Elevated Leptin Levels Are Associated with Excess ... - Oxford Journals

9 downloads 0 Views 174KB Size Report
M. L. AHMED, K. K. L. ONG, A. P. WATTS, D. J. MORRELL, M. A. PREECE, AND .... David. B. Dunger, University Department of Pediatrics, Level 8, Box 116, Ad-.
0021-972X/01/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 2001 by The Endocrine Society

Vol. 86, No. 3 Printed in U.S.A.

Elevated Leptin Levels Are Associated with Excess Gains in Fat Mass in Girls, But Not Boys, with Type 1 Diabetes: Longitudinal Study during Adolescence M. L. AHMED, K. K. L. ONG, A. P. WATTS, D. J. MORRELL, M. A. PREECE, D. B. DUNGER

AND

University Department of Pediatrics, Addenbrooke’s Hospital (M.L.A., K.K.L.O., A.P.W., D.B.D.), Cambridge, United Kingdom CB2 2QQ; Reproductive Medicine Laboratory, University of Edinburgh Centre for Reproductive Biology (D.J.M.), EH11 9EW Edinburgh, United Kingdom; and Institute of Child Health (M.A.P.), WC1N 1AH London, United Kingdom ABSTRACT Adolescents, in particular girls, with type 1 diabetes may gain excessive weight during puberty. We present the results of a longitudinal study aimed to determine the roles of leptin and insulin in changes in body composition in subjects with type 1 diabetes and controls. Forty-six children (23 boys) with type 1 diabetes and 40 controls (20 boys) were followed from 8 –17 yr of age. Height, weight, and sc skinfolds were assessed every 6 months, and a blood sample taken for leptin determination. Throughout the age range, body mass index (mean ⫾ SEM) was greater by 1.45 ⫾ 0.69 kg/m2 in girls and

A

NUMBER OF recent studies (1– 6) have shown that pubertal growth may be impaired, particularly in girls with type 1 insulin-dependent diabetes mellitus. Greater gains in weight or weight for height during puberty relative to control values have been reported either in both sexes (7–10) or only in girls (4, 11–14). Most of these studies were cross-sectional, and analysis was based on assessment of weight or body mass index (BMI); the inference being that there are excessive increases in fat mass in adolescents with type 1 diabetes. However, with the exception of Gregory et al. (13), measurements of body composition have rarely been undertaken, and the longitudinal changes during puberty in adolescents with type 1 diabetes have not been determined. The mechanism by which there could be excessive weight gain for height in subjects with type 1 diabetes has not been identified. It has been generally accepted that it might relate to insulin dose or frequency of insulin injections (8, 11–18), but it remains a paradox as to why the same insulin dose should be associated with impaired statural growth yet increasing body fat, particularly in girls. The recent discovery of the hormone leptin (19), which is produced by adipocytes and regulates appetite, food intake, and energy metabolism in rodents (20 –22) and humans (23), could provide a clue to the pathogenesis of excess weight gain in subjects with type 1 diabetes. Raised leptin levels have been observed in chil-

Received January 4, 2000. Revision received August 29, 2000. Accepted October 26, 2000. Address all correspondence and requests for reprints to: Prof. David B. Dunger, University Department of Pediatrics, Level 8, Box 116, Addenbrooke’s Hospital, Hills Road, Cambridge, United Kingdom CB2 2QQ. E-mail: [email protected].

1.46 ⫾ 0.55 kg/m2 in boys with type 1 diabetes compared with control values. In girls with type 1 diabetes, this reflected greater percent body fat (3.2 ⫾ 1.0%; P ⫽ 0.002), whereas in boys it related to differences in fat-free mass. Both boys and girls with type 1 diabetes had higher leptin levels adjusted for percent body fat than controls; in the girls this was related to insulin dose (regression coefficient B ⫽ 0.006 ⫾ 0.003; P ⫽ 0.04) and greater gains in fat mass. Hyperinsulinemia and raised leptin levels are associated with gains in fat mass throughout puberty in girls, but not boys, with type 1 diabetes. (J Clin Endocrinol Metab 86: 1188 –1193, 2001)

dren and adolescents with type 1 diabetes by several investigators (24, 25). It has been postulated that this might reflect leptin resistance resulting from the peripheral hyperinsulinemia that invariably occurs in type 1 subjects treated with sc rather than portal insulin administration (26). However, Verrotti (27) did not find any difference in leptin levels between children and young adults with type 1 diabetes and matched controls. Previous studies of the relationships among insulin dose, leptin levels, and fat mass in type 1 subjects have been crosssectional or have used a mixed cross-sectional/longitudinal design. In these studies data have often been contrasted with population control data instead of contemporary control values, and assessment of fat mass has usually been inferred from BMI. We report the first study of children and adolescents with type 1 diabetes in which subjects were followed prospectively to assess changes in body composition and leptin levels. Data are compared with those from a parallel control group of normal subjects studied following an identical protocol. Experimental Subjects

Forty-six children (23 boys and 23 girls) attending the pediatric diabetes clinic at the John Radcliffe Hospital (Oxford, UK) were recruited for a longitudinal study of growth and puberty. The median age (range) at recruitment was 11.27 (10.53–12.83) yr in the boys and 10.36 (8.54 –11.36) yr in the girls. The durations of diabetes at recruitment were 6.17 (1.87–10.16) and 2.69 (0.55–9.15) yr, respectively. This difference in duration of diabetes was significant (P ⬍ 0.001). The median (range) duration of follow-up was 6 (3– 8) yr.

1188

ELEVATED LEPTIN AND GAIN IN FAT MASS IN DIABETIC GIRLS

All of the children were receiving twice daily injections of intermediate and soluble insulin at recruitment, and 75% were changed to multiple injection therapy during puberty. This consisted of three preprandial injections of soluble insulin, with the intermediate acting insulin given at bedtime. None of the children had any other significant medical condition, except one girl with treated hypothyroidism that was subsequently verified as being transient. Forty healthy normal children (20 boys and 20 girls) were followed from the beginning of puberty to assess normal growth and endocrine function during puberty (28). Their median age (range) at recruitment was 9.52 (8.63–10.09) yr in the boys and 9.57 (8.67–10.16) yr in the girls. The median (range) duration of follow-up was 6.4 (5.0 – 6.4) yr. The children were seen every 6 months, with measurements made, puberty staging performed, and blood specimens taken as described for the diabetic children. Materials and Methods All children with type 1 diabetes were seen every 3 months. Height, weight, and skinfold measurements were made by one observer (M.L.A.) using standard techniques (29). Puberty stages were assessed at each visit by the method of Tanner (30). Nonfasting blood samples obtained between 1000 and 1400 h were taken at 6-month intervals, spun, separated, and stored at ⫺20 C. Control subjects were assessed every 6 months, and measurements and blood samples were made as previously described (28). Height was measured using a Harpenden stadiometer (Holtain Ltd., Crymych, Wales). Weight was measured on a beam balance (diabetes clinic) and electronic scales (control study). Subcutaneous skinfold measurements were made on the left side of the body at four sites (biceps, triceps, subscapula, and suprailiac) using a Harpenden skinfold caliper (Holtain Ltd., Crymych, Wales). BMI was calculated as weight/height (kilograms per meter2) (31) for each child at each visit. The BMI sd score (BMISDS) was calculated based on the 1995 British standards (32). Body density was calculated from four skinfold measurements using the equation reported by Brook (33) in prepubertal children up to the age of 11 yr and using the equation reported by Durnin and Rahaman (34) for pubertal children. Although these are not the most recent published equations available, they were chosen because they are populationspecific models. Percent body fat was then derived using the equation reported by Siri (35), and fat mass (kilograms) was calculated as percent body fat ⫻ weight. Fat-free mass (kg) was obtained by subtraction (weight ⫺ fat mass).

1189

by regression coefficient (B) ⫾ sem. Using the mean for each child within a puberty stage (where more than one value was available), differences between type 1 subjects and controls were examined using Student’s t test. All analyses were performed using SPSS for Windows (version 9.0, SPSS, Inc., Chicago, IL). Multilevel modeling (MLwiN, Institute of Education, London, UK) was used to further examine longitudinal differences in leptin and body composition between the two cohorts. This method is an extension of multiple regression; it optimizes the use of repeated measures data by obviating the requirements for equal numbers of measuring occasions for each subject and allows the analysis of within-individual as well as between-individual effects.

Results BMI

BMI increased with age in all children: type 1 girls: B ⫽ 1.0005 ⫾ 0.027 (P ⬍ 0.0005); type 1 boys: B ⫽ 0.773 ⫾ 0.024 (P ⬍ 0.0005); control girls: B ⫽ 0.895 ⫾ 0.031 (P ⬍ 0.0005); and control boys: B ⫽ 0.681 ⫾ 0.026 (P ⬍ 0.0005). BMI was higher in both boys and girls with type 1 diabetes compared with control children throughout the age range studied. Overall, girls with type 1 diabetes had 1.45 ⫾ 0.69 kg/m2 higher BMI than control girls. Boys with type 1 diabetes had 1.46 ⫾ 0.55 kg/m2 BMI higher than the control boys (MLwiN). When analyzed by puberty stage, boys with type 1 diabetes had higher BMI at all stages compared with control boys. This was significant at each stage except puberty stage 3 (P ⫽ 0.09). In the girls, however, a significant difference in BMIs between subjects with type 1 diabetes and controls was only seen at stage 5 (P ⫽ 0.04; Fig. 1). Body composition

Analyses by age. In both girls with type 1 diabetes and control girls, percent body fat increased with age: B ⫽ 0.127 ⫾ 0.007 (P ⬍ 0.0005) and B ⫽ 0.180 ⫾ 0.012 (P ⬍ 0.0005), respectively. Percent body fat also increased with age in the control boys (B ⫽ 0.019 ⫾ 0.008; P ⫽ 0.03), but decreased in the boys with type 1 diabetes (B ⫽ ⫺0.047 ⫾ 0.008; P ⬍ 0.0005; Fig. 2). Overall, girls with type 1 diabetes had a higher percent body fat than the control girls by 3.2 ⫾ 1.0% (P ⫽ 0.002), whereas there was no difference between the type 1 and control boys (0.68 ⫾ 1.27%; P ⫽ 0.6; MLwiN).

Hemoglobin A1c (HbA1c) was measured by high pressure liquid chromatography (Diamat, Bio-Rad Laboratories, Inc., Hemel Hempstead, UK). The intraassay coefficients of variation (CVs) were 1.9% and 2.2% at HbA1c levels of 6.9% and 11.5%, respectively. Interassay CVs were 2.7% and 2.3% at HbA1c levels of 7.0% and 11.6%, respectively. Serum leptin was measured by RIA (Linco Research Co., St Charles, MO). The detection limit of the assay was 0.5 ng/mL. The intraassay CVs were 8.3%, 3.9%, and 3.4% at leptin levels of 4.9, 10.4, and 25.6 ng/mL, respectively. Interassay CVs were 6.2%, 4.7%, and 3.6% at leptin levels of 4.9, 10.4, and 25.6 ng/mL, respectively.

Analyses by puberty stage. Between pubertal stages 1–5, girls with type 1 diabetes gained more body fat than control girls (P ⫽ 0.04; Table 1). In contrast to analysis by age (Fig. 2), no significant change in percent body fat was seen by pubertal stage in control boys; in fact, a marginal loss was seen (Table 1). This difference may relate to the shortened time frame of analyses by pubertal stage in contrast to age. However, compared with control boys, the decrease in percent body fat in boys with type 1 diabetes was still evident by pubertal stage (P ⫽ 0.008; Table 1).

Statistical analysis

Leptin by age and puberty

Leptin concentrations and body composition values were transformed to approximate normal distributions by log transformation or square roots as appropriate. Parametric tests were then used, with P ⬍ 0.05 considered significant. The data were analyzed longitudinally using an analysis of covariance method (36). This examines the association between variables within an individual by fitting parallel lines with a common slope for all subjects. Intersubject variation is shown by differences in the constants. Associations between variables are described

Leptin levels increased with age in both girls with type 1 diabetes and control girls: type 1 diabetes, B ⫽ 0.179 ⫾ 0.02 (P ⬍ 0.0005); and controls, B ⫽ 0.143 ⫾ 0.012 (P ⬍ 0.0005; Fig. 3A). In contrast, in both the boys with type 1 diabetes and control boys, leptin levels decreased with age: type 1 boys, B ⫽ ⫺0.040 ⫾ 0.015 (P ⫽ 0.01); and control boys, B ⫽ ⫺0.063 ⫾ 0.01 (P ⬍ 0.0005; Fig. 3B).

Assays

1190

AHMED ET AL.

FIG. 1. A, BMI (mean ⫾ 95% confidence interval) in girls by puberty stage. Type 1 diabetes (f and —-) vs. controls (F and - - -): P ⫽ NS at all stages except stage 5, where P ⫽ 0.04. B, BMI (mean ⫾ 95% confidence interval) in boys by puberty stage. Type 1 diabetes (f and —-) vs. controls (F and - - -): stage 1, P ⬍ 0.0005; stage 2, P ⫽ 0.003; stage 3, P ⫽ 0.09; stage 4, P ⫽ 0.004; stage 5, P ⫽ 0.001. *, P ⬍ 0.05; **, P ⬍ 0.005, ***, P ⬍ 0.0005 (type 1 vs. controls).

Leptin levels in subjects with type 1 diabetes were significantly greater in both sexes at each stage of puberty than in the control children. In the girls P ⬍ 0.0005 for all stages of puberty of type 1 diabetes vs. controls; in the boys P ⫽ 0.001 at stage 2, 0.008 at stage 3, and less than 0.0005 between the two cohorts at stages 4 and 5 (stage 1 in the boys was omitted because there were only two boys with type 1 diabetes with leptin levels; Fig. 4). Leptin and body composition

Both fat mass and fat-free mass were individually related to leptin levels in girls with type 1 diabetes and control girls

JCE & M • 2001 Vol. 86 • No. 3

FIG. 2. A, Percent body fat in girls plotted against age. F and —-, Type 1 diabetes (B ⫽ 0.127 ⫾ 0.007); E and - - -, controls (B ⫽ 0.180 ⫾ 0.012 (P ⬍ 0.0005 for both groups). B, Percent body fat in boys plotted against age. F and —-, Type 1 diabetes (B ⫽ ⫺0.047 ⫾ 0.008; P ⬍ 0.0005); E and - - -, controls (B ⫽ 0.0186 ⫾ 0.008; P ⬍ 0.03).

(P ⬍ 0.0005). On entering both variables into the same covariance model, a significant and positive relation of fat mass on leptin was seen in the girls of both groups: type 1 diabetes, B ⫽ 0.694 ⫾ 0.135 (P ⬍ 0.0005); and controls, B ⫽ 0.312 ⫾ 0.106 (P ⫽ 0.004). Although fat-free mass in the control girls was negatively related to leptin levels (P ⬍ 0.0005), this relationship was not seen in the girls with type 1 diabetes (P ⫽ 0.6). In both boys with type 1 diabetes and controls, fat mass was positively related to leptin (P ⬍ 0.0005), whereas fat-free mass was negatively related (P ⬍ 0.0005). In a case nested multilevel model using the whole dataset, controls and fat-free mass had negative effects on leptin levels, and fat mass had a positive effect. No significant sex difference was seen: Lnleptin ⫽ 0.481(0.100) ⫺ [(0.536 ⫾

ELEVATED LEPTIN AND GAIN IN FAT MASS IN DIABETIC GIRLS

1191

TABLE 1. Change in fat mass, percent body fat, and fat-free mass during puberty: comparison between adolescents with type 1 diabetes and controls Type 1

Fat mass (kg) Girls Boys % Body fat Girls Boys Fat-free mass Girls Boys

Controls

P

10.7 ⫾ 1.03 3.5 ⫾ 0.53

8.2 ⫾ 0.41 4.7 ⫾ 0.77

8.1 ⫾ 1.19 ⫺3.7 ⫾ 0.78

7.6 ⫾ 1.08 ⫺0.2 ⫾ 0.96

0.8 0.008

13.7 ⫾ 0.81 24.0 ⫾ 1.1

13.6 ⫾ 0.67 21.9 ⫾ 0.59

0.9 0.1

Values are the mean ⫾

0.04 0.2

SEM.

0.047) ⫻ controls] ⫺ [(0.029 ⫾ 0.002) ⫻ fat-free mass] ⫹ [(0.699 ⫾ 0.034) ⫻ fat mass1/2)] ⫺ [(0.069 ⫾ 0.058) ⫻ male]. In the girls with type 1 diabetes, from puberty stages 1– 4, the total daily insulin dose adjusted for percent body fat was positively related to leptin levels (B ⫽ 0.006 ⫾ 0.003; P ⫽ 0.04). No such relationship was seen in the boys (P ⫽ 0.1). Discussion

Previous studies have shown the propensity for subjects with diabetes to gain more weight for height during puberty (7–11, 13). The BMI data from our study demonstrate a similar trend. However, neither BMI nor weight unravels differences in the accumulation of fat mass and fat-free mass. In this respect our data were interesting in that the increases in BMI in girls were largely due to the recruitment of fat mass, whereas the higher BMI in boys was accompanied by a reduction in percent body fat. These differences highlight important sexual dimorphism in subjects with diabetes and parallel similar dimorphism observed in the reduced, blunted pubertal growth in girls, but not in boys, with diabetes (5). There is much debate about the mechanisms by which pubertal growth is blunted and weight gain enhanced in girls, but not in boys, with type 1 diabetes. It has been argued that the abnormalities of the GH/insulin-like growth factor I (IGF-I) axis (elevated GH levels and reduced IGF-I levels) commonly observed in type 1 subjects (37) could be involved. These abnormalities are thought to arise because of inadequate portal delivery of insulin with impaired generation of IGF-I and subsequent feedback drive for enhanced GH secretion. With conventional insulin therapy, the sc administration of insulin means that portal levels may not reach normal levels, resulting in impaired IGF-I generation despite peripheral hyperinsulinemia (26). Reduced IGF-I production could therefore explain the impaired growth, whereas the peripheral hyperinsulinemia could lead to increased risk for obesity. However, as IGF-I levels are equally low in boys and girls, this model does not explain why reduced statural growth and increased weight gain are only seen in girls with type 1 diabetes, whereas boys appear to be relatively protected (1). Discovery of the hormone leptin (19), which could regulate satiety, energy expenditure, and weight gain, added complexity to these relationships. Leptin is produced by adipocytes and is thought to feed back through the hypothalamic

FIG. 3. Leptin (log scale) in girls plotted against age. F and —-, Type 1 diabetes (B ⫽ 0.179 ⫾ 0.02); E and - - -, controls (B ⫽ 0.143 ⫾ 0.012; P ⬍ 0.0005 for both groups). B, Leptin (log scale) in boys plotted against age. F and —-, Type 1 diabetes (B ⫽ ⫺0.040 ⫾ 0.015; P ⬍ 0.01); E and - - -, controls (B ⫽ ⫺0.063 ⫾ 0.01; P ⬍ 0.0005).

receptors to regulate weight gain and energy expenditure (38). Whereas it was initially anticipated that leptin deficiency might be a cause of obesity (39), this has proved to be very rare. Paradoxically, however, common adult obesity is related to elevated leptin levels, possibly indicative of leptin resistance (40). Leptin resistance has been inferred from cross-sectional data in subjects with type 1 diabetes in whom leptin levels are higher than would be predicted for the degree of fat mass (24, 25). Our data confirm these reports that children with type 1 diabetes have higher leptin levels that are not explained by their differences in body composition. All attempts at intensification of insulin therapy lead to weight gain far in excess of that observed with more standard ther-

1192

JCE & M • 2001 Vol. 86 • No. 3

AHMED ET AL.

levels even at the start of puberty, girls with type 1 diabetes still subsequently gain more body fat compared with normal girls. However, the term leptin resistance needs to be used with caution, as the role of leptin in the regulation of normal pubertal weight gain has yet to be confirmed (28). Furthermore, the boys with type 1 diabetes also had elevated leptin levels, yet subsequently showed higher gains in fat-free mass. Sexual dimorphism in the regulation of leptin levels and gains in fat mass have been reported in normal subjects (28, 43– 46). In our own longitudinal study of the normal cohort presented here we observed that fat mass had a positive effect on leptin levels, whereas fat-free mass had a negative effect on leptin during puberty (28). The trend for leptin levels to rise in girls and fall in boys reflected these relationships, as girls tend to gain more fat mass and boys more fat-free mass during puberty (28). Similar observations exist for the type 1 subjects. Despite the continuing decrease in percent body fat in the boys and increasing fat mass in the girls, leptin levels are consistently higher in both sexes than in the controls. It is likely that hyperinsulinemia is a major determinant of the elevated leptin levels that we observed. However, alterations in sex steroids could also contribute, as estrogen may enhance leptin levels (47) and weight gain (48). In contrast, testosterone is related to fat-free mass and lower leptin levels (49, 50). Thus, although their role is unclear, sex steroids may make a major contribution to the differences in leptin levels and changes in body composition observed in children and adolescents with type 1 diabetes. Acknowledgments We thank all the children who participated in the studies over the years. We are very grateful for their cooperation and endurance. We also thank all the parents and hospital and school staff who gave their help and support. Special mention to Janet Baines-Preece, Noel Cameron, and Les Cox for their care in the measuring and data collection of the control study, and to Sally Strang and Dot Harris for their help and support with the children and blood samples in the children’s diabetic clinic. We are very grateful to Huiqi Pan who gave time and advice on the use of MLwiN. And a warm thanks to Joanna Hilken for her help in typing the manuscript. FIG. 4. A, Mean leptin ⫾ 95% confidence interval (log scale) in girls by puberty stage. Type 1 diabetes (F and —-) vs. controls (f and - - -): P ⬍ 0.0005 for all stages. B, Mean leptin ⫾ 95% confidence intervals (log scale) in boys by puberty stage. Type 1 (F and —-) vs. controls (f and - -): stage 2, P ⫽ 0.001; stage 3, P ⫽ 0.008; stages 4 and 5, P ⬍ 0.0005.

apy (8, 11, 15–18), and it has been postulated that peripheral hyperinsulinemia could lead to inappropriately enhanced leptin secretion. Evidence to support this hypothesis comes from studies in which continuous administration of insulin leads to elevated leptin levels in excess of those predicted by gains in fat mass (41). However, excessive food intake may also lead to leptin resistance (42). One study (8) indicated that the weight gain in girls with type 1 diabetes was triggered by the frequency of insulin injections (and thus perhaps the induced hunger) rather than the overall insulin dose. Thus, insulin therapy or enhanced food intake may contribute to excess leptin secretion and putative leptin resistance. Leptin resistance may also be inferred by our novel longitudinal data, which show that despite having higher leptin

References 1. Ahmed ML, Connors MH, Drayer NM, Jones JS, Dunger DB. 1998 Pubertal growth in IDDM is determined by HbA1c levels, sex, and bone age. Diabetes Care. 21:831– 835. 2. Salerno M, Argenziano A, Di-Maio S, et al. 1997 Pubertal growth, sexual maturation, and final height in children with IDDM. Effects of age at onset and metabolic control. Diabetes Care. 20:721–724. 3. Jos J, Meteyer I, Farkas D, Oberkampf B. 1996 [Growth of children with insulin-dependent diabetes. Study of 104 cases] Croissance des enfants atteints de diabete insulinodependant. Etude de 104 observations. Arch Pediatr. 3:218 –226. 4. Du Caju MV, Rooman RP, Op de Beeck L. 1995 Longitudinal data on growth and final height in diabetic children. Pediatr Res. 38:607– 611. 5. Brown M, Ahmed ML, Clayton KL, Dunger DB. 1994 Growth during childhood and final height in type 1 diabetes. Diabet Med. 11:182–187. 6. Salardi S, Tonioli S, Tassoni P, et al. 1987 Growth and growth factors in diabetes mellitus. Arch Dis Child. 62:57– 62. 7. Holl RW, Heinze E, Seifert M, Grabert M, Teller WM. 1994 Longitudinal analysis of somatic development in paediatric patients with IDDM: genetic influences on height and weight. Diabetologia. 37:925–929. 8. Holl RW, Grabert M, Sorgo W, Heinze E, Debatin KM. 1998 Contributions of age, gender and insulin administration to weight gain in subjects with IDDM. Diabetologia. 41:542–547. 9. Pitukcheewanont P, Alemzadeh R, Jacobs WR, Jones BH, Eberle AJ. 1995

ELEVATED LEPTIN AND GAIN IN FAT MASS IN DIABETIC GIRLS

10. 11. 12. 13. 14. 15.

16. 17.

18. 19. 20. 21. 22. 23. 24. 25. 26. 27.

Does glycemic control affect growth velocity in children with insulin-dependent diabetes mellitus [published erratum appears in Acta. Diabetol 1996 Mar;33(1):57]. Acta Diabetol. 32:148 –152. Thon A, Heinze E, Feilen KD, et al. 1992 Development of height and weight in children with diabetes mellitus: report on two prospective multicentre studies, one cross-sectional, one longitudinal. Eur J Pediatr. 151:258 –262. Danne T, Kordonouri O, Enders I, Weber B. 1997 Factors influencing height and weight development in children with diabetes. Results of the Berlin Retinopathy Study. Diabetes Care. 20:281–285. Bognetti E, Macellaro P, Novelli D, et al. 1995 Prevalence and correlates of obesity in insulin dependent diabetic patients. Arch Dis Child. 73:239 –242. Gregory JW, Wilson AC, Greene SA. 1992 Body fat and overweight among children and adolescents with diabetes mellitus. Diabet Med. 9:344 –348. Mortensen HB, Hartling SG, Petersen KE. 1988 A nation-wide cross-sectional study of glycosylated haemoglobin in Danish children with type 1 diabetes. Diabet Med. 5:871– 876. Mortensen HB, Hougaard P. 1997 Comparison of metabolic control in a cross-sectional study of 2,873 children and adolescents with IDDM from 18 countries. The Hvidore Study Group on Childhood Diabetes [published erratum appears in Diabetes Care 1997 Jul;20(7):1216]. Diabetes Care. 20:714 –720. The Diabetes Control and Complications Trial Research Group. 1995 Adverse events and their association with treatment regimens in the diabetes control and complications trial. Diabetes Care. 18:1415–1427. Reichard P, Berglund B, Britz A, et al. 1991 Intensified conventional insulin treatment retards the microvascular complications of insulin-dependent diabetes mellitus (IDDM): the Stockholm Diabetes Intervention Study (SDIS) after 5 years. J Intern Med. 230:101–108. Diabetes Control and Complications Research Group. 1988 Weight gain associated with intensive therapy in the diabetes control and complications trial. The DCCT Research Group. Diabetes Care. 11:567–573. Zhang Y, Proenca R, Maffei M, et al. 1994 Positional cloning of the mouse obese gene and its human homologue. Nature. 372:425– 432. Stephens TW, Basinski M, Bristow PK, et al. 1995 The role of neuropeptide Y in the antiobesity action of the obese gene product. Nature. 377:530 –532. Maffei M, Fei H, Lee GH, et al. 1995 Increased expression in adipocytes of ob RNA in mice with lesions of the hypothalamus and with mutations at the db locus. Proc Natl Acad Sci USA. 92:6957– 6960. Vaisse C, Halaas JL, Horvath CM, et al. 1996 Leptin activation of Stat3 in the hypothalamus of wild-type and ob/ob mice but not db/db mice. Nat Genet. 14:95–97. Montague CT, Farooqi IS, Whitehead JP, et al. 1997 Congenital leptin deficiency is associated with severe early-onset obesity in humans. Nature. 387:903–908. Kiess W, Anil M, Blum WF, et al. 1998 Serum leptin levels in children and adolescents with insulin-dependent diabetes mellitus in relation to metabolic control and body mass index. Eur J Endocrinol. 138:501–509. Kamoda T, Saitoh H, Nakahara S, et al. 1998 Serum leptin and insulin concentrations in prepubertal lean, obese and insulin-dependent diabetes mellitus children. Clin Endocrinol (Oxf). 49:385–389. Dunger DB. 1992 Diabetes in puberty. Arch Dis Child. 67:569 –570. Verrotti A, Basciani F, Morgese G, Chiarelli F. 1998 Leptin levels in non-obese and obese children and young adults with type 1 diabetes mellitus. Eur J Endocrinol. 139:49 –53.

1193

28. Ahmed ML, Ong KK, Morrell DJ, et al. 1999 Longitudinal study of leptin concentrations during puberty: sex differences and relationship to changes in body composition. J Clin Endocrinol Metab. 84:899 –905. 29. Cameron N. 1984 The measurement of human growth. Kent: Croom Helm. 30. Tanner JM. 1962 Growth at adolescence, 2nd Ed. Oxford: Blackwell. 31. Garrow JS, Webster J. 1985 Quetelet’s index (W/H2) as a measure of fatness. Int J Obes. 9:147–153. 32. Cole TJ, Freeman JV, Preece MA. 1995 Body mass index reference curves for the UK, 1990. Arch Dis Child. 73:25–29. 33. Brook CG. 1971 Determination of body composition of children from skinfold measurements. Arch Dis Child. 46:182–184. 34. Durnin JV, Rahaman MM. 1967 The assessment of the amount of fat in the human body from measurements of skinfold thickness. Br J Nutr. 21:681– 689. 35. Siri W. 1956 The gross composition of the body. Adv Biol Med Phys. 4:239 –280. 36. Bland JM, Altman DG. 1995 Calculating correlation coefficients with repeated observations. I. Correlation within subjects. Br Med J. 310:446. 37. Dunger DB, Edge JA, Pal R, et al. 1991 Impact of increased growth hormone secretion on carbohydrate metabolism in adolescents with diabetes. Acta Paediatr Scand. 377(Suppl):69 –77. 38. Rohner JF, Cusin I, Sainsbury A, Zakrzewska KE, Jeanrenaud B. 1996 The loop system between neuropeptide Y and leptin in normal and obese rodents. Horm Metab Res. 28:642– 648. 39. Campfield LA, Smith FJ, Burn P. 1996 The OB protein (leptin) pathway–a link between adipose tissue mass and central neural networks. Horm Metab Res. 28:619 – 632. 40. Caro JF, Sinha MK, Kolaczynski JW, Zhang PL, Considine RV. 1996 Leptin: the tale of an obesity gene. Diabetes. 45:1455–1462. 41. Kolaczynski JW, Nyce MR, Considine RV, et al. 1996 Acute and chronic effects of insulin on leptin production in humans: studies in vivo and in vitro. Diabetes. 45:699 –701. 42. Halaas JL, Boozer C, Blair WJ, et al. 1997 Physiological response to long-term peripheral and central leptin infusion in lean and obese mice. Proc Natl Acad Sci USA. 94:8878 – 8883. 43. Clayton PE, Gill MS, Hall CM, et al. 1997 Serum leptin through childhood and adolescence. Clin Endocrinol (Oxf). 46:727–733. 44. Garcia MR, Andrade MA, Rios M, et al. 1997 Serum leptin levels in normal children: relationship to age, gender, body mass index, pituitary-gonadal hormones, and pubertal stage. J Clin Endocrinol Metab. 82:2849 –2855. 45. Carlsson B, Ankarberg C, Rosberg S, et al. 1997 Serum leptin concentrations in relation to pubertal development. Arch Dis Child. 77:396 – 400. 46. Blum WF, Englaro P, Hanitsch S, et al. 1997 Plasma leptin levels in healthy children and adolescents: dependence on body mass index, body fat mass, gender, pubertal stage, and testosterone. J Clin Endocrinol Metab. 82:2904 –2910. 47. Shimizu H, Shimomura Y, Nakanishi Y, et al. 1997 Estrogen increases in vivo leptin production in rats and human subjects. J Endocrinol. 154:285–292. 48. Bray GA. 1997 Obesity and reproduction. Hum Reprod. 12(Suppl 1):26 –32. 49. Wabitsch M, Blum WF, Muche R, et al. 1997 Contribution of androgens to the gender difference in leptin production in obese children and adolescents. J Clin Invest. 100:808 – 813. 50. Palmert MR, Radovick S, Boepple PA. 1998 The impact of reversible gonadal sex steroid suppression on serum leptin concentrations in children with central precocious puberty. J Clin Endocrinol Metab. 83:1091–1096.