ~glucose concentrations remain equivalent, but insulin concentrations are markedly higher in adolescents compared with preadolescent children. 5 This evidence for evolving insulin resistance was inferential, however, because the insulin response tO Supported by research grants HD 12613 and RR 00123 from the National Institutes of Health. Submitted for publication Aug. 25, 1986; accepted Oct. 22, 1986. Reprint requests: Mark A. Sperling, M.D., Division of Endocrinology, Children's Hospital Research Foundation, Cincinnati, OH 45229.
OGTT may depend in part on the dose of administered glucose, which is commonly calculated on the basis of unit mass. 6 Thus the larger glucose load given to older children might impose a need for more insulin to maintain euglycemia, notwithstanding the recent evidence that there is a real decline in insulin sensitivity associated with puberty. 7 The purpose of this study was twofold. First, we evaluDHEA-S IGF-I OGTT Sm-C
Dehydroepiandrosterone sulphate Insulinlike growth factor I Oral glucose tolerance test Somatomedin C
ated insulin responses during OGTT using two doses of glucose, one calculated conventionally per unit mass a n d the other per unit surface area. The latter has a more constant relationship to various metabolic processes than does weightS; dosage per unit surface area is advocated in pediatric therapeutics to avoid the pitfall of underdosing in younger children. 8 Second, we documented the extent of
481
48 2
Bloch, Clemons, and Sperling
The Journal of Pediatrics March 1987
T a b l e . Glucose and insulin responses to oral glucose tolerance test in prepubertal and pubertal subjects Prepubertal
Pubertal
Glucose dose
Glucose dose
1.75 g / k g (n = 9)
Fasting blood glucose (rag/alL) Peak blood glucose (mg/dL) Area under glucose curve (mg/dL X 4 hr)* Area under insulin curve (/zU/mL • 4 hr)t
82.0 •
55 g / m 2 (n = 8)
1.75 g / k g (n = 10)
55 g / m 2 (n = 9)
3.1
75.6 +
4.1
84.3 _+ 3.0
83.2 +
3.7
151.6 + 8.5
143.3 •
7.5
152.2 _+ 7.9
148.8 +
11.5
421.1 • 17.0
409.5 • 16.3
432.7 + 17.9
429.1 _+ 22.9
118.5 + 17.3~
115.4 +_ 16.5w
299.1 • 77.6:~
365.4 + 114.8w
Values represent mean + SEM. *To convert glucose values to mmol/L, multiply by 0.055. tTo convert insulin values to pmol/L, multiply by 7.175. :~w symbols compared, P <0.05.
the true change in insulin sensitivity as measured via the hyperinsulinemic-euglycemicclamp procedure9 in peripuber~al children. We also evaluated the relationship of insulin sensitivity to sex, body mass index, clinical stage of puberty, and biochemical indices of adrenarche and growth hormone secretion as reflected by, respectively, serum dehydroepiandrosterone sulphate and somatomedin C/insuliniike growth factor I levels. METHODS The subjects were 22 healthy children who are siblings of patients with insulin-dependent diabetes mellitus attending the Diabetes Clinic at our medical center. Informed consent was obtained from the parents, and all proposed procedures were explained to the children. Each participant was assessed clinically and assigned a pubertal~ rating according to the method of Marshall and Tanner.10.1~ Body mass index was calculated as the weight in kilograms divided by the square of the height in meters. 12 Body surface area was determined from the formula of Haycock et al. la The 22 children were equally divided into two groups, half prepubertal and half pubertal, with seven boys and four girls in each group. Mean (_+ SD) ages of the two groups were 8.1 + 1.7 and 12.4 _+ 1.7 years, respectively (P <0.001). There were no differences in body mass index between boys and girls within the prepubertal or pubertal groups, but the pubertal subjects had a significantly greater index (20.1 + 4.3 vs 16.1 _+ 2.3; P <0.02). Within the pubertal group, five children were ranked at Tanner stage II, five at stage III or IV, and one at stage V. Subjects were admitted to the Clinical Research Center
after an overnight fast. A 4-hour OGTT was performed with random assignment to a glucose dose calculated per unit mass (1.75 g/kg) or unit surface area (55 g/m2), with a maximum dose in each instance of 75 g. Blood was withdrawn from a peripheral vein at 0, 30, 60, 120, 180, and 240 minutes for immediate measurement of whole blood glucose and later assay of serum insulin. The subjects remained in the Clinical Research Center and were fed an appropriate diet for age. After a second overnight fast, each subject underwent a hyperinsulinemie-euglycemic clamp procedure. 9 Soluble insulin was infused into a peripheral vein of one hand at a constant rate of 40 mU/m2/min, a rate designed to raise the insulin concentration to approximately 100 # U / m L (717.5 pmol/L). Blood was sampled every 10 minutes from a dorsal hand vein in the contralateral limb, at ambient temperature, for immediate measurement of glucose and later measurement of insulin and mean DHEA-S and Sm-C/IGF-I. Five minutes after commencing the insulin infusion, an infusion of 15% dextrose in water was commenced via a Harvard infusio~ pump at a variable rate determined by an algorithm programmed into a hand-held calculator. The algorithm was modified by the substitution of surface area for body mass and an adjustment interval of 10 minutes. The blood glucose was clamped at 90 mg/dL and maintained at this steady state level for a period of 90 minutes. The insulin infusion was then discontinued and blood glucose monitored for a further 60 minutes to avoid hypoglycemia. Subjects were then discharged from the Clinical Research Center, and readmitted approximately 1 month later for the repeat OGTT with the alternate dose of glucose calculated per unit mass or
Volume 110 Number 3
Pubertal insulin resistance
483
2000--
300 --
I--q BOYS P < O . 0 1 ~
GIRLS
P
I
i BOYS
._L.._.
1500--
~ " ~ GIRLS
////~
200 2 & OD 1 0 0 0 - i < w -1G3 >, 9~
100
500 --
N ~'//z
t-
-///, ~///,
////, ,111~
t-
Prepubertal
Prepubertal
=ubertal
Fig. ,1, Insulin sensitivity index, calculated as glucose utilizafioi'i rate (mg/m 2 9 min) divided by log insulin concentration (#U/ mL), in prepubertal and pubertal children. (Mean _+ SEM.)
per surface area as indicated. This protocol had been reviewed and approved by the Institutional Review Board on Human Investigations. Biochemical measures. Whole blood glucose was immediately measured via the glucose oxidase method using a YSI glucose analyzer (Yellow Springs instrument Co., Inc., Yellow Springs, Ohio). Serum was immediately separated from the remainder of the samples and stored at - 2 0 ~ C foi" later assay of insulin, DHEA-S, and SmC / I G F - i by appropriate double-antibody radioimmunoassaysJ 4-'6 Glucose and insulin responses during OGTT were analyzed as the respective areas Under the curves, as previously described? The derived rate of glucose utilization at each 10-minute interval, corrected for unit surface area, was divided by the log of the insulin concentration to yield the "insulin Sensitivity index," because the relationship between glucose utilization and insulin concenti'ation is log linear between insulin concentrations of 20 and 200 # U / m L (143.5 to 1435 pmol/L)/7 The insulin sensitivity index for each subject is the mean of the calculated indices during the period of steady-state blood glucose. The values for S m - C / I G F - I were expressed logarithmically because a
Pubertal
Fig. 2. Dehydroepiandrosterone sulphate concentrations in prepubertal and pubertal children. (Mean _+ SEM.) To convert DHEA-S Concentration to nmol/L, multiply by 0.26.
normal frequency distribution occurs only with its log concentration.'8 Statistical analyses. Statistical analyses were performed via analysis of variance using the Student Newman-Keuls test if the F value was significant, or simple linear regression with the Pearson correlation coefficient? 9 Multivariate analysis was also used to examine the relationship between insulin sensitivity, age, body mass index, DHEAS, and Sm-C/IGF2I as independent variables. 2~Results are expressed as mean • SEM except where otherwise indicated. RESULTS Glucose and insulin responses during OGTT are indicated in the Table. The mean area under the glucose curves was similar between and within each group irrespective of the dose of glucose administered. Insulin area under the curve also bore no relationship to glucose dose; however, in each instance, the pubertal group had a significantly greater insulin response (P <0.05). Eight children (five prepubertal, three pubertal) did not return for repeat OGTT.
484
Bloch, Clemons, and Sperling
3--
2
~
The Journal of Pediatrics March 1987
p-0.oo3 .--._L_
--
5E
r---1BOYS ~
GIRLS
I-E i U_
//////~
-~ o.5 (..) u E
//////
basis of DHEA-S concentration less than or greater than 300 ng/dL (78 nmol/L), respectively,z~ The preadrenarchal group had an insulin sensitivity index of 232.8 _+ 8.2, versus 161.7 _+ 16.2 for the adrenarchal group (P <0.001). The prepubertal group had significantly lower Sm-C/ IGF-I values than those of the pubertal group (P <0.003, Fig. 3). Linear regression revealed that the insulin sensitivity index correlated inversely with body mass index (r = -0.49, P <0.02), DHEA-S (r = -0.57, P <0.01), and log Sm-C/IGF-I concentration (r = -0.45, P <0.05) (Fig. 4, A through C), but not with age (r = -0.32). None of these correlations was significant by multivariate analysis when considered as independent variables.
Y~44~ / / / / H
//////
~//////
"/////~
N
"///.~ /MH, /////~
z, ////i/-/ // /, ,
~/.~;/./.o •1//i.<
Prepubertal
HH~*
Pub(,rta]
Fig. 3. Somatomedin C/insulinlike growth factor l concentrations in prepubertal and pubertal Children, drawn to logarithmic scale. (Mean + SEM.)
With the hyperinsulinemic-euglycemicclamp (n = 22), mean steady-state blood glucose concentration was 89.7 _+ 0.3 and 89.4 _+ 0.3 mg/dL (4.93 + .02 and 4.92 _+ 0.02 mmol/L) in the prepubertal and pubertal children, respectively. The time to attain this steady-state blood glucose concentration was 91 _+ 6 and 94 _+ 10 minutes for the two groups, respectively, and the mean insulin concentrations at steady state were 90.3 _+_+4.6 and 106.1 _+7.5 ~ U / m L (647.9 _+ 33 and 761.3 _+ 53.8 pmol/L). These differences were not significant. The average glucose infusion rates to maintain the steady-state blood glucose level were 402.4 + 31.5 and 287.7 _+ 29.6 mg/min 9 m 2 in the prepubertal and pubertal children, respectively (P <0.02). The prepubertal group had a greater insulin sensitivity index (glucose infusion rate divided by log serum insulin concentration) than that in the pubertal group (P <0.01; Fig. 1). Overall, the insulin sensitivity index correlated inversely with the insulin response to OGTT (r = -0.62, P <0.01). The serum DHEA-S values in the pubertal group were, as expected, significantly higher than those in the prepubertal group (P <0.001, Fig. 2). We grouped the subjects as to biochemical preadrenarche and adrenarche on the
DISCUSSION The increasing insulin requirement of diabetic adolescents as they progress through puberty has been attributed to several factors, including psychologic adjustment problems and poor compliance, especially inasmuch as glycosylated hemoglobin values increase in these patients despite their increasing insulin dose. ~-3We have demonstrated that physiologic insulin resistance occurs peripubertally in children without diabetes. Thus, part of the increasing ir~sulin requirement of diabetic adolescents is attributable to a decline in insulin sensitivity, as recently reported. 7 We have demonstrated that in adolescents without diabetes the insulin resistance is compensated for by increased insulin secretion. The increased insulin secretion of adolescents during OGTT cannot be explained solely on the basis of administered glucose dose, as demonstrated by the comparison of insulin responses of prepubertal and pubertal children tested with glucose given per unit mass or per unit surface area. When glucose is given per unit mass, the differences in load may be considerable; at the 50th percentile for weight, a typical 14-year-old child would receive more than twice as much glucose as a typical 7-year-old.8 Hence it could be proposed that for equivalent glycemia, the greater glucose load would elicit more insulin secretion. By contrast, the difference in surface area of these representative individuals is only about 60%, with a corresponding reduction in the difference between administered glucose loads. Nevertheless, insulin responses during OGTT were greater in adolescents, irrespective of the absolute glucose load. Moreover, the insulin response to OGTT correlated inversely with insulin sensitivity, assessed via the hyperinsulinemic-euglycemic clamp procedure. Thus the higher insulin responses of adolescents during OGTT appear to reflect reduced insulin sensitivity, as previously suggested2 During the euglycemic clamp procedure, the actual
Volume 110 Number 3
mean blood glucose concentrations at steady state, time to achieve the clamp, and the steady-state insulin concentrations were equivalent in both groups, and similar to those reported in studies done at equal insulin infusion rates. 7,]7 With insulin concentrations of approximately 100 #U/mL, hepatic glucose output is negligible, so the rate of glucose infusion necessary to maintain euglycemia is a measure of glucose utilization, and hence sensitivity to the ambient insulin concentration.17 This relationship is log-linear between insulin concentrations of 20 and 200 p U / m L (143.5 to 1435 pmol/L). J7 By these measures, insulin sensitivity was reduced by approximately 30% in adolescents. This extent of insulin resistance is similar to that recently reported in a group of children without diabetes. 7 Because of the log-linear relationship that exists between insulin concentration and glucose utilization, a substantial increase in serum insulin concentration would be expected to result in only a modest increase in glucose utilization. This would explain the threefold increase in the insulin area under the curve observed in the adolescents during OGTT, but only a 30% decrease in the insulin sensitivity index. However, diabetic adolescents may have an additional element of insulin resistance. Insulin resistance is a prominent feature of insulin-dependent diabetes, 22 and recent data indicate that diabetic adolescents have a further reduction of insulin sensitivity compared with that of the pubertal nondiabetic children we report. 7,23 The reduced insulin sensitivity of pubertal children could result from a number of causes. First, body mass increases at about the time of puberty, as demonstrated in our study. Obesity causes insulin resistance at the insulin receptor and postreceptor sites24; however, our subjects were not obese) 1 Linear regression demonstrated a significant inverse correlation between body mass index and insulin sensitivity. Although an increase in body mass index may partly account for the peripubertal decline in insulin sensitivity, it fails to exp!ain the apparent sex difference in insulin sensitivity observed in prepubertal children, in whom there was no difference of body mass index. Second, there was an inverse relationship between insulin sensitivity and levels of DHEA-S, the predominant adrenal hormone of adrenarche, 2~ and preadrenarchal children were more insulin sensitive than adrenarchal children. These findings are intriguing in that they suggest that adrenarche, rather than puberty, is an earlier predictor of insulin resistance. Several syndromes of hyperandrogenemia are associated with insulin resistance.~5 Thus adrenarche, which occurs earlier in girls than in boys,21and the resultant rise of adrenal androgens may be a factor responsible for the decline in insulin sensitivity. Preliminary in vivo and in vitro evidence suggests that androgens
Pubertal insulin resistance
485
28 9
==
0 Q3
iy=.O.22x+23.SSI
9
~
24
. "
20
16
12
2000
l o
I
I lO0
I
l 200
I
I 300
I
-B 9
1500
E
[ I
I~=-o.49 I~
9
9
9
1000
y=-4.g4x + 1584.4 r =-0.57 D~O.01
%
<
o
500
I 0
I 100
1
1 200
1
I 300
I
C _~ E
9
2
~
9
[ Y=-O'OO2x+0'411
9
9
I r=-o,4s
I
1 ,~
..
0.5
E
O3
0.1
I 0
I
I 100
I
I 200
1
I 300
Insulin Sensitivity Index Fig. 4. Simplelinear regression of insulin sensitivity index versus body mass index (A), DHEA-S (B), and Sm-C/IGF-I (C).
may alter carbohydrate metabolism by reducing insulin sensitivity?6. 27 Third, growth hormone secretion increases in the peripubertal period? 8 Thus, increased growth hormone secretion, a known cause of insulin resistance, 29 may also be a factor responsible for the peripubertal decline in insulin sensitivity. Although we did not measure growth hormone secretion directly, we did measure Sm-C/IGF-I, which is a marker of growth hormone secretory status. Unlike growth hormone, Sm-C/IFG-I concentration in serum remains stable throughout the day because of its relatively long
486
Bloch, Clemons, and Sperling
half-life. TM S m - C / I G F - I is a valid measure of growth hormone secretion in well-nourished, healthy children, and when plotted as its logarithmic concentration, there is a normal frequency distribution. 18 Although S m - C / I G F - I is an indirect measure of growth hormone action, it does not mediate insulin resistance, which is attributed to a direct effect of growth hormone independent of S m - C / I G F - I 5 9 Our results indicate that the log S m - C / I G F - I concentration correlates inversely with insulin sensitivity. These data would be consistent with reports that correlate S m - C / IGF-I, and by implication growth hormone, with pubertal status.30,3L They are also consistent with the recent report in which integrated growth hormone concentrations were measured and found to correlate inversely with insulin sensitivity. 7 Although we did demonstrate multiple inverse correlations with insulin sensitivity via univariate analyses for each of the listed factors, multivariate analysis failed to demonstrate any single factor that independently correlated with the insulin sensitivity index. This indicates that the loss of insulin sensitivity during the peripubertal period i s multifactorial. It is possible that the three factors we have identified--increasing adrenal androgens, body mass index, and growth hormone as reflected by S m - C / I G F I - - i n t e r a c t to mediate insulin resistance. Thus, adrenal androgens increase the body mass index and augment growth hormone secretion. 32 The physiologic decrease of insulin sensitivity at the time of adrenarche/puberty implies that the clinician managing type I insulin-dependent diabetes mellitus in children should anticipate an increase in insulin requirement at the time of adrenarche/ puberty. We thank the nursing staff of the Clinical Research Center; Steven D. Chernausek, M.D., Division of Endocrinology, Department of Pediatrics, University of Cincinnati, for assay of SmC/1GF~I; and Jane Khoury, B.Sc., Department of Biostatistics, University of Cincinnati, for statistical analysis of the data. REFERENCES
1. Mann NP, Johnston DI. Improvement in metabolic control in diabetic adolescents by the use of increased insulin dose. Diabetes Care 1984;7:460-4. 2. Blethen SL, Sargeant DT, Whitlow MG, Santiago JV. Effect of pubertal stage and recent blood glucose control on plasma somatomedin C in children with insulin-dependent diabetes mellitus. Diabetes 1981;30:868-72. 3. Daneman D, Wolfson DH, Becker D J, Drash AL. Factors affecting glycosylated hemoglobin values in children with insulin-dependent diabetes. J PEDIATR 1981;99:847-53. 4. Wing RR, Nowalk MP, Marcus MD, Koeske R, Finegold D. Subclinical eating disorders and glycemic control in adolescents with type I diabetes. Diabetes Care 1986;9:162-7. 5. Rosenbloom AL, Wheeler L, Bianchi R, Chin FT, Tiwary CM, Grgic A. Age-adjusted analysis of insulin responses
The Journal of Pediatrics March 1987
6.
7.
8. 9.
10. 11. 12. 13.
14.
15.
16.
17.
18.
19. 20. 21.
22.
23.
24.
25.
during normal and abnormal glucose tolerance tests in children and adolescents. Diabetes 1975;24:820-8. National Diabetes Data Group: Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28:1039-57. Amiel SA, Sherwin RS, Simonson DC, Lauritano AA, Tamborlane WV. Impaired insulin action in puberty. N Engl J Med I986;315:215-9. Shirkey HC, Dosage (posology). In: Shirkey HC, ed. Pediatric therapy, 5th ed. St. Louis: CV Mosby, 1975;19-33. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol 1979;237:E214-23. Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in girls. Arch Dis Child 1969;44:291-303. Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in boys. Arch Dis Child 1970;45:13-23. James WPT. Treatment of obesity: the constraints for success. Clin Endocrinol Metab 1984;13:635-59. Haycock GB, Schwartz G J, Wisotsky DH. Geometric method for measuring body surface area: a height-weight formula validated in infants, children, and adults. J PEDIATR 1978; 93:62-6. Sperling MA, DeLamater PV, Phelps D, Fiser RH, Oh W, Fisher DA. Spontaneous and amino acid-stimulated glucagon secretion in the immediate postnatal period: relation to glucose and insulin. J Clin Invest 1974;53:1159-66. Korth-Schultz S, Levine LS, NewMI. Dehydroepiandrosterone sulphate (DS) levels: a rapid test for abnormal adrenal androgen secretion. J Clin Endocrinol Metab 1976;42:t00513. Furlanetto RW, Underwood LE, Van Wyk J J, D'Ercole AJ. Estimation of somatomedin-C levels in normals and patients with pituitary disease by radioimmunoassay. J Clin Invest 1977;60:648-57. Rizza RA, Mandarino L J, Gerich JE. Dose-response characteristics for effects of insulin on production and utilization of glucose in man. Am J Physiol 1981;240:E630-39. Clemmons DR, Van Wyk JJ. Somatomedin: physiological control and effects on cell proliferation. In: Baserga R, ed. Handbook of experimental pharmacology. New York: Springer-Verlag, 1981;57:t61-208. Winer BJ. Statistical principles in experimental design. New York: McGraw-Hill, 1971. Morrison DF. Multivariate statistical methods, 2nd ed. New York: McGraw-Hill, 1970;159-206. Pang S. Premature adrenarche. In: New MI, Levine LS, eds. Adrenal diseases in childhood. Pediatric Adolescent Endocrinology. New York: Karger, 1984;13:173-84. DeFronzo RA, Hendler R, Simonson D. Insulin resistance is a prominent feature of insulin-dependent diabetes. Diabetes 1982;31:795-801. Landt KW, Campaigne BN, James FW, Sperling MA. Effects of exercise training on insulin sensitivity in adolescents with type I diabetes. Diabetes Care 1985;8:461-5. Kolterman O, Insel J, Saekow M, Olefsky JM. Mechanisms of insulin resistance in human obesity. J Clin Invest 1980;65:1272-84. Richards GE, Cavallo A, Meyer W J III, et al. Obesity, acanthosis nigricans, insulin resistance, and hyperandrogenemia: pediatric perspective and natural history. J PEDIATR 1985;107:893-7.
Volume 110 Number 3
Pubertal insulin resistance
26. Peiris A, Smith GA, Kissebah AH. Relationship of sex hormones to splanchnic insulin metabolism and peripheral insulin sensitivity. Endocrinology 1986;118(Suppl):309A. 27. Mueller PL, Kwan AH. Dehydroepiandrosterone inhibits insulin-stimulated glycogen synthesis in minimal deviation hepatoma cells. Endocrinology 1986;118(Suppl):1026A. 28. Zadik Z, Chalew SA, McCarter R J, Meistas M, Kowarski AA. The influence of age on the twenty-four-hour integrated concentration of growth hormone in normal individuals. J Clin Endocrinol Metab 1985;60:513-6. 29. Rizza RA, Mandarino L J, Gerich JE. Effects of growth hormone on insulin action in man. Diabetes 1982;31:663-9.
487
30. Rosenfield RI, Furlanetto R, Bock D. Relationship of somatomedin-C concentrations to pubertal changes. J PEDIATR 1983;103:723-8. 31. Luna AM, Wilson DM, Wibbelsman C J, 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 1983;57:268-71. 32. Link K, Blizzard RM, Evans WS, Kaiser DL, Parker MW, Rogol AD. The effect of androgens on the pulsatile release and the twenty-four-hour mean concentration of growth hormone in peripubertal males. J Clin Endocrinol Metab 1986;62:159-64.
FELLOWSHIPS Available fellowships in pediatric subspeciaIties and those for general academic pediatric training are listed once a year, in May, in The JourrLal of Pediatrics. Each October, forms for listing such fellowships are sent to the Chairman of the' Department of Pediatrics at most major hospitals in the United States and Canada. Should you desire to list fellowships, a separate application must be made each year for each position. All applications must be returned to The C. V. Mosby Company by February 15 of the listing year to ensure publication. Additional forms will be supplied on request from the Journal Editing Department, The C. V. Mosby Company, 11830 Westline Industrial Drive, St. Louis, M O 63146/314-872-8370.
Exogenous insulin requirements tend to increase during pubertal maturation in children with insulin-dependent diabetes. 1"~This increase in insulin requirement in diabetic adolescents may reflect poor compliance, abnormal eating behavior that disrupts metabolic control) increased insulin requirement for pubertal growth, or a physiologic change in insulin sensitivity. A real change in insulin sensitivity has been invoked because in normal children undergoing oral glucose tolerance testing, ~glucose concentrations remain equivalent, but insulin concentrations are markedly higher in adolescents compared with preadolescent children. 5 This evidence for evolving insulin resistance was inferential, however, because the insulin response tO Supported by research grants HD 12613 and RR 00123 from the National Institutes of Health. Submitted for publication Aug. 25, 1986; accepted Oct. 22, 1986. Reprint requests: Mark A. Sperling, M.D., Division of Endocrinology, Children's Hospital Research Foundation, Cincinnati, OH 45229.
OGTT may depend in part on the dose of administered glucose, which is commonly calculated on the basis of unit mass. 6 Thus the larger glucose load given to older children might impose a need for more insulin to maintain euglycemia, notwithstanding the recent evidence that there is a real decline in insulin sensitivity associated with puberty. 7 The purpose of this study was twofold. First, we evaluDHEA-S IGF-I OGTT Sm-C
Dehydroepiandrosterone sulphate Insulinlike growth factor I Oral glucose tolerance test Somatomedin C
ated insulin responses during OGTT using two doses of glucose, one calculated conventionally per unit mass a n d the other per unit surface area. The latter has a more constant relationship to various metabolic processes than does weightS; dosage per unit surface area is advocated in pediatric therapeutics to avoid the pitfall of underdosing in younger children. 8 Second, we documented the extent of
481
48 2
Bloch, Clemons, and Sperling
The Journal of Pediatrics March 1987
T a b l e . Glucose and insulin responses to oral glucose tolerance test in prepubertal and pubertal subjects Prepubertal
Pubertal
Glucose dose
Glucose dose
1.75 g / k g (n = 9)
Fasting blood glucose (rag/alL) Peak blood glucose (mg/dL) Area under glucose curve (mg/dL X 4 hr)* Area under insulin curve (/zU/mL • 4 hr)t
82.0 •
55 g / m 2 (n = 8)
1.75 g / k g (n = 10)
55 g / m 2 (n = 9)
3.1
75.6 +
4.1
84.3 _+ 3.0
83.2 +
3.7
151.6 + 8.5
143.3 •
7.5
152.2 _+ 7.9
148.8 +
11.5
421.1 • 17.0
409.5 • 16.3
432.7 + 17.9
429.1 _+ 22.9
118.5 + 17.3~
115.4 +_ 16.5w
299.1 • 77.6:~
365.4 + 114.8w
Values represent mean + SEM. *To convert glucose values to mmol/L, multiply by 0.055. tTo convert insulin values to pmol/L, multiply by 7.175. :~w symbols compared, P <0.05.
the true change in insulin sensitivity as measured via the hyperinsulinemic-euglycemicclamp procedure9 in peripuber~al children. We also evaluated the relationship of insulin sensitivity to sex, body mass index, clinical stage of puberty, and biochemical indices of adrenarche and growth hormone secretion as reflected by, respectively, serum dehydroepiandrosterone sulphate and somatomedin C/insuliniike growth factor I levels. METHODS The subjects were 22 healthy children who are siblings of patients with insulin-dependent diabetes mellitus attending the Diabetes Clinic at our medical center. Informed consent was obtained from the parents, and all proposed procedures were explained to the children. Each participant was assessed clinically and assigned a pubertal~ rating according to the method of Marshall and Tanner.10.1~ Body mass index was calculated as the weight in kilograms divided by the square of the height in meters. 12 Body surface area was determined from the formula of Haycock et al. la The 22 children were equally divided into two groups, half prepubertal and half pubertal, with seven boys and four girls in each group. Mean (_+ SD) ages of the two groups were 8.1 + 1.7 and 12.4 _+ 1.7 years, respectively (P <0.001). There were no differences in body mass index between boys and girls within the prepubertal or pubertal groups, but the pubertal subjects had a significantly greater index (20.1 + 4.3 vs 16.1 _+ 2.3; P <0.02). Within the pubertal group, five children were ranked at Tanner stage II, five at stage III or IV, and one at stage V. Subjects were admitted to the Clinical Research Center
after an overnight fast. A 4-hour OGTT was performed with random assignment to a glucose dose calculated per unit mass (1.75 g/kg) or unit surface area (55 g/m2), with a maximum dose in each instance of 75 g. Blood was withdrawn from a peripheral vein at 0, 30, 60, 120, 180, and 240 minutes for immediate measurement of whole blood glucose and later assay of serum insulin. The subjects remained in the Clinical Research Center and were fed an appropriate diet for age. After a second overnight fast, each subject underwent a hyperinsulinemie-euglycemic clamp procedure. 9 Soluble insulin was infused into a peripheral vein of one hand at a constant rate of 40 mU/m2/min, a rate designed to raise the insulin concentration to approximately 100 # U / m L (717.5 pmol/L). Blood was sampled every 10 minutes from a dorsal hand vein in the contralateral limb, at ambient temperature, for immediate measurement of glucose and later measurement of insulin and mean DHEA-S and Sm-C/IGF-I. Five minutes after commencing the insulin infusion, an infusion of 15% dextrose in water was commenced via a Harvard infusio~ pump at a variable rate determined by an algorithm programmed into a hand-held calculator. The algorithm was modified by the substitution of surface area for body mass and an adjustment interval of 10 minutes. The blood glucose was clamped at 90 mg/dL and maintained at this steady state level for a period of 90 minutes. The insulin infusion was then discontinued and blood glucose monitored for a further 60 minutes to avoid hypoglycemia. Subjects were then discharged from the Clinical Research Center, and readmitted approximately 1 month later for the repeat OGTT with the alternate dose of glucose calculated per unit mass or
Volume 110 Number 3
Pubertal insulin resistance
483
2000--
300 --
I--q BOYS P < O . 0 1 ~
GIRLS
P
I
i BOYS
._L.._.
1500--
~ " ~ GIRLS
////~
200 2 & OD 1 0 0 0 - i < w -1G3 >, 9~
100
500 --
N ~'//z
t-
-///, ~///,
////, ,111~
t-
Prepubertal
Prepubertal
=ubertal
Fig. ,1, Insulin sensitivity index, calculated as glucose utilizafioi'i rate (mg/m 2 9 min) divided by log insulin concentration (#U/ mL), in prepubertal and pubertal children. (Mean _+ SEM.)
per surface area as indicated. This protocol had been reviewed and approved by the Institutional Review Board on Human Investigations. Biochemical measures. Whole blood glucose was immediately measured via the glucose oxidase method using a YSI glucose analyzer (Yellow Springs instrument Co., Inc., Yellow Springs, Ohio). Serum was immediately separated from the remainder of the samples and stored at - 2 0 ~ C foi" later assay of insulin, DHEA-S, and SmC / I G F - i by appropriate double-antibody radioimmunoassaysJ 4-'6 Glucose and insulin responses during OGTT were analyzed as the respective areas Under the curves, as previously described? The derived rate of glucose utilization at each 10-minute interval, corrected for unit surface area, was divided by the log of the insulin concentration to yield the "insulin Sensitivity index," because the relationship between glucose utilization and insulin concenti'ation is log linear between insulin concentrations of 20 and 200 # U / m L (143.5 to 1435 pmol/L)/7 The insulin sensitivity index for each subject is the mean of the calculated indices during the period of steady-state blood glucose. The values for S m - C / I G F - I were expressed logarithmically because a
Pubertal
Fig. 2. Dehydroepiandrosterone sulphate concentrations in prepubertal and pubertal children. (Mean _+ SEM.) To convert DHEA-S Concentration to nmol/L, multiply by 0.26.
normal frequency distribution occurs only with its log concentration.'8 Statistical analyses. Statistical analyses were performed via analysis of variance using the Student Newman-Keuls test if the F value was significant, or simple linear regression with the Pearson correlation coefficient? 9 Multivariate analysis was also used to examine the relationship between insulin sensitivity, age, body mass index, DHEAS, and Sm-C/IGF2I as independent variables. 2~Results are expressed as mean • SEM except where otherwise indicated. RESULTS Glucose and insulin responses during OGTT are indicated in the Table. The mean area under the glucose curves was similar between and within each group irrespective of the dose of glucose administered. Insulin area under the curve also bore no relationship to glucose dose; however, in each instance, the pubertal group had a significantly greater insulin response (P <0.05). Eight children (five prepubertal, three pubertal) did not return for repeat OGTT.
484
Bloch, Clemons, and Sperling
3--
2
~
The Journal of Pediatrics March 1987
p-0.oo3 .--._L_
--
5E
r---1BOYS ~
GIRLS
I-E i U_
//////~
-~ o.5 (..) u E
//////
basis of DHEA-S concentration less than or greater than 300 ng/dL (78 nmol/L), respectively,z~ The preadrenarchal group had an insulin sensitivity index of 232.8 _+ 8.2, versus 161.7 _+ 16.2 for the adrenarchal group (P <0.001). The prepubertal group had significantly lower Sm-C/ IGF-I values than those of the pubertal group (P <0.003, Fig. 3). Linear regression revealed that the insulin sensitivity index correlated inversely with body mass index (r = -0.49, P <0.02), DHEA-S (r = -0.57, P <0.01), and log Sm-C/IGF-I concentration (r = -0.45, P <0.05) (Fig. 4, A through C), but not with age (r = -0.32). None of these correlations was significant by multivariate analysis when considered as independent variables.
Y~44~ / / / / H
//////
~//////
"/////~
N
"///.~ /MH, /////~
z, ////i/-/ // /, ,
~/.~;/./.o •1//i.<
Prepubertal
HH~*
Pub(,rta]
Fig. 3. Somatomedin C/insulinlike growth factor l concentrations in prepubertal and pubertal Children, drawn to logarithmic scale. (Mean + SEM.)
With the hyperinsulinemic-euglycemicclamp (n = 22), mean steady-state blood glucose concentration was 89.7 _+ 0.3 and 89.4 _+ 0.3 mg/dL (4.93 + .02 and 4.92 _+ 0.02 mmol/L) in the prepubertal and pubertal children, respectively. The time to attain this steady-state blood glucose concentration was 91 _+ 6 and 94 _+ 10 minutes for the two groups, respectively, and the mean insulin concentrations at steady state were 90.3 _+_+4.6 and 106.1 _+7.5 ~ U / m L (647.9 _+ 33 and 761.3 _+ 53.8 pmol/L). These differences were not significant. The average glucose infusion rates to maintain the steady-state blood glucose level were 402.4 + 31.5 and 287.7 _+ 29.6 mg/min 9 m 2 in the prepubertal and pubertal children, respectively (P <0.02). The prepubertal group had a greater insulin sensitivity index (glucose infusion rate divided by log serum insulin concentration) than that in the pubertal group (P <0.01; Fig. 1). Overall, the insulin sensitivity index correlated inversely with the insulin response to OGTT (r = -0.62, P <0.01). The serum DHEA-S values in the pubertal group were, as expected, significantly higher than those in the prepubertal group (P <0.001, Fig. 2). We grouped the subjects as to biochemical preadrenarche and adrenarche on the
DISCUSSION The increasing insulin requirement of diabetic adolescents as they progress through puberty has been attributed to several factors, including psychologic adjustment problems and poor compliance, especially inasmuch as glycosylated hemoglobin values increase in these patients despite their increasing insulin dose. ~-3We have demonstrated that physiologic insulin resistance occurs peripubertally in children without diabetes. Thus, part of the increasing ir~sulin requirement of diabetic adolescents is attributable to a decline in insulin sensitivity, as recently reported. 7 We have demonstrated that in adolescents without diabetes the insulin resistance is compensated for by increased insulin secretion. The increased insulin secretion of adolescents during OGTT cannot be explained solely on the basis of administered glucose dose, as demonstrated by the comparison of insulin responses of prepubertal and pubertal children tested with glucose given per unit mass or per unit surface area. When glucose is given per unit mass, the differences in load may be considerable; at the 50th percentile for weight, a typical 14-year-old child would receive more than twice as much glucose as a typical 7-year-old.8 Hence it could be proposed that for equivalent glycemia, the greater glucose load would elicit more insulin secretion. By contrast, the difference in surface area of these representative individuals is only about 60%, with a corresponding reduction in the difference between administered glucose loads. Nevertheless, insulin responses during OGTT were greater in adolescents, irrespective of the absolute glucose load. Moreover, the insulin response to OGTT correlated inversely with insulin sensitivity, assessed via the hyperinsulinemic-euglycemic clamp procedure. Thus the higher insulin responses of adolescents during OGTT appear to reflect reduced insulin sensitivity, as previously suggested2 During the euglycemic clamp procedure, the actual
Volume 110 Number 3
mean blood glucose concentrations at steady state, time to achieve the clamp, and the steady-state insulin concentrations were equivalent in both groups, and similar to those reported in studies done at equal insulin infusion rates. 7,]7 With insulin concentrations of approximately 100 #U/mL, hepatic glucose output is negligible, so the rate of glucose infusion necessary to maintain euglycemia is a measure of glucose utilization, and hence sensitivity to the ambient insulin concentration.17 This relationship is log-linear between insulin concentrations of 20 and 200 p U / m L (143.5 to 1435 pmol/L). J7 By these measures, insulin sensitivity was reduced by approximately 30% in adolescents. This extent of insulin resistance is similar to that recently reported in a group of children without diabetes. 7 Because of the log-linear relationship that exists between insulin concentration and glucose utilization, a substantial increase in serum insulin concentration would be expected to result in only a modest increase in glucose utilization. This would explain the threefold increase in the insulin area under the curve observed in the adolescents during OGTT, but only a 30% decrease in the insulin sensitivity index. However, diabetic adolescents may have an additional element of insulin resistance. Insulin resistance is a prominent feature of insulin-dependent diabetes, 22 and recent data indicate that diabetic adolescents have a further reduction of insulin sensitivity compared with that of the pubertal nondiabetic children we report. 7,23 The reduced insulin sensitivity of pubertal children could result from a number of causes. First, body mass increases at about the time of puberty, as demonstrated in our study. Obesity causes insulin resistance at the insulin receptor and postreceptor sites24; however, our subjects were not obese) 1 Linear regression demonstrated a significant inverse correlation between body mass index and insulin sensitivity. Although an increase in body mass index may partly account for the peripubertal decline in insulin sensitivity, it fails to exp!ain the apparent sex difference in insulin sensitivity observed in prepubertal children, in whom there was no difference of body mass index. Second, there was an inverse relationship between insulin sensitivity and levels of DHEA-S, the predominant adrenal hormone of adrenarche, 2~ and preadrenarchal children were more insulin sensitive than adrenarchal children. These findings are intriguing in that they suggest that adrenarche, rather than puberty, is an earlier predictor of insulin resistance. Several syndromes of hyperandrogenemia are associated with insulin resistance.~5 Thus adrenarche, which occurs earlier in girls than in boys,21and the resultant rise of adrenal androgens may be a factor responsible for the decline in insulin sensitivity. Preliminary in vivo and in vitro evidence suggests that androgens
Pubertal insulin resistance
485
28 9
==
0 Q3
iy=.O.22x+23.SSI
9
~
24
. "
20
16
12
2000
l o
I
I lO0
I
l 200
I
I 300
I
-B 9
1500
E
[ I
I~=-o.49 I~
9
9
9
1000
y=-4.g4x + 1584.4 r =-0.57 D~O.01
%
<
o
500
I 0
I 100
1
1 200
1
I 300
I
C _~ E
9
2
~
9
[ Y=-O'OO2x+0'411
9
9
I r=-o,4s
I
1 ,~
..
0.5
E
O3
0.1
I 0
I
I 100
I
I 200
1
I 300
Insulin Sensitivity Index Fig. 4. Simplelinear regression of insulin sensitivity index versus body mass index (A), DHEA-S (B), and Sm-C/IGF-I (C).
may alter carbohydrate metabolism by reducing insulin sensitivity?6. 27 Third, growth hormone secretion increases in the peripubertal period? 8 Thus, increased growth hormone secretion, a known cause of insulin resistance, 29 may also be a factor responsible for the peripubertal decline in insulin sensitivity. Although we did not measure growth hormone secretion directly, we did measure Sm-C/IGF-I, which is a marker of growth hormone secretory status. Unlike growth hormone, Sm-C/IFG-I concentration in serum remains stable throughout the day because of its relatively long
486
Bloch, Clemons, and Sperling
half-life. TM S m - C / I G F - I is a valid measure of growth hormone secretion in well-nourished, healthy children, and when plotted as its logarithmic concentration, there is a normal frequency distribution. 18 Although S m - C / I G F - I is an indirect measure of growth hormone action, it does not mediate insulin resistance, which is attributed to a direct effect of growth hormone independent of S m - C / I G F - I 5 9 Our results indicate that the log S m - C / I G F - I concentration correlates inversely with insulin sensitivity. These data would be consistent with reports that correlate S m - C / IGF-I, and by implication growth hormone, with pubertal status.30,3L They are also consistent with the recent report in which integrated growth hormone concentrations were measured and found to correlate inversely with insulin sensitivity. 7 Although we did demonstrate multiple inverse correlations with insulin sensitivity via univariate analyses for each of the listed factors, multivariate analysis failed to demonstrate any single factor that independently correlated with the insulin sensitivity index. This indicates that the loss of insulin sensitivity during the peripubertal period i s multifactorial. It is possible that the three factors we have identified--increasing adrenal androgens, body mass index, and growth hormone as reflected by S m - C / I G F I - - i n t e r a c t to mediate insulin resistance. Thus, adrenal androgens increase the body mass index and augment growth hormone secretion. 32 The physiologic decrease of insulin sensitivity at the time of adrenarche/puberty implies that the clinician managing type I insulin-dependent diabetes mellitus in children should anticipate an increase in insulin requirement at the time of adrenarche/ puberty. We thank the nursing staff of the Clinical Research Center; Steven D. Chernausek, M.D., Division of Endocrinology, Department of Pediatrics, University of Cincinnati, for assay of SmC/1GF~I; and Jane Khoury, B.Sc., Department of Biostatistics, University of Cincinnati, for statistical analysis of the data. REFERENCES
1. Mann NP, Johnston DI. Improvement in metabolic control in diabetic adolescents by the use of increased insulin dose. Diabetes Care 1984;7:460-4. 2. Blethen SL, Sargeant DT, Whitlow MG, Santiago JV. Effect of pubertal stage and recent blood glucose control on plasma somatomedin C in children with insulin-dependent diabetes mellitus. Diabetes 1981;30:868-72. 3. Daneman D, Wolfson DH, Becker D J, Drash AL. Factors affecting glycosylated hemoglobin values in children with insulin-dependent diabetes. J PEDIATR 1981;99:847-53. 4. Wing RR, Nowalk MP, Marcus MD, Koeske R, Finegold D. Subclinical eating disorders and glycemic control in adolescents with type I diabetes. Diabetes Care 1986;9:162-7. 5. Rosenbloom AL, Wheeler L, Bianchi R, Chin FT, Tiwary CM, Grgic A. Age-adjusted analysis of insulin responses
The Journal of Pediatrics March 1987
6.
7.
8. 9.
10. 11. 12. 13.
14.
15.
16.
17.
18.
19. 20. 21.
22.
23.
24.
25.
during normal and abnormal glucose tolerance tests in children and adolescents. Diabetes 1975;24:820-8. National Diabetes Data Group: Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance. Diabetes 1979;28:1039-57. Amiel SA, Sherwin RS, Simonson DC, Lauritano AA, Tamborlane WV. Impaired insulin action in puberty. N Engl J Med I986;315:215-9. Shirkey HC, Dosage (posology). In: Shirkey HC, ed. Pediatric therapy, 5th ed. St. Louis: CV Mosby, 1975;19-33. DeFronzo RA, Tobin JD, Andres R. Glucose clamp technique: a method for quantifying insulin secretion and resistance. Am J Physiol 1979;237:E214-23. Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in girls. Arch Dis Child 1969;44:291-303. Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in boys. Arch Dis Child 1970;45:13-23. James WPT. Treatment of obesity: the constraints for success. Clin Endocrinol Metab 1984;13:635-59. Haycock GB, Schwartz G J, Wisotsky DH. Geometric method for measuring body surface area: a height-weight formula validated in infants, children, and adults. J PEDIATR 1978; 93:62-6. Sperling MA, DeLamater PV, Phelps D, Fiser RH, Oh W, Fisher DA. Spontaneous and amino acid-stimulated glucagon secretion in the immediate postnatal period: relation to glucose and insulin. J Clin Invest 1974;53:1159-66. Korth-Schultz S, Levine LS, NewMI. Dehydroepiandrosterone sulphate (DS) levels: a rapid test for abnormal adrenal androgen secretion. J Clin Endocrinol Metab 1976;42:t00513. Furlanetto RW, Underwood LE, Van Wyk J J, D'Ercole AJ. Estimation of somatomedin-C levels in normals and patients with pituitary disease by radioimmunoassay. J Clin Invest 1977;60:648-57. Rizza RA, Mandarino L J, Gerich JE. Dose-response characteristics for effects of insulin on production and utilization of glucose in man. Am J Physiol 1981;240:E630-39. Clemmons DR, Van Wyk JJ. Somatomedin: physiological control and effects on cell proliferation. In: Baserga R, ed. Handbook of experimental pharmacology. New York: Springer-Verlag, 1981;57:t61-208. Winer BJ. Statistical principles in experimental design. New York: McGraw-Hill, 1971. Morrison DF. Multivariate statistical methods, 2nd ed. New York: McGraw-Hill, 1970;159-206. Pang S. Premature adrenarche. In: New MI, Levine LS, eds. Adrenal diseases in childhood. Pediatric Adolescent Endocrinology. New York: Karger, 1984;13:173-84. DeFronzo RA, Hendler R, Simonson D. Insulin resistance is a prominent feature of insulin-dependent diabetes. Diabetes 1982;31:795-801. Landt KW, Campaigne BN, James FW, Sperling MA. Effects of exercise training on insulin sensitivity in adolescents with type I diabetes. Diabetes Care 1985;8:461-5. Kolterman O, Insel J, Saekow M, Olefsky JM. Mechanisms of insulin resistance in human obesity. J Clin Invest 1980;65:1272-84. Richards GE, Cavallo A, Meyer W J III, et al. Obesity, acanthosis nigricans, insulin resistance, and hyperandrogenemia: pediatric perspective and natural history. J PEDIATR 1985;107:893-7.
Volume 110 Number 3
Pubertal insulin resistance
26. Peiris A, Smith GA, Kissebah AH. Relationship of sex hormones to splanchnic insulin metabolism and peripheral insulin sensitivity. Endocrinology 1986;118(Suppl):309A. 27. Mueller PL, Kwan AH. Dehydroepiandrosterone inhibits insulin-stimulated glycogen synthesis in minimal deviation hepatoma cells. Endocrinology 1986;118(Suppl):1026A. 28. Zadik Z, Chalew SA, McCarter R J, Meistas M, Kowarski AA. The influence of age on the twenty-four-hour integrated concentration of growth hormone in normal individuals. J Clin Endocrinol Metab 1985;60:513-6. 29. Rizza RA, Mandarino L J, Gerich JE. Effects of growth hormone on insulin action in man. Diabetes 1982;31:663-9.
487
30. Rosenfield RI, Furlanetto R, Bock D. Relationship of somatomedin-C concentrations to pubertal changes. J PEDIATR 1983;103:723-8. 31. Luna AM, Wilson DM, Wibbelsman C J, 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 1983;57:268-71. 32. Link K, Blizzard RM, Evans WS, Kaiser DL, Parker MW, Rogol AD. The effect of androgens on the pulsatile release and the twenty-four-hour mean concentration of growth hormone in peripubertal males. J Clin Endocrinol Metab 1986;62:159-64.
FELLOWSHIPS Available fellowships in pediatric subspeciaIties and those for general academic pediatric training are listed once a year, in May, in The JourrLal of Pediatrics. Each October, forms for listing such fellowships are sent to the Chairman of the' Department of Pediatrics at most major hospitals in the United States and Canada. Should you desire to list fellowships, a separate application must be made each year for each position. All applications must be returned to The C. V. Mosby Company by February 15 of the listing year to ensure publication. Additional forms will be supplied on request from the Journal Editing Department, The C. V. Mosby Company, 11830 Westline Industrial Drive, St. Louis, M O 63146/314-872-8370.