- Email: [email protected]
Insulin resistance in healthy prepubertal twins
INSULIN RESISTANCE IN HEALTHY PREPUBERTAL TWINS CRAIG A. JEFFERIES, FRACP, PAUL L. HOFMAN, FRACP, HANS KNOBLAUCH, MD, FRIEDRICH C. LUFT, MD, ELIZABETH M. ROBINSON, MS, AND WAYNE S. CUTFIELD, FRACP
Objectives To evaluate insulin sensitivity (SI) in prepubertal twins and to examine the relation to reduced birth weight, prematurity, and peroxisome proliferator-activated receptor-g (PPARg) polymorphism. Study design Fifty twins (birth weight SDS, 0.7 ± 0.2; gestation, 33.5 ± 0.5 weeks; and body mass index SDS, 0.04 ± 0.2) were studied at 8.2 ± 0.3 years. SI was measured by Bergman’s minimal model from a 90 minutes frequently sampled intravenous glucose test. Twenty control children (height SDS, 1.7 ± 0.3; birth weight SDS, 0.3 ± 0.2; and gestation of 39.2 ± 0.7 weeks) were also evaluated at 7.0 ± 0.4 years. The PPARg T-variant polymorphism was evaluated in 41 twins. Values are expressed as mean ± SEM, or 95% confidence intervals. Results SI was reduced in twins compared with control subjects, (12.7 [11-15] versus 23.0 [16.8-31.4] 104 min1 mU/mL, respectively, P = .005). The reduction in SI was independent of prematurity and birth weight and zygosity (P<.0001). There was no difference in SI, even in twin pairs with >20% difference in birth weight (P = .9). The PPARg heterozygote T-variant polymorphism was present in 7 of 41, with a further reduction in SI (P = .03) and a later gestation (P = .03). These twins also had increased fat mass (P = .02) but with similar fat free mass (P = .14). Conclusions Twin children, independent of prematurity or birth weight, had a marked reduction in SI. To use twins as a model to study the fetal origins of adult diseases for glucose homeostasis is not valid. (J Pediatr 2004;144:608-13)
rogressive reduction in birth weight has been shown to be associated with type 2 diabetes mellitus (T2DM), hypertension, dyslipidemia, coronary artery disease, and Syndrome X in adult life.1,2 Because twins have increased rates of low birth weight, these diseases would be expected to occur more commonly in twins. Adult twins have high rates of impaired glucose tolerance or T2DM; however, there is conflicting evidence as to the effect of birth weight in twins and T2DM.3-5 Although Poulsen et al5 linked low birth weight and T2DM in adult twins, the Birmingham twin study6 showed no such association in middle-aged twins. Insulin resistance in first-degree relatives with T2DM has been shown to be an early and key predictor for the later development of T2DM. Insulin resistance is observed in all the adult diseases associated with low birth weight.1,7,8 Healthy prepubertal children born prematurely or small for gestational age (SGA) have been shown to be insulin-resistant, indicating that insulin resistance exists decades before the onset of these adult diseases.9-11 Recently, the phenotype of being a dizygotic twin has been linked to a specific peroxisome proliferator-activated receptor gamma (PPARc) heterozygote C-to-T polymorphism, associated with reduced survival.12 PPARc is a key regulator of adipocyte function and an intracellular regulator of insulin action. Major PPARc mutations have resulted in severe insulin resistance, hypertension, and diabetes.13 Although twins appear to be an ideal group to examine the subsequent effects of SGA and prematurity on insulin sensitivity (SI), a comparison to singletons is essential to
P
BMI DZ FSIGT MZ PPARc
608
Body mass index Dizygotic Frequently sampled intravenous glucose test Monozygotic Peroxisome proliferator-activated receptor-c
SGA SI T2DM
Small for gestational age Insulin sensitivity Type 2 diabetes mellitus
See editorial, p 567.
From Liggins Institute, Health Research Council Biostatistics Unit, Department of Community Health, University of Auckland, Auckland, New Zealand; and Medical Faculty of the Charite´, Humboldt University of Berlin, Max Delbrueck Center for Molecular Medicine, Berlin, Germany. Supported by grants from the New Zealand Lottery Board, National Heart Foundation, Joan Mary Reynolds Trust, and Novo Nordisk. Submitted for publication June 5, 2003; last revision received Nov 20, 2003; accepted Jan 28, 2004. Reprint requests: Wayne S. Cutfield, FRACP, Liggins Institute, University of Auckland, Private Bag 92019, Auckland, New Zealand. E-mail: w.cutfield @auckland.ac.nz. 0022-3476/$ - see front matter Copyright ª 2004 Elsevier Inc. All rights reserved. 10.1016/j.jpeds.2004.01.059
ensure that they are representative of the general population. An effect of being born a twin may influence postnatal insulin sensitivity independent of birth weight and prematurity, obscuring these relations. A singleton group has not been previously used in twin studies examining the association between birth weight and later diseases despite the variable results obtained with this paradigm. The aims of this study were to compare SI in prepubertal twin children with a group of short normal control and to examine the influence of reduced birth weight, prematurity, and the presence of the PPARc T-variant polymorphism on SI in twins.
METHODS Healthy, prepubertal monozygotic (MZ) and dizygotic (DZ) twin subjects were recruited either by random mail-out to the parents of children living in the Auckland region, who were members of the New Zealand Multiple Birth Association, or by personal contact. Zygosity was determined by parental reporting, which is highly accurate compared with DNA analysis.14 The control group subjects were normal birth weight (birth weight >10th percentile), term (37-42 weeks’ gestation), singleton children. These children were either being evaluated for short stature or were healthy children of normal stature. Height in healthy prepubertal children does not influence S.9,15 All had normal growth hormone response to clonidine stimulation (defined as growth hormone peak level >7 lg/L) and normal height velocity. Short control subjects had a family history of familial short stature or constitutional growth delay. Exclusion criteria included a firstdegree relative with T2DM or hypertension, a chromosomal or syndromic diagnosis, documented intrauterine infection, presence of type 1 diabetes mellitus antibodies (islet cell, glutamic acid decarboxylase, insulin autoantibodies, and IA2), any medication known to influence SI, obesity (defined as body mass index [BMI] SDS >+2 for age), chronic illness, or extreme fitness.15-18 SGA was defined as birth weight <10th percentile for gestational age and prematurity as <36 weeks.19 Birth weight and height were converted into SD scores to correct for gestational age and sex. BMI was used as a measure of relative obesity and expressed as BMI SDS.20 Free fat mass and fat percentage were calculated from the bioelectrical impedance data as previously published.21 The North Health Ethics Committee provided approval for the study, and signed consent was obtained by all subjects and parents.
Study Methods After an overnight fast, subjects were evaluated between 8 and 9 AM in the Children’s Research Centre, Starship Children’s Hospital, with a 90-minute frequently sampled intravenous glucose test modified for children, as previously validated.9,16 Body composition was measured by foot-to-foot bioelectrical impedance (Tanita/Stellar Innovations Inc, Palos Verdes, Calif ). Each blood sample was immediately collected in lithium tubes on ice, then immediately centrifuged; plasma Insulin Resistance in Healthy Prepubertal Twins
was separated on completion of the frequently sampled intravenous glucose test (FSIGT). Plasma samples for glucose were kept on ice and analyzed the same day; samples for insulin were frozen for later analysis. The data were used in Bergman’s minimal model to measure SI, acute insulin response, the glucose effectiveness, and glucose disposal coefficient.9 SI reflects the ability of insulin to increase net glucose disposal. The acute insulin response is an estimation of the first-phase insulin response and is calculated as the total insulin response in the first 10 minutes after the dextrose bolus, corrected for the baseline insulin value. Sg is a measure of glucose to enhance its own disposal and suppress glucose production at basal insulin levels. The glucose disposal coefficient was calculated from the slope of the natural logarithm of the glucose concentration between 10 and 19 minutes.22 DNA was collected, and the C- to T-substitution in exon 6 of PPARc23 was genotyped with the use of an oligonucleotide ligation–based assay.24
Assays Plasma glucose was measured with the use of a Hitachi 911 automated random access analyzer (Hitachi, Tokyo, Japan), with an interassay coefficient variation of 1.2%.25 Insulin was determined by Abbott’s IMX microparticle enzyme immunoassay, with an interassay coefficient variation of <5% (Abbott Laboratories, Abbott Park, Ill). Insulin autoantibodies were measured by a competitive radioimmunoassay, and islet cell antibodies were measured by indirect immunoflourescence.26
Statistical Analysis A mixed model with the statistical package SAS (v. 8, SAS Institute, Cary, NC) for personal computers was used to investigate the differences in the parameters of glucose regulation within the twin pairs and between the twin and control groups. The model included birth weight, gestational age, sex, BMI, age, and twin pair as a random effect to evaluate SI. Mixed models were then used to estimate the component of variance in SI of twin pairs that was attributable to zygosity; zygosity was treated as a fixed effect, with twins nested as a random effect. All t tests were 2-tailed, with a P value of <.05 considered significant.
RESULTS Fifty twin children were studied, of whom 43 were white, 5 were New Zealand Maori, and 2 were New Zealand– born Indian (Table I). Birth weight SDS and BMI SDS were similar; as expected, the twins were more premature (14 of 50 born <36 weeks). The control group was younger and shorter than the twins, though neither of these factors has previously been shown to influence SI in the 5- to 12-year age range.16 In particular, stature alone has never been shown to influence SI in prepubertal children.9,16 There were 18 MZ and 32 DZ twins, with no differences in either demographic (age, sex, birth weight SDS, or gestational age) or anthropometric variables (height SDS, midparental height SDS, or BMI 609
Table I. Clinical characteristics of the twin and control groups
Age Sex (M:F) Birth weight SDS Gestational age (wk) BMI SDS Height SDS
Twins
Control subjects
8.2 ± 0.3 27:23 0.7 ± 0.2 33.5 ± 0.5 0.1 ± 0.2 0.2 ± 0.2
7.0 ± 0.4* 14:5 0.3 ± 0.2 39.2 ± 0.7y 0.5 ± 0.3 1.7 ± 0.3y
*P = .03 yP < .001.
SDS) between these zygosity groups. Two sets of triplets were also included, one with three DZ the other with two MZ and one DZ subjects. The triplets did not differ from either the DZ or MZ twins for clinical measures, anthropometric data, or any parameter of insulin action. The 50 twin children therefore represented 22 twin pairs and 2 sets of triplets. To ascertain whether twins were representative of singletons for glucose homeostasis, the twins were compared with the singleton controls. During the FSIGT, the twin and control subjects had very similar glucose profiles; as shown in Figure 1, A, there was no difference in fasting insulin levels (4.03 ± 0.02 vs 4.2 ± 0.2 mU/L for twins and control subjects, respectively, P = not significant). In marked contrast, the twins had 3-fold higher insulin levels during the FSIGT than control subjects, as shown in Figure 1, B. Accordingly, SI was markedly reduced in twins compared with control subjects (P = .005), though it was compensated for by a predicted, >3-fold increase in the acute insulin response to maintain euglycemia (see Table II). By model analysis, the reduction in SI in twins compared with singletons was independent of gestational age, birth weight, zygosity, age, sex, and BMI (P<.0001). There was no difference between the twin and control groups for either glucose effectiveness or the glucose disposal coefficient, and no twin had an abnormality of glucose tolerance (defined as a glucose disposal coefficient of <1).31 The well-established hyperbolic relation between SI and acute insulin response was demonstrated in twins and control subjects (SI vs 1/acute insulin response: r2 = 0.83, P< .0001).13 Thus, the twins had an isolated reduction in SI. The correlation between SI and fasting insulin in twins was similar to previous studies (r2 = 0.43, P<.0001).20 Sex, maternal smoking, or antenatal glucocorticoid treatment was not associated with SI in twins. Birth weight and SI were analyzed in twins to determine whether birth weight influenced SI, as previously demonstrated in SGA children. When analyzed separately, there was no influence of SI on birth weight in twins (P = .5). When further assessed by model analysis, SI was neither associated with birth weight (P = .08) nor birth weight SDS (P = .14) as shown in Figure 2. In twins, there was also no association between SI and gestation (P = .09). We then examined the effect of birth weight on SI within twin pairs, hypothesizing that the twins born smaller 610
Jefferies et al
Fig 1. A, Plasma glucose values during a frequently sampled intravenous glucose test in twins (closed squares) and control subjects (open squares), expressed as mean ± SEM. B, Plasma insulin values during a frequently sampled intravenous glucose test in twins (closed squares) and control subjects (open squares) (mean ± SEM).
would have a lower SI. Despite a difference in birth weight (P = .007), there was no difference in SI between the twins born heavier and lighter (SI = 13.5 ± 0.9 vs 12.7 ± 0.9 104/ min1 (lU/mL), P = .6). Furthermore, even when those twin pairs who had a >20% difference in birth weight were studied, there was still no difference in SI (P = .9). To determine the genetic influence on SI, the genetically identical MZ twins were compared with the DZ twins. There was no difference in SI between MZ and DZ twins (P = .7), which provides evidence against a major genetic influence on reducing SI. The influence of the T-variant PPARc polymorphism on SI was examined, as shown in Table III. The polymorphism was found in 7 of 41 twins evaluated, and, although not more common in the DZ twins, there was an association with an The Journal of Pediatrics May 2004
Table III. Comparison of clinical characteristics and insulin regulation parameters with regard to the presence of the T-variant PPARg polymorphism C-to-T No polymorphism polymorphism P value
Fig 2. Relation between birth weight SDS and insulin sensitivity in twin (closed squares) and control subjects (open squares) (mean± SEM).
Age (mo) Birth weight (g) Gestation (wk) Height SDS BMI Fat-free mass (kg)* Fat mass (kg)* Percentage body fat* Insulin sensitivity Acute insulin response
8.5 ± 0.4 2629 ± 252 36.9 ± 0.6 0.6 ± 0.2 17.7 ± 0.5 21.6 ± 1.4 10.0 ± 1.0 31.6 ± 1.1
8.2 ± 0.3 1912 ± 105 32.9 ± 0.6 0.1 ± 0.2 15.9 ± 0.3 19.0 ± 0.7 7.0 ± 0.5 26.4 ± 1.2
.2 .01 .003 .08 .02 .14 .02 .06
10.0 ± 1.8 497 ± 0.07
14.1 ± 0.7 314 ± 21
.03 .007
*Derived from foot-to-foot bioelectrical impedance measurement.26
Table II. Glucose and insulin regulation parameters in the twin and control groups Units Insulin sensitivity 104/min1 index (lU/mL) Acute insulin (mU/L) response Glucose 102/min1 effectiveness Glucose disposal 102(mg/d) min1
Twins
Control subjects
12.7 (11.0–14.8) 23.0 (16.8–31.4)* 301 (247–369)
147 (96–224)y
2.6 (2.3–2.9)
2.4 (2.1–2.6)
2.9 (2.7–3.2)
2.1 (1.5–2.8)
Values are expressed as mean ± SEM. *P = .005. yP = .01.
even further reduction in SI (P = .03). Compared with the twins without the T-variant PPARc polymorphism, these 7 children were also born at a later gestation (P = .03) and were heavier as the result of increased fat mass when tested (P = .02).
DISCUSSION We have demonstrated that twin children were insulinresistant, with a marked reduction in SI in twins compared with singleton control subjects that was not due to prematurity nor growth retardation. The twin children also did not have any of the known correlates of insulin resistance that include acanthosis nigricans, race, obesity, nor family history of T2DM. Insulin resistance has been shown to exist in subjects before the onset of both T2DM and hypertension.28,29 In a 25-year longitudinal study, Martin et al7 demonstrated that an isolated reduction in SI had a 40% cumulative risk for the development of T2DM in those with a first-degree relative with T2DM. Similarly, adult twins have a high incidence of Insulin Resistance in Healthy Prepubertal Twins
T2DM. Adult twin studies have shown a prevalence of impaired glucose tolerance or T2DM of 35%, which was considerably higher than the background population prevalence of 10%.30 The reduced SI in twin children confirms these observations that twins are at risk of development of T2DM in later life and demonstrates that abnormalities in SI are present at a very early age. With puberty and later metabolic stresses such as truncal obesity, inactivity, and a high fat diet, insulin resistance would be amplified, thus increasing the risk of T2DM.16,31 There have been several mechanisms postulated to explain the relation between SGA and the development of insulin resistance that may still be valid in twins. Fetal or early postnatal undernutrition is thought to be a primary event that leads to a programmed and persistent reduction in SI in childhood and ultimately to T2DM in later adulthood.8,32 The fetal salvage hypothesis has been proposed to explain the link between undernutrition and insulin resistance. After fetal or neonatal undernutrition, insulin resistance occurs, allowing redistribution of limited nutrients from insulin-dependent tissues such as muscle and fat to the brain and heart, organs essential for fetal survival. The insulin resistance increases the survival chance of the affected fetus. However, this in utero adaptation is maladaptive after birth, as the persistent insulin resistance results in hyperinsulinism, chronic b-cell stress, and eventual b-cell failure and T2DM. Our study indicates that undernutrition does not appear to be the initiating event for insulin resistance in twins. Despite this, the in utero protective benefits of insulin resistance proposed for SGA and prematurity would equally apply to twins.9,33 Insulin resistance in twin fetuses would improve their utilization of the available nutrient resources and maximize survival. In our study, all twins were insulin-resistant, suggesting that large differences in SI within twin pairs may not be compatible with survival of both twins in utero. The reduced SI in twins was not principally due to a combination of prematurity and/or reduced birth weight; 611
rather, it was dependent on being a twin. This suggests that glucose homeostasis is fundamentally different in twins. These findings are in direct contrast to observations made in SGA or premature children, in whom a direct correlation is found between birth weight or prematurity and SI.9 In addition, SI was reduced in twins even when compared with the insulinresistant SGA or premature children previously described.9,10 Given the lack of effect of birth weight and prematurity on SI and the magnitude of the reduction of SI seen in twins, it is likely that other mechanisms are involved. Lack of influence of zygosity on SI is consistent with very similar prevalences of impaired glucose tolerance and T2DM for MZ and DZ adult twins from the Danish Twin Registry.30 To study common diseases, twins must be representative of the general population for the disease being studied.36,37 This study indicates that for glucose homeostasis, twins are different than singletons, and therefore twins should not be used in studying diseases that have reduced insulin sensitivity involved in their pathogenesis. Twins to date have been proposed as a useful model to evaluate the effect of reduced birth weight on diseases in adult life, although in twin children the relation with blood pressure is debated.34,35 We were unable to find any relation between birth weight or gestation and SI in twins, challenging the validity of twins in examination of the influence of birth weight on later glucose homeostasis. Recently, genetic regulation of DZ twinning has been examined.12,37 In a large German cohort of adult twins, heterozygosity for the C to T PPARc polymorphism was underrepresented in dizygotic twins. Furthermore, the frequency distribution of the PPARc polymorphism did not conform to Hardy-Weinberg equilibrium. Conversely, in parents of dizygotic twins, Hardy-Weinberg equilibrium was maintained.12 We did not demonstrate deviations from Hardy-Weinberg equilibrium for this polymorphism, although our twin group was too small to accurately determine the frequency of this polymorphism. Busjahn et al12 postulated that heterozygosity for this polymorphism is associated with reduced survival of DZ twins in utero, which appears to conflict with our finding of prolonged gestation in twins, suggesting improved twin survival, especially before improvements in obstetric and neonatal care. To reconcile these conflicting findings regarding twin survival with the C to T PPARc polymorphism, we propose that the polymorphism is associated with early pregnancy fetal loss, perhaps at the implantation stage. PPARc has a role in successful trophoblast formation in the placenta.38 However, in those twins with the polymorphism, if survival through early pregnancy is achieved, the fetus is more likely to reach a longer gestation, enhancing late pregnancy survival. In this study, we observed an association between the PPARc polymorphism and a further reduction in SI and also greater weight in childhood.12 PPARc is a nuclear hormone receptor and is a key regulator of adipocyte function.23 Severe but rare abnormalities of PPARc have resulted in marked insulin resistance and hypertension.13 However, it is important to note that the major reduction in SI in all twins is not 612
Jefferies et al
due to the PPARc polymorphism. Nevertheless, the effect of the polymorphism in further reducing SI and increasing adiposity may contribute to the later development of T2DM. The long period between impaired SI in childhood and development of T2DM in adult life provides a window to implement intervention strategies to delay or prevent the risk of subsequent disease. The use of twins to examine fetal programming of glucose regulation parameters to explain the fetal origins of adult disease would not appear valid. We thank the Joan Mary Reynolds Trust and Novo Nordisk for their financial support of Dr C. A. Jefferies and Pharmacia Corporation for the generous donation of Orinase.
REFERENCES 1. Hales CN, Barker DJ, Clark PM. Foetal and infant growth and impaired glucose tolerance at age 64. BMJ 1991;303:1019-22. 2. Carlson S, Persson P-G, Alvarsson M, Efendic S, Norman A, Svanstrom L, et al. Low birth weight, family history of diabetes, and glucose intolerance in Swedish middle aged men. Diabetes Care 1999;22:1043-7. 3. Medici F, Hawa M, Ianari A, Pyke DA, Leslie RDG. Concordance rate for type 2 diabetes mellitus in monozygotic twins: actuarial analysis. Diabetologia 1999;42:146-50. 4. Barnett AH, Eff C, Leslie RDG, Pyke DA. Diabetes in identical twins. Diabetologia 1981;20:87-93. 5. Poulsen P, Vaag A, Kyvik K, Jensen D, Beck-Nielsen H. Low birth weight is associated with NIDDM in discordant monozygotic and dizygotic twin pairs. Diabetologia 1997;40:439-46. 6. Baird J, Osmond C, MacGregor A, Sneider H, Hales CN, Phillips DIW. Testing the fetal origins hypothesis in twins: the Birmingham twin study. Diabetologia 2001;44:33-9. 7. Martin BC, Warram JH, Krolwski AJ, Bergman RN, Soeldner JS, Khan CR. Role of glucose and insulin resistance in the development of type 2 diabetes mellitus: results of a 25-year follow up study. Lancet 1992;340:925-9. 8. Barker DJP. Mothers, babies and health in later life. 2nd ed. Edinburgh, Scotland: Churchill Livingstone; 1998. 9. Hofman PL, Cutfield WS, Robinson EM, Bergman RN, Menon RK, Sperling MA, et al. Insulin resistance in short children with intrauterine growth retardation. J Clin Endocrinol Metab 1997;82:402-6. 10. Hofman PL, Cutfield WS, Robinson EM. Premature children are insulin resistant [Abstract]. Sydney, Australia: ICE; 2000. 11. Wright NP, Gibson AT, Carney S, Wales JK. Does being born prematurely have a long term effect on blood pressure and glucose tolerance [abstract]? Montreal, Canada: ESPE Lawson Wilkins; 2001. 12. Busjahn A, Knoblauch H, Faulhaber H, Luft FC, Mu¨ller-Myhsok BC. A region on chromosome 3 is linked to dizygotic twinning. Nat Genet 2000;26:398-9. 13. Barroso J, Gurnell M, Crowley VEF, Agnostini M, Schwabe JW, Soos MA, et al. Dominant negative mutations in human PPARy associated with severe insulin resistance, diabetes mellitus and hypertension [letter]. Nature 1999;402:880-3. 14. Rietveld MJ, van Der Valk JC, Bongers IL, Stroet TM, Slagboom PE, Boomsma DI. Zygosity diagnosis in young twins by parental report. Twin Res 2000;3:134-41. 15. Tanner JR, Whitehouse RH, Takaishi M. Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965, I. Arch Dis Child 1966;41:454-471 and 613-35. 16. Cutfield WS, Bergman RN, Menon RK, Sperling MA. The modified minimal model: application to measurement of insulin sensitivity in children. J Clin Endocrinol Metab 1990;70:1644-50. 17. Atkinson MA, MacLaren NK, Riley WJ, Winter WE, Fisk DD, Spillar RP. Are insulin autoantibodies markers for insulin dependent diabetes mellitus? Diabetes 1986;35:894-8. 18. Gorsuch AN, Spencer KM, Lister J, McNally JM, Dean BM, Bottazzo GF, et al. Evidence for a long prediabetic period in type 1 (insulin-dependent) diabetes mellitus. Lancet 1981;2:1363-5.
The Journal of Pediatrics May 2004
19. Guaran R, Wein R, Sheedy M, Walstab J, Beischer N. Update of growth percentiles for infants born in an Australian population. Aust N Z J Obstet Gynecol 1995;34:39-50. 20. Hammer LD, Kraemer HC, Wilson DM, Ritter PL, Dornbusch SM. Standardized percentile curves of body-mass index for children and adolescents. Am J Dis Child 1991;145:259-63. 21. Tyrrell VJ, Richards G, Hofman P, Gilles GF, Robinson E, Cutfield WS. Foot to foot bioelectrical impedance analysis: a valuable tool for the measurement of body composition in children. Int J Obes 2001;25:273-8. 22. Lee A, Ader M, Bray GA, Bergman RN. Diurnal variation in glucose tolerance: cyclic suppression of insulin action and insulin secretion in normal weight, but not obese, subjects. Diabetes 1992;41:742-9. 23. Meirhaeghe A, Fajas L, Helbecque N, Cottel D, Lebel P, Dallongeville J, et al. A genetic polymorphism of the peroxisome proliferator activated receptor gamma gene influences plasma leptin levels in obese humans. Hum Mol Med 1998;26:519-25. 24. Baron H, Fung S, Aydin A, Bahring S, Luft FC, Schuster H. Oligonucleotide ligation assay (OLA) for the diagnosis of familial hypercholesterolemia. Nat Biotechnol 1996;14:1279-82. 25. Trinder P. Determination of blood glucose using an oxidase-peroxidase system with a non-carcinogenic chromogen. J Clin Pathol 1995;22:158-61. 26. Pilcher C, Elliot R. A sensitive and reproducible method for the assay of human islet cell antibodies. J Immunol Methods 1990;129:111-7. 27. Bergman RN. Lilly Lecture 1989: Towards physiological understanding of glucose tolerance. Diabetes 1989;38:1512-27. 28. Osei K, Cottrell DA, Orabella MM. Insulin sensitivity, glucose effectiveness and body fat pattern in non diabetic offspring of patients with NIDDM. Diabetes Care 1991;14:890-6.
Insulin Resistance in Healthy Prepubertal Twins
29. Grunfeld B, Balzareti M, Romo M, Gimenez M, Gutman R. Hyperinsulinaemia in normotensive offspring of hypertensive parents. Hypertension 1994;23:112-5. 30. Poulsen P, Vaag A, Beck-Nielsen H. Does zygosity influence the metabolic profile of twins? A population based cross sectional study. Br Med J 1999;319:151-4. 31. Cook JS, Hoffman RP, Stene MA, Hansen JR. Effects of maturational stage on insulin sensitivity during puberty. J Clin Endocrinol Metab 1993; 77:725-30. 32. Barker DJP. Early growth and cardiovascular disease. Arch Dis Child 1999;80:305-7. 33. Simmons RA, Templeton LJ, Gertz SJ. Intrauterine growth retardation leads to the development of type 2 diabetes in the rat. Diabetes 2001;5: 2279-86. 34. Dwyer T, Blizzard L, Morley R, Ponsonby A-L. Within pair association between birth weight and blood pressure at age 8 in twins from a cohort study. Br Med J 1999;319:1325-9. 35. Zhang J, Brenner RA, Klebanoff MA. Differences in birth weight and blood pressure at age 7 years among twins. Am J Epidemiol 2001;153:779-82. 36. Hrubec Z, Robinette CD. Study of human twins in medical research. N Engl J Med 1984;10:435-41. 37. Kyvik KO. Twin research: nature versus nurture in common diseases. In: Day INM, Humphries SE, editors. Genetics of common diseases. Oxford: BIOS Scientific Publishers LTD; 1997. 38. Tarrade A, Schoonjans K, Guibourdenche J, Bidart JM, Vidaud M, Auwerx J, et al. PPAR gamma/RXR alpha heterodimers are involved in human CG beta synthesis and human trophoblast differentiation. Endocrinology 2001;142:4504-14.
613
Objectives To evaluate insulin sensitivity (SI) in prepubertal twins and to examine the relation to reduced birth weight, prematurity, and peroxisome proliferator-activated receptor-g (PPARg) polymorphism. Study design Fifty twins (birth weight SDS, 0.7 ± 0.2; gestation, 33.5 ± 0.5 weeks; and body mass index SDS, 0.04 ± 0.2) were studied at 8.2 ± 0.3 years. SI was measured by Bergman’s minimal model from a 90 minutes frequently sampled intravenous glucose test. Twenty control children (height SDS, 1.7 ± 0.3; birth weight SDS, 0.3 ± 0.2; and gestation of 39.2 ± 0.7 weeks) were also evaluated at 7.0 ± 0.4 years. The PPARg T-variant polymorphism was evaluated in 41 twins. Values are expressed as mean ± SEM, or 95% confidence intervals. Results SI was reduced in twins compared with control subjects, (12.7 [11-15] versus 23.0 [16.8-31.4] 104 min1 mU/mL, respectively, P = .005). The reduction in SI was independent of prematurity and birth weight and zygosity (P<.0001). There was no difference in SI, even in twin pairs with >20% difference in birth weight (P = .9). The PPARg heterozygote T-variant polymorphism was present in 7 of 41, with a further reduction in SI (P = .03) and a later gestation (P = .03). These twins also had increased fat mass (P = .02) but with similar fat free mass (P = .14). Conclusions Twin children, independent of prematurity or birth weight, had a marked reduction in SI. To use twins as a model to study the fetal origins of adult diseases for glucose homeostasis is not valid. (J Pediatr 2004;144:608-13)
rogressive reduction in birth weight has been shown to be associated with type 2 diabetes mellitus (T2DM), hypertension, dyslipidemia, coronary artery disease, and Syndrome X in adult life.1,2 Because twins have increased rates of low birth weight, these diseases would be expected to occur more commonly in twins. Adult twins have high rates of impaired glucose tolerance or T2DM; however, there is conflicting evidence as to the effect of birth weight in twins and T2DM.3-5 Although Poulsen et al5 linked low birth weight and T2DM in adult twins, the Birmingham twin study6 showed no such association in middle-aged twins. Insulin resistance in first-degree relatives with T2DM has been shown to be an early and key predictor for the later development of T2DM. Insulin resistance is observed in all the adult diseases associated with low birth weight.1,7,8 Healthy prepubertal children born prematurely or small for gestational age (SGA) have been shown to be insulin-resistant, indicating that insulin resistance exists decades before the onset of these adult diseases.9-11 Recently, the phenotype of being a dizygotic twin has been linked to a specific peroxisome proliferator-activated receptor gamma (PPARc) heterozygote C-to-T polymorphism, associated with reduced survival.12 PPARc is a key regulator of adipocyte function and an intracellular regulator of insulin action. Major PPARc mutations have resulted in severe insulin resistance, hypertension, and diabetes.13 Although twins appear to be an ideal group to examine the subsequent effects of SGA and prematurity on insulin sensitivity (SI), a comparison to singletons is essential to
P
BMI DZ FSIGT MZ PPARc
608
Body mass index Dizygotic Frequently sampled intravenous glucose test Monozygotic Peroxisome proliferator-activated receptor-c
SGA SI T2DM
Small for gestational age Insulin sensitivity Type 2 diabetes mellitus
See editorial, p 567.
From Liggins Institute, Health Research Council Biostatistics Unit, Department of Community Health, University of Auckland, Auckland, New Zealand; and Medical Faculty of the Charite´, Humboldt University of Berlin, Max Delbrueck Center for Molecular Medicine, Berlin, Germany. Supported by grants from the New Zealand Lottery Board, National Heart Foundation, Joan Mary Reynolds Trust, and Novo Nordisk. Submitted for publication June 5, 2003; last revision received Nov 20, 2003; accepted Jan 28, 2004. Reprint requests: Wayne S. Cutfield, FRACP, Liggins Institute, University of Auckland, Private Bag 92019, Auckland, New Zealand. E-mail: w.cutfield @auckland.ac.nz. 0022-3476/$ - see front matter Copyright ª 2004 Elsevier Inc. All rights reserved. 10.1016/j.jpeds.2004.01.059
ensure that they are representative of the general population. An effect of being born a twin may influence postnatal insulin sensitivity independent of birth weight and prematurity, obscuring these relations. A singleton group has not been previously used in twin studies examining the association between birth weight and later diseases despite the variable results obtained with this paradigm. The aims of this study were to compare SI in prepubertal twin children with a group of short normal control and to examine the influence of reduced birth weight, prematurity, and the presence of the PPARc T-variant polymorphism on SI in twins.
METHODS Healthy, prepubertal monozygotic (MZ) and dizygotic (DZ) twin subjects were recruited either by random mail-out to the parents of children living in the Auckland region, who were members of the New Zealand Multiple Birth Association, or by personal contact. Zygosity was determined by parental reporting, which is highly accurate compared with DNA analysis.14 The control group subjects were normal birth weight (birth weight >10th percentile), term (37-42 weeks’ gestation), singleton children. These children were either being evaluated for short stature or were healthy children of normal stature. Height in healthy prepubertal children does not influence S.9,15 All had normal growth hormone response to clonidine stimulation (defined as growth hormone peak level >7 lg/L) and normal height velocity. Short control subjects had a family history of familial short stature or constitutional growth delay. Exclusion criteria included a firstdegree relative with T2DM or hypertension, a chromosomal or syndromic diagnosis, documented intrauterine infection, presence of type 1 diabetes mellitus antibodies (islet cell, glutamic acid decarboxylase, insulin autoantibodies, and IA2), any medication known to influence SI, obesity (defined as body mass index [BMI] SDS >+2 for age), chronic illness, or extreme fitness.15-18 SGA was defined as birth weight <10th percentile for gestational age and prematurity as <36 weeks.19 Birth weight and height were converted into SD scores to correct for gestational age and sex. BMI was used as a measure of relative obesity and expressed as BMI SDS.20 Free fat mass and fat percentage were calculated from the bioelectrical impedance data as previously published.21 The North Health Ethics Committee provided approval for the study, and signed consent was obtained by all subjects and parents.
Study Methods After an overnight fast, subjects were evaluated between 8 and 9 AM in the Children’s Research Centre, Starship Children’s Hospital, with a 90-minute frequently sampled intravenous glucose test modified for children, as previously validated.9,16 Body composition was measured by foot-to-foot bioelectrical impedance (Tanita/Stellar Innovations Inc, Palos Verdes, Calif ). Each blood sample was immediately collected in lithium tubes on ice, then immediately centrifuged; plasma Insulin Resistance in Healthy Prepubertal Twins
was separated on completion of the frequently sampled intravenous glucose test (FSIGT). Plasma samples for glucose were kept on ice and analyzed the same day; samples for insulin were frozen for later analysis. The data were used in Bergman’s minimal model to measure SI, acute insulin response, the glucose effectiveness, and glucose disposal coefficient.9 SI reflects the ability of insulin to increase net glucose disposal. The acute insulin response is an estimation of the first-phase insulin response and is calculated as the total insulin response in the first 10 minutes after the dextrose bolus, corrected for the baseline insulin value. Sg is a measure of glucose to enhance its own disposal and suppress glucose production at basal insulin levels. The glucose disposal coefficient was calculated from the slope of the natural logarithm of the glucose concentration between 10 and 19 minutes.22 DNA was collected, and the C- to T-substitution in exon 6 of PPARc23 was genotyped with the use of an oligonucleotide ligation–based assay.24
Assays Plasma glucose was measured with the use of a Hitachi 911 automated random access analyzer (Hitachi, Tokyo, Japan), with an interassay coefficient variation of 1.2%.25 Insulin was determined by Abbott’s IMX microparticle enzyme immunoassay, with an interassay coefficient variation of <5% (Abbott Laboratories, Abbott Park, Ill). Insulin autoantibodies were measured by a competitive radioimmunoassay, and islet cell antibodies were measured by indirect immunoflourescence.26
Statistical Analysis A mixed model with the statistical package SAS (v. 8, SAS Institute, Cary, NC) for personal computers was used to investigate the differences in the parameters of glucose regulation within the twin pairs and between the twin and control groups. The model included birth weight, gestational age, sex, BMI, age, and twin pair as a random effect to evaluate SI. Mixed models were then used to estimate the component of variance in SI of twin pairs that was attributable to zygosity; zygosity was treated as a fixed effect, with twins nested as a random effect. All t tests were 2-tailed, with a P value of <.05 considered significant.
RESULTS Fifty twin children were studied, of whom 43 were white, 5 were New Zealand Maori, and 2 were New Zealand– born Indian (Table I). Birth weight SDS and BMI SDS were similar; as expected, the twins were more premature (14 of 50 born <36 weeks). The control group was younger and shorter than the twins, though neither of these factors has previously been shown to influence SI in the 5- to 12-year age range.16 In particular, stature alone has never been shown to influence SI in prepubertal children.9,16 There were 18 MZ and 32 DZ twins, with no differences in either demographic (age, sex, birth weight SDS, or gestational age) or anthropometric variables (height SDS, midparental height SDS, or BMI 609
Table I. Clinical characteristics of the twin and control groups
Age Sex (M:F) Birth weight SDS Gestational age (wk) BMI SDS Height SDS
Twins
Control subjects
8.2 ± 0.3 27:23 0.7 ± 0.2 33.5 ± 0.5 0.1 ± 0.2 0.2 ± 0.2
7.0 ± 0.4* 14:5 0.3 ± 0.2 39.2 ± 0.7y 0.5 ± 0.3 1.7 ± 0.3y
*P = .03 yP < .001.
SDS) between these zygosity groups. Two sets of triplets were also included, one with three DZ the other with two MZ and one DZ subjects. The triplets did not differ from either the DZ or MZ twins for clinical measures, anthropometric data, or any parameter of insulin action. The 50 twin children therefore represented 22 twin pairs and 2 sets of triplets. To ascertain whether twins were representative of singletons for glucose homeostasis, the twins were compared with the singleton controls. During the FSIGT, the twin and control subjects had very similar glucose profiles; as shown in Figure 1, A, there was no difference in fasting insulin levels (4.03 ± 0.02 vs 4.2 ± 0.2 mU/L for twins and control subjects, respectively, P = not significant). In marked contrast, the twins had 3-fold higher insulin levels during the FSIGT than control subjects, as shown in Figure 1, B. Accordingly, SI was markedly reduced in twins compared with control subjects (P = .005), though it was compensated for by a predicted, >3-fold increase in the acute insulin response to maintain euglycemia (see Table II). By model analysis, the reduction in SI in twins compared with singletons was independent of gestational age, birth weight, zygosity, age, sex, and BMI (P<.0001). There was no difference between the twin and control groups for either glucose effectiveness or the glucose disposal coefficient, and no twin had an abnormality of glucose tolerance (defined as a glucose disposal coefficient of <1).31 The well-established hyperbolic relation between SI and acute insulin response was demonstrated in twins and control subjects (SI vs 1/acute insulin response: r2 = 0.83, P< .0001).13 Thus, the twins had an isolated reduction in SI. The correlation between SI and fasting insulin in twins was similar to previous studies (r2 = 0.43, P<.0001).20 Sex, maternal smoking, or antenatal glucocorticoid treatment was not associated with SI in twins. Birth weight and SI were analyzed in twins to determine whether birth weight influenced SI, as previously demonstrated in SGA children. When analyzed separately, there was no influence of SI on birth weight in twins (P = .5). When further assessed by model analysis, SI was neither associated with birth weight (P = .08) nor birth weight SDS (P = .14) as shown in Figure 2. In twins, there was also no association between SI and gestation (P = .09). We then examined the effect of birth weight on SI within twin pairs, hypothesizing that the twins born smaller 610
Jefferies et al
Fig 1. A, Plasma glucose values during a frequently sampled intravenous glucose test in twins (closed squares) and control subjects (open squares), expressed as mean ± SEM. B, Plasma insulin values during a frequently sampled intravenous glucose test in twins (closed squares) and control subjects (open squares) (mean ± SEM).
would have a lower SI. Despite a difference in birth weight (P = .007), there was no difference in SI between the twins born heavier and lighter (SI = 13.5 ± 0.9 vs 12.7 ± 0.9 104/ min1 (lU/mL), P = .6). Furthermore, even when those twin pairs who had a >20% difference in birth weight were studied, there was still no difference in SI (P = .9). To determine the genetic influence on SI, the genetically identical MZ twins were compared with the DZ twins. There was no difference in SI between MZ and DZ twins (P = .7), which provides evidence against a major genetic influence on reducing SI. The influence of the T-variant PPARc polymorphism on SI was examined, as shown in Table III. The polymorphism was found in 7 of 41 twins evaluated, and, although not more common in the DZ twins, there was an association with an The Journal of Pediatrics May 2004
Table III. Comparison of clinical characteristics and insulin regulation parameters with regard to the presence of the T-variant PPARg polymorphism C-to-T No polymorphism polymorphism P value
Fig 2. Relation between birth weight SDS and insulin sensitivity in twin (closed squares) and control subjects (open squares) (mean± SEM).
Age (mo) Birth weight (g) Gestation (wk) Height SDS BMI Fat-free mass (kg)* Fat mass (kg)* Percentage body fat* Insulin sensitivity Acute insulin response
8.5 ± 0.4 2629 ± 252 36.9 ± 0.6 0.6 ± 0.2 17.7 ± 0.5 21.6 ± 1.4 10.0 ± 1.0 31.6 ± 1.1
8.2 ± 0.3 1912 ± 105 32.9 ± 0.6 0.1 ± 0.2 15.9 ± 0.3 19.0 ± 0.7 7.0 ± 0.5 26.4 ± 1.2
.2 .01 .003 .08 .02 .14 .02 .06
10.0 ± 1.8 497 ± 0.07
14.1 ± 0.7 314 ± 21
.03 .007
*Derived from foot-to-foot bioelectrical impedance measurement.26
Table II. Glucose and insulin regulation parameters in the twin and control groups Units Insulin sensitivity 104/min1 index (lU/mL) Acute insulin (mU/L) response Glucose 102/min1 effectiveness Glucose disposal 102(mg/d) min1
Twins
Control subjects
12.7 (11.0–14.8) 23.0 (16.8–31.4)* 301 (247–369)
147 (96–224)y
2.6 (2.3–2.9)
2.4 (2.1–2.6)
2.9 (2.7–3.2)
2.1 (1.5–2.8)
Values are expressed as mean ± SEM. *P = .005. yP = .01.
even further reduction in SI (P = .03). Compared with the twins without the T-variant PPARc polymorphism, these 7 children were also born at a later gestation (P = .03) and were heavier as the result of increased fat mass when tested (P = .02).
DISCUSSION We have demonstrated that twin children were insulinresistant, with a marked reduction in SI in twins compared with singleton control subjects that was not due to prematurity nor growth retardation. The twin children also did not have any of the known correlates of insulin resistance that include acanthosis nigricans, race, obesity, nor family history of T2DM. Insulin resistance has been shown to exist in subjects before the onset of both T2DM and hypertension.28,29 In a 25-year longitudinal study, Martin et al7 demonstrated that an isolated reduction in SI had a 40% cumulative risk for the development of T2DM in those with a first-degree relative with T2DM. Similarly, adult twins have a high incidence of Insulin Resistance in Healthy Prepubertal Twins
T2DM. Adult twin studies have shown a prevalence of impaired glucose tolerance or T2DM of 35%, which was considerably higher than the background population prevalence of 10%.30 The reduced SI in twin children confirms these observations that twins are at risk of development of T2DM in later life and demonstrates that abnormalities in SI are present at a very early age. With puberty and later metabolic stresses such as truncal obesity, inactivity, and a high fat diet, insulin resistance would be amplified, thus increasing the risk of T2DM.16,31 There have been several mechanisms postulated to explain the relation between SGA and the development of insulin resistance that may still be valid in twins. Fetal or early postnatal undernutrition is thought to be a primary event that leads to a programmed and persistent reduction in SI in childhood and ultimately to T2DM in later adulthood.8,32 The fetal salvage hypothesis has been proposed to explain the link between undernutrition and insulin resistance. After fetal or neonatal undernutrition, insulin resistance occurs, allowing redistribution of limited nutrients from insulin-dependent tissues such as muscle and fat to the brain and heart, organs essential for fetal survival. The insulin resistance increases the survival chance of the affected fetus. However, this in utero adaptation is maladaptive after birth, as the persistent insulin resistance results in hyperinsulinism, chronic b-cell stress, and eventual b-cell failure and T2DM. Our study indicates that undernutrition does not appear to be the initiating event for insulin resistance in twins. Despite this, the in utero protective benefits of insulin resistance proposed for SGA and prematurity would equally apply to twins.9,33 Insulin resistance in twin fetuses would improve their utilization of the available nutrient resources and maximize survival. In our study, all twins were insulin-resistant, suggesting that large differences in SI within twin pairs may not be compatible with survival of both twins in utero. The reduced SI in twins was not principally due to a combination of prematurity and/or reduced birth weight; 611
rather, it was dependent on being a twin. This suggests that glucose homeostasis is fundamentally different in twins. These findings are in direct contrast to observations made in SGA or premature children, in whom a direct correlation is found between birth weight or prematurity and SI.9 In addition, SI was reduced in twins even when compared with the insulinresistant SGA or premature children previously described.9,10 Given the lack of effect of birth weight and prematurity on SI and the magnitude of the reduction of SI seen in twins, it is likely that other mechanisms are involved. Lack of influence of zygosity on SI is consistent with very similar prevalences of impaired glucose tolerance and T2DM for MZ and DZ adult twins from the Danish Twin Registry.30 To study common diseases, twins must be representative of the general population for the disease being studied.36,37 This study indicates that for glucose homeostasis, twins are different than singletons, and therefore twins should not be used in studying diseases that have reduced insulin sensitivity involved in their pathogenesis. Twins to date have been proposed as a useful model to evaluate the effect of reduced birth weight on diseases in adult life, although in twin children the relation with blood pressure is debated.34,35 We were unable to find any relation between birth weight or gestation and SI in twins, challenging the validity of twins in examination of the influence of birth weight on later glucose homeostasis. Recently, genetic regulation of DZ twinning has been examined.12,37 In a large German cohort of adult twins, heterozygosity for the C to T PPARc polymorphism was underrepresented in dizygotic twins. Furthermore, the frequency distribution of the PPARc polymorphism did not conform to Hardy-Weinberg equilibrium. Conversely, in parents of dizygotic twins, Hardy-Weinberg equilibrium was maintained.12 We did not demonstrate deviations from Hardy-Weinberg equilibrium for this polymorphism, although our twin group was too small to accurately determine the frequency of this polymorphism. Busjahn et al12 postulated that heterozygosity for this polymorphism is associated with reduced survival of DZ twins in utero, which appears to conflict with our finding of prolonged gestation in twins, suggesting improved twin survival, especially before improvements in obstetric and neonatal care. To reconcile these conflicting findings regarding twin survival with the C to T PPARc polymorphism, we propose that the polymorphism is associated with early pregnancy fetal loss, perhaps at the implantation stage. PPARc has a role in successful trophoblast formation in the placenta.38 However, in those twins with the polymorphism, if survival through early pregnancy is achieved, the fetus is more likely to reach a longer gestation, enhancing late pregnancy survival. In this study, we observed an association between the PPARc polymorphism and a further reduction in SI and also greater weight in childhood.12 PPARc is a nuclear hormone receptor and is a key regulator of adipocyte function.23 Severe but rare abnormalities of PPARc have resulted in marked insulin resistance and hypertension.13 However, it is important to note that the major reduction in SI in all twins is not 612
Jefferies et al
due to the PPARc polymorphism. Nevertheless, the effect of the polymorphism in further reducing SI and increasing adiposity may contribute to the later development of T2DM. The long period between impaired SI in childhood and development of T2DM in adult life provides a window to implement intervention strategies to delay or prevent the risk of subsequent disease. The use of twins to examine fetal programming of glucose regulation parameters to explain the fetal origins of adult disease would not appear valid. We thank the Joan Mary Reynolds Trust and Novo Nordisk for their financial support of Dr C. A. Jefferies and Pharmacia Corporation for the generous donation of Orinase.
REFERENCES 1. Hales CN, Barker DJ, Clark PM. Foetal and infant growth and impaired glucose tolerance at age 64. BMJ 1991;303:1019-22. 2. Carlson S, Persson P-G, Alvarsson M, Efendic S, Norman A, Svanstrom L, et al. Low birth weight, family history of diabetes, and glucose intolerance in Swedish middle aged men. Diabetes Care 1999;22:1043-7. 3. Medici F, Hawa M, Ianari A, Pyke DA, Leslie RDG. Concordance rate for type 2 diabetes mellitus in monozygotic twins: actuarial analysis. Diabetologia 1999;42:146-50. 4. Barnett AH, Eff C, Leslie RDG, Pyke DA. Diabetes in identical twins. Diabetologia 1981;20:87-93. 5. Poulsen P, Vaag A, Kyvik K, Jensen D, Beck-Nielsen H. Low birth weight is associated with NIDDM in discordant monozygotic and dizygotic twin pairs. Diabetologia 1997;40:439-46. 6. Baird J, Osmond C, MacGregor A, Sneider H, Hales CN, Phillips DIW. Testing the fetal origins hypothesis in twins: the Birmingham twin study. Diabetologia 2001;44:33-9. 7. Martin BC, Warram JH, Krolwski AJ, Bergman RN, Soeldner JS, Khan CR. Role of glucose and insulin resistance in the development of type 2 diabetes mellitus: results of a 25-year follow up study. Lancet 1992;340:925-9. 8. Barker DJP. Mothers, babies and health in later life. 2nd ed. Edinburgh, Scotland: Churchill Livingstone; 1998. 9. Hofman PL, Cutfield WS, Robinson EM, Bergman RN, Menon RK, Sperling MA, et al. Insulin resistance in short children with intrauterine growth retardation. J Clin Endocrinol Metab 1997;82:402-6. 10. Hofman PL, Cutfield WS, Robinson EM. Premature children are insulin resistant [Abstract]. Sydney, Australia: ICE; 2000. 11. Wright NP, Gibson AT, Carney S, Wales JK. Does being born prematurely have a long term effect on blood pressure and glucose tolerance [abstract]? Montreal, Canada: ESPE Lawson Wilkins; 2001. 12. Busjahn A, Knoblauch H, Faulhaber H, Luft FC, Mu¨ller-Myhsok BC. A region on chromosome 3 is linked to dizygotic twinning. Nat Genet 2000;26:398-9. 13. Barroso J, Gurnell M, Crowley VEF, Agnostini M, Schwabe JW, Soos MA, et al. Dominant negative mutations in human PPARy associated with severe insulin resistance, diabetes mellitus and hypertension [letter]. Nature 1999;402:880-3. 14. Rietveld MJ, van Der Valk JC, Bongers IL, Stroet TM, Slagboom PE, Boomsma DI. Zygosity diagnosis in young twins by parental report. Twin Res 2000;3:134-41. 15. Tanner JR, Whitehouse RH, Takaishi M. Standards from birth to maturity for height, weight, height velocity, and weight velocity: British children, 1965, I. Arch Dis Child 1966;41:454-471 and 613-35. 16. Cutfield WS, Bergman RN, Menon RK, Sperling MA. The modified minimal model: application to measurement of insulin sensitivity in children. J Clin Endocrinol Metab 1990;70:1644-50. 17. Atkinson MA, MacLaren NK, Riley WJ, Winter WE, Fisk DD, Spillar RP. Are insulin autoantibodies markers for insulin dependent diabetes mellitus? Diabetes 1986;35:894-8. 18. Gorsuch AN, Spencer KM, Lister J, McNally JM, Dean BM, Bottazzo GF, et al. Evidence for a long prediabetic period in type 1 (insulin-dependent) diabetes mellitus. Lancet 1981;2:1363-5.
The Journal of Pediatrics May 2004
19. Guaran R, Wein R, Sheedy M, Walstab J, Beischer N. Update of growth percentiles for infants born in an Australian population. Aust N Z J Obstet Gynecol 1995;34:39-50. 20. Hammer LD, Kraemer HC, Wilson DM, Ritter PL, Dornbusch SM. Standardized percentile curves of body-mass index for children and adolescents. Am J Dis Child 1991;145:259-63. 21. Tyrrell VJ, Richards G, Hofman P, Gilles GF, Robinson E, Cutfield WS. Foot to foot bioelectrical impedance analysis: a valuable tool for the measurement of body composition in children. Int J Obes 2001;25:273-8. 22. Lee A, Ader M, Bray GA, Bergman RN. Diurnal variation in glucose tolerance: cyclic suppression of insulin action and insulin secretion in normal weight, but not obese, subjects. Diabetes 1992;41:742-9. 23. Meirhaeghe A, Fajas L, Helbecque N, Cottel D, Lebel P, Dallongeville J, et al. A genetic polymorphism of the peroxisome proliferator activated receptor gamma gene influences plasma leptin levels in obese humans. Hum Mol Med 1998;26:519-25. 24. Baron H, Fung S, Aydin A, Bahring S, Luft FC, Schuster H. Oligonucleotide ligation assay (OLA) for the diagnosis of familial hypercholesterolemia. Nat Biotechnol 1996;14:1279-82. 25. Trinder P. Determination of blood glucose using an oxidase-peroxidase system with a non-carcinogenic chromogen. J Clin Pathol 1995;22:158-61. 26. Pilcher C, Elliot R. A sensitive and reproducible method for the assay of human islet cell antibodies. J Immunol Methods 1990;129:111-7. 27. Bergman RN. Lilly Lecture 1989: Towards physiological understanding of glucose tolerance. Diabetes 1989;38:1512-27. 28. Osei K, Cottrell DA, Orabella MM. Insulin sensitivity, glucose effectiveness and body fat pattern in non diabetic offspring of patients with NIDDM. Diabetes Care 1991;14:890-6.
Insulin Resistance in Healthy Prepubertal Twins
29. Grunfeld B, Balzareti M, Romo M, Gimenez M, Gutman R. Hyperinsulinaemia in normotensive offspring of hypertensive parents. Hypertension 1994;23:112-5. 30. Poulsen P, Vaag A, Beck-Nielsen H. Does zygosity influence the metabolic profile of twins? A population based cross sectional study. Br Med J 1999;319:151-4. 31. Cook JS, Hoffman RP, Stene MA, Hansen JR. Effects of maturational stage on insulin sensitivity during puberty. J Clin Endocrinol Metab 1993; 77:725-30. 32. Barker DJP. Early growth and cardiovascular disease. Arch Dis Child 1999;80:305-7. 33. Simmons RA, Templeton LJ, Gertz SJ. Intrauterine growth retardation leads to the development of type 2 diabetes in the rat. Diabetes 2001;5: 2279-86. 34. Dwyer T, Blizzard L, Morley R, Ponsonby A-L. Within pair association between birth weight and blood pressure at age 8 in twins from a cohort study. Br Med J 1999;319:1325-9. 35. Zhang J, Brenner RA, Klebanoff MA. Differences in birth weight and blood pressure at age 7 years among twins. Am J Epidemiol 2001;153:779-82. 36. Hrubec Z, Robinette CD. Study of human twins in medical research. N Engl J Med 1984;10:435-41. 37. Kyvik KO. Twin research: nature versus nurture in common diseases. In: Day INM, Humphries SE, editors. Genetics of common diseases. Oxford: BIOS Scientific Publishers LTD; 1997. 38. Tarrade A, Schoonjans K, Guibourdenche J, Bidart JM, Vidaud M, Auwerx J, et al. PPAR gamma/RXR alpha heterodimers are involved in human CG beta synthesis and human trophoblast differentiation. Endocrinology 2001;142:4504-14.
613